The video did not make predictions about the future, it just presented the current reality at the time. I think it is still a good introduction to set the right expectations when encountering this technology.
Maybe. However there is fundamental physics in play and so it is likely someone (not me!) can tell you the most efficiency we get from that system. I'd be curious what those people say.
I haven't watched it recently, but here are the main takeaways I remember:
Peltier coolers are neat because they're very small and quiet - as opposed to vapor compression systems solutions. However, they are an order of magnitude less energy efficient.
Also Peltier coolers still have to obey the laws of thermodynamics, which means that to cool one side of the mechanism, you must heat the other side. In order to do any substantial cooling, you need a way to dispose of that heat on the other side. This usually involves the use of radiators and fans, which negate much of the size and noise benefits.
As a result, Peltier coolers are pretty niché. Your use case would have to require only a little bit of cooling. You'd have to need a form factor that cannot accomidate a vapor cooling solution. And you'd have to be willing to make the system very energy inefficient.
AFAIK, no one has tried to build a Peltier cell paired with a heat pump. I am not an expert, but I would imagine that it's a path that could bring higher efficiencies. Thoughts?
Also not an expert, but I’m struggling to find a combination where one of those couldn’t be replaced with a passive thermoconductive element. It’s hard to beat the efficiency of “free”.
I think there are some applications though. I remember PWMing a peltier element to cool something to more or less exactly 35°C. I didn't need to be efficient cooling, it just needed to be reliable under space constraints.
I am no hardware guy and I remember there was a giant heatsink despite the constraints. It was some kind of photosensor + lamp if I remember correctly.
I think there was also some software logic to reduce water condensing at something too cool compared to the ambient temperature.
• Fundamental Design and Coefficient of Performance (COP): Peltier elements are technically heat pumps. However, their efficiency, measured by the Coefficient of Performance (COP), is inherently low, ranging from "somewhere between zero and very bad" in practical applications. In theory, with perfect conditions and very low current, a COP between 1 and 2 might be achieved, but this generates hardly any cooling. In contrast, vapor-compression heat pumps used in traditional refrigerators can achieve a COP of 3 or more, meaning they can move significantly more heat energy than they consume in power.
• Impact of Temperature Difference: The efficiency of a Peltier element varies wildly depending on the temperature difference it is working against. The materials used to construct a Peltier element conduct heat even when not running. The greater the temperature difference between the hot and cold sides, the more heat energy leaks back through the element itself, dramatically reducing its efficiency.
• Internal Heat Generation from Current: The more current you attempt to push through a Peltier element, the more heat it generates within itself, which further diminishes its cooling performance.
• Comparison to Traditional Refrigeration: A small personal refrigerator using a Peltier element was found to consume about 55 watts of power constantly. This is significantly more energy than a standard mini-fridge, which might average only 21.7 watts to keep items cold over three hours because its compressor doesn't run all the time due to a thermostat. Even a much larger refrigerator, holding over 50 times more items, uses slightly less energy annually than the Peltier-based toy fridge.
For those worried about tiny COPs from these gizmos, trawling through the actual paper -- as well as the PR from JHU APL -- in this HN post [1] shows claims of COPs of ~15 for Delta Ts of 1.3°C.
A compressor based cooler gets a COP of about 4 in the real world. I'm pretty sure this is an apples to oranges comparison to an expert (I am not one of those) but a factor of 3+ increase in COP is fairly noteworthy -- if it holds up.
ugh, reading the paper, their methodology is kinda crap. They basically just guess what the thermal resistances in their system are, and use air temperature measurements to figure out the heat flow. This is not how to measure heat flow accurately. It might be OK as a comparison with whatever TEC they tested with, maybe, but it's not at all something I would trust to compare to another test setup. If their box is more insulative than they think it is, their results are gonna look better than reality. This can be validated at least approximately by just putting a heater in the box that's dissipating a known amount of heat and looking at the temperature rise, but it seems they didn't even do this. And in general the regime where you've got small temperature differences is where your systematic error in a system like this can become huge and distort the results by multiples.
(This is an area which is really hard and details matter. Heat is basically impossible to measure directly, and the indirect measurements are fraught with peril. Getting it wrong was a large part of why people thought they had demonstrated cold fusion)
Why, what would happen to my lettuce if I set the thermostat near the cooling element to e.g. 3°C and have the cabbage be somewhere between that and 4.3°C?
Unless I'm very mistaken, this 1.5 degree delta is between the hot side (the condenser at almost your room temp) and the cold side (evaporator inside the fridge).
That is, if you want to chill your cabbage to 4.3 degrees C, your room temp should be not more than 5.6 degree C. Or... if it's 25 degrees outside, in an ideal case your cabbage is 23.7 degrees cool.
Btw, max theoretical COP increases rapidly as T'hot - T'cold shrinks to zero. Your practical AC or fridge delta T is much more than that, depending on the weather.
The delta of 1.3C is critical there - peltier cooling drops precipitously in efficiency as the delta increases, and struggles to hit a COP of even 1 in real world scenarios. Their figure works out at about 6.5% Carnot efficiency, whereas a normal heat pump is usually nearer 45% over a much broader range of temperatures, as you can separate the hot and cold sides completely. Not so with a peltier wafer.
What they’ve done here is add a point of failure, use additional materials as well as a traditional heat pump, and called it “AI” and “eco friendly”.
They're already doing great. I have a portable neck AC unit that uses peltiers and it keeps my head and neck cool in otherwise nasty desert conditions when mining. Yes the radiative heating from the sun is still a bitch but a hat basically minimizes that, and also redirects the partially-chilled air more efficiently around my head and face.
The direct-contact neck cooling plates are an absolute lifesaver. Keep the sun off the back of your neck and chill one of the best heat sink locations exposed on your clothed body.
There's a bajillion of them around if you search "neck cooler" or similar. Very simple product sticking a few commodity items together. Some do only have fans though.
When I looked a while ago there wasn't really a clear winner or high quality unit. There is the "Coolify" series that are much more expensive but still somewhat middling reviews overall.
The idea was just so astonishing that I ordered some from American Science and Surplus. I connected the leads to a battery and poof, one side got hot and the other got cold. Blew my mind.
I didn't actually have a use for it. It was just neat that it actually worked.
I understand the basic physics of it perfectly well. It's just one of those things where you expect basic physics to be overwhelmed by friction or something.
I don't know how people can convince themselves that they can understand these effects.
Unusually effective heat insulation is exactly how they work. (This is tempered by eg radiative losses, so making them thicker doesn't work better.) placing poorer heat (vs electrical) insulators between peltier material is counterproductive, similar to using resistors to improve conduction between copper wires
Don't ask me for a better explanation :)
As to why COP or even Carnot efficiency hasn't been thrown out in favor of temp-difference independent efficiency metrics like exergy.
Indeed, but in practical use (like a mini refrigerator), one will want to insulate the hot and cold sides from each other.
If one makes the walls thick, then they end up with a hole for the Peltier device and somehow sandwich two heatsinks on the device while maintaining insulation around it.
Perhaps easier to deal with in CPU cooling and such since one side is simply smacked into the thing being cooled.
What is exergy? The one time a mechanical engineering colleagues tried to explain it to me, he reached the wrong conclusion on the problem we were working on.
I haven't seen it in any physics thermodynamics book, and only mech eng. seem to know what it is, and then only in the US.
Faires (undergrad MIT Thermo book from the 50s) makes no mention of it as far as I can tell.
But that isn't a mathematical expression. At best, it would appear to be energy * maximum_Carnot_efficiency (for heat engines anyway)
But it seems not to be adding very much, since Carnot efficiency depends on delta_(T). The OPs point that exergy doesn't depend on T is tautological since T has already been accounted for by the Carnot expression.
One can only hope some kind of phonon diode material can exist that a slight voltage can overcome something so inescapable as entropy by providing it only lanes that suit us.
I believe the “apples to oranges” is the temperature gradient. AC units would routinely manage 15-20c and are rated for more than that. And some freezers manage up to 50c. The greater the gradient the worse the efficiency in general.
What about a ΔT of say 20°C? I'd reckon most refrigerators and air conditioners are around there (temp difference of refrigerant between evaporator and condenser).
Stacking a bunch of these Peltiers to give more temperature difference would give a pretty low CoP. Say, for a 13°C temperature difference you'd have to stack 10 of them and use 10x the power. It's even worse actually as the hotter ones have to also pump the waste heat from the cooler ones.
Note that a small temperature difference that is sustained very consistently over a long time using a tiny amount of electricity (let's say half of what the parent post cited, so like a COP of 8) could add up to a lot of nearly-free cooling. You'd chill your walls for weeks and when a heat wave comes with hot nights for a week, if you(r home automation) close(s) the blinds during the heat of the day, the more-powerful AC might barely have to do anything
Just an idea of course, but I'd not write new tech off as "ok but just 1.3 degrees who cares" when the claimed COP is so insanely good without first trying it out
Chill for weeks? What kind of thermal mass are you thinking of there?
A brick wall is on the heavy side and you'd be able to store a whopping 1/8 of a ton of AC in ten feet of brick wall. It'd save about 0.1kWh of AC power use. A whole house will have many lengths of wall but also multiple tons of cooling requirements, so that doesn't help much. And how are you going to distribute those small temperature differences without wasting a ton of power?
And that's after you figure out how to trap a temperature difference in your walls for more than a day.
The ΔT between evaporator (typically 2-6°C while cooling) and condenser (often 40-50°C in cooling mode) is much higher than 20°C. The condenser is often almost 20°C above ambient outside temperature.
The design ΔT of ~10°C is the typical return-to-supply air ΔT.
>in this HN post [1] shows claims of COPs of ~15 for Delta Ts of 1.3°C.
>A compressor based cooler gets a COP of about 4 in the real world.
Real life refrigeration usually isn't very interested in a difference of 1.3 C. The Carnot COP for this temperature drop near ambient conditions is, I believe, around 200. When you consider a cooling technology relative to the Carnot efficiency (or COP) you get a better idea of what the efficiency means in practice. For an AC unit blowing 10 C air on a 40 C day, the Carnot COP is about 10, while real units get less than half that. But I think that's still better than the Peltier effect getting less than 10% performance relative to Carnot limits.
You can get a COP of ~3-4 out of regular TECs, but only at pretty low temperature differences. That's the killer, fundamentally the TEC material itself is thermally conductive and heat really wants to flow back the other way, so no matter how well it moves the heat, it winds up fighting against the heat load generated by itself. A refrigerant based heat pump works much better because the heat basically only moves in the direction the refrigerant itself is moving.
TECs are wonderful little devices with operating characteristics unlike comparable devices.
They can be designed to move a specific amount of heat or to cool at some delta-T below the hot side (and due to inefficiencies the hot side can climb above ambient temperatures too, raising the “cold side” above ambient!)
I ran through a design exercise with a high quality TEC and at 8°C delta-T for a wine cooler you could expect a COP of around 3.5–4 (theoretically). This is pretty good! But below the 2.5V max to do that you’re only able to exhaust up to around 40W. For a wine cooler this is not so bad. For a refrigerator it’s a harder challenge because the temperature drops when the door opens, and if someone sticks in a pot of hot soup, it’s important to eject that heat before it raises the temperature inside to levels where food safety becomes a problem. For a CPU it’s basically untenable under load because it’s too much heat entering the cold side thus temperatures will rise.
- Most TECs are cheap and small and come without data sheets, so people tend to become disillusioned after running them too hot.
- You have to keep the hot side cool or else the delta-T doesn’t help you. For a wine cooler this is probably no big deal: you can add a sizable fan and heat sink. For CPU cooling it becomes a tighter problem. You basically can’t win by mounting on the CPU; they are best at mediating two independent water-cooling loops.
- Q ratings are useless without performance graphs. It’s meaningless to talk about a “100W” TEC other than to estimate that it has a higher capacity than a “20W” TEC.
- Ratings and data sheets are hypothetical best cases. Reality constrains the efficiency through a thousand cuts.
When I think about TECs I think more about heat transfer than temperature drops. If you open a well-insulated wine cooler once a week then once it cools it will only need to maintain its temperature, and that requires very little heat movement. Since nothing inside is generating heat you basically have zero watts as a first-order approximation. For the same device mentioned above, it stops working below 1V, and at 8° delta-T that’s a drop in COP to around zero but it’s also nearly zero waste. If you were to maintain a constant 2.5V, however, it would continue to try and pull 40W to the hot side. This would cause the internal temperature to drop and your COP would decrease even though the TEC is using constant power. The delta-T would in fact increase until the inefficiencies match the heat transfer and everything stabilizes. In this case that’s around a 20° drop from the hot side, assuming perfect insulation.
Unlike compressors, TECs have this convenient ability to scale up and down and maintain consistent temperatures; they just can’t respond quickly and dump a ton of heat in the same way.
COP of peltier elements can be large only when the temperature difference is small, such as the measly 1.3 degrees you quoted. When do you want to cool something by only 1.3°C compared to the surrounding temperature?
I don't understand how they could quote him saying: “This thin-film technology has the potential to grow from powering small-scale refrigeration systems to supporting large building HVAC applications, similar to the way lithium-ion batteries have been scaled to power devices as small as mobile phones and as large as electric vehicles,”
Then the entire article foregoes comparing their peltier device to traditional compressor based heat pumps.
Only if you have a cold environment into which to dump that heat into. You could maybe trickle charge your phone in the winter, if you don't mind a cold spot where the device sits on your body.
I wish they had gone into detail about what makes this device different from Peltier coolers that are available today. You can already get mini fridges that use them, but they're not very good - the total amount of cooling you can get from them is not enough to maintain the temperatures that a regular refrigerator does.
Yeah I have a cooler box with a peltier cooler, but it will only cool a few degrees, not sure I'd trust it as a standalone fridge. Plus its energy use is much higher.
They claim 75% better. I doubt this is better than compressors in general but they seem to be going for a niche where a high power compressor does the high load part but when only a litte cooling is needed the compressor is now very inefficient and so the peltier is better. Normal fridges just let the temperature have a wider swing which is good enough for most needs.
> Normal fridges just let the temperature have a wider swing which is good enough for most needs.
These wide swings annoy me. You hear that you shouldn't let your fridge go above 4°C, because that's dangerous. And you obviously don't want your fridge to go below 0°C. But finding a setting where the hottest part of the fridge doesn't go above 4°C (or even 5°C or 6°C) during the hottest part of the cycle and the coldest part of the fridge doesn't go near 0°C during the coldest part of the cycle is really pretty difficult, in my experience.
Put a thermometer in a plastic bag (to keep it dry) in a jar of water in your fridge, you'll likely find the temperature of items inside the fridge much more stable than the air temperature.
The yogurts at the back of my fridge still freeze up.
There's going to be a temperature gradient in a typical fridge, if only because one side is getting opened every now and then, and the other is separated by all the products which are inside.
I suppose what they're trying to do here is even out that gradient without running the compressor.
The time it stays above 4C unless you're opening the door of your refrigerator all the fucking time is mostly negligible in terms of the difference it will make for bacterial growth.
The foodstuff itself is loaded with water, it won't have an excursion dangerously above 4C just because you opened the door and the air temperature raised a few degrees.
If you are really worried about it (you shouldn't), and you don't keep your refrigerator full, add a few water bottles for thermal mass.
Why should that be dangerous? I have never heard that.
I have always had my fridge at 8°C and never had something dangerous happen to me. I have never come across fridges that were way cooler, apart from fridges of friends in Canada and the US. What's the reasoning?
The recommendation I've always heard in the Netherlands is 7°C, it's more recent that I've been seeing 4°C on meat packaging in Germany (where I live now). I doubt anyone's fridge is consistently at or below 4° without freezing things constantly, so I've been assuming this is wishful thinking and/or ass-covering on the manufacturer's part and not what anybody actually does. Your 8°C is close enough that it probably makes little difference, though afaik this is an exponential curve (at 14°C it would last far less than half as long) so I'd not be surprised if things spoil a bit sooner than they otherwise would
Even if your products generally meet their "should be safe at least until" date (Mindesthaltbarkeitsdatum, idk if it's the same as "best before"), you might exceed that longevity more often than you do now and thus have less food waste by setting the fridge colder - if food waste is a thing you have in the first place (I'm the type of person that is hungry all the time, opens the fridge when hungry, and isn't super selective (among what I've bought anyway), so food I buy ~always gets eaten before it spoils, but then when I see food waste numbers, apparently that's not the case for everyone so I'm just throwing this out there)
Edit: trying to fact-check myself, I can't find any trustworthy source in Dutch saying your foodstuffs fridge should be more than 4°C. I measure new fridges when moving in and again at least once during the first summer to make sure they stay at or below 7°C when we had the door open a normal amount of times, so I know they're that (and not much cooler, to not freeze items or waste energy). So far, products meet their minimum shelf life date thingy and almost always exceed it. Strange. Maybe this recommendation I heard predates the internet (showing my age here), or maybe every page on the internet assumes that nobody actually measures it properly and so they recommend a value that's half of what's actually safe?
Crazy, I'd think the fluctuations would be way too big. How long does it normally run for when it does a cooling cycle? Is that temperature throughout the fridge or only on certain shelves?
Ours (well, our landlord's) runs very consistently for 40 +/- 3 minutes, with just over 1h30min in between during normal use, or 1h54min if it was closed the whole time (like at night). The temperature of products, as measured with a cheap infrared thermometer that's probably off like 10%, varies between 2 and 7 degrees, but it's not very consistent between shelves (top seems warmer but then the very bottom one, that is only half deep, is as warm as the top again). The products I checked have all been in there for days; reflectivity may be part of the difference, not sure. I don't know what the air temperature difference is at the beginning and end of a cooling cycle though
You should buy a fridge with a fan in the cooling chamber. My Samsung fridge has a nice fan and associated ducting to circulate air and keep temperatures ~uniform.
I think the 4C recommendation is an average. There is not magic in ~1% temperature difference from 3 to 6. Just slightly different rates. Which again slow, but do not stop when it is at low end compared to high.
I suspect it's that. Most inexpensive refrigerators don't have thermostats. Which seems insane; it cannot add that much to the price.
8C is a perfectly fine temperature for a wine fridge. And they usually have thermostats because a wine fridge is a luxury item. As opposed to keeping people from getting foodborne illness.
> Which seems insane; it cannot add that much to the price.
Be careful what you wish for.
When buying my current fridge, I specifically tried to go out of my way to avoid complexity, but the opening for my fridge is an odd size so my choices were limited. The only fridge I could buy that didn’t have a bunch of crap that’s was guaranteed to break (ice maker, water dispenser [seriously? aren’t most fridges right next to a faucet?], LCD-covered glass panel, etc) that also had a thermostat had a digitally-controlled thermostat. No knob or physical buttons, just a capacitive surface for temperature adjustment and some LCD screens showing the fridge and freezer set temps. (Not the actual measured temps, that would be too useful, just the set temps.)
In hindsight, I probably should’ve just gotten one with a regular dial, but I was a bit fixated on the “real” thermostat. So now I’ve got that to look forward to breaking in 4-5 years and figuring out where the hell to source a discontinued fridge LCD panel from.
Many people prefer a cold glass of water compared to the cool to room-temperature water that comes out of the tap (Hence the refrigerator for cooling). Filtering at the individual level is typically done for flavor, because depending on where you live, while the water is probably safe to drink (assuming HN readership demographics), it may or may not be particularly pleasant. Even here in the Bay Area, where we have that sweet, sweet Hetch-Hetchy water, I know people who live in buildings with pipes bad enough that they filter their water.
There are many applications for peltier cooling in industrial settings, i.e. sensor devices (or, actually, microscopes with CMOS sensors). Those have to be temperature stabilized to minimize noise while also be IP rated, i.e. no air cooling. 75% efficiency gains would be awesome here.
Would a peltier element allow for a more constant temperature? Or can you turn a peltier up and down easily? With a compressor based system it's always been "on or off", something that can ease up and down would be nice.
Traditional Peltier devices operate at ~10% efficiency (COP of 0.5-0.7) compared to vapor-compression systems (COP of 2-4), but recent advances in thermoelectric materials like bismuth telluride alloys and segmented elements have pushed lab efficiencies to ~15-20%.
OK, a typical Peltier device has 3.5% coefficient of performance, that is, it produces 35 W of cooling per 1 kW consumed.
Fine, let's expect that the new tech doubles the efficiency, to 7%. Still, to my mind, pretty wasteful, on par with a steam railway engine. A Peltier element is good in cases where you can afford a large heat removal device, but need precise temperature control and no moving parts. For a home fridge, I'll take the sound of the compressor and the temperature fluctuations of a 400% efficient compressor-based heat pump over a Peltier element any day.
A common use case: coolers. You don't want a whole compressor, but you do want to keep your hot dogs safe and your beer drinkable. It's not very efficient but it's enough for short periods.
There are mini-compressor based coolers available now if you look for them. They cost a little bit more, and obviously have more weight than a peltier, but I think are worth it if a bag of ice isn't viable by itself because they will run for way longer than a peltier setup.
I considered putting some in our under-sink reverse osmosis tank, to cool the water. Couldn’t come up with a way to exhaust the heat well enough to even look at how much electricity it would have cost me though - probably too much to make it worth it.
Oh, indeed I was wrong; COP of Peltier elements is much higher than I had gleaned from online charts, not the order of 0.03 (3%) but can easily reach values above 1.0 (100%) and be e.g. 0.5 (50%) at ΔT = 30C, enough for a home fridge.
Still a bit far from compressor-based designs, but not negligible, and almost doubling the efficiency is indeed a serious advance.
That is wrong, COP is expressed as a ratio, not a percentage (efficiency is expressed as a percentage, which is COP * 100). And as others have said, both efficiency and COP are dependent on ΔT both in refrigeration and thermoelectric cooling.
A steam railway engine is a lot better than you would think. They were more efficient than diesel engines when diesel took over - the diesel engine needs much less human labor and so was cheaper overall, but for efficiency steam was better. (Note that diesel technology has improved since the 1950s, so I don't know how they compare now)
> In the Bespoke AI Hybrid Refrigerator Samsung launched in 2024, the compressor operates under normal conditions such as routine storage and retrieval, while the Peltier device activates alongside the compressor during high-load situations — like when storing large amounts of groceries or placing hot food inside — thereby enhancing both cooling performance and energy efficiency
With home HVAC, fridges, water heaters, and dryers all using now able to use of dependent on heat pumps I wonder how long it be before we see modular appliances that connect to coolant lines where the temperature differential is supplied by a central high efficiency heat pump.
Cars already have heat scavenging that can move heat from where it's being created through losses to places where it's valuable, like the cabin or battery pre-heating. Especially in cold climates it feels like homes should be next.
There's some commercial options for this, but it's not common. Usually, these devices just have their own compressors, because they all pale in comparison to the heat pump(s) used for climate control. For example, I have a HP water heater, and its heat pump is about 1/3 of a ton, whereas most homes need 3+ tons for climate control. Fridges are a fraction of that.
For HP clothes dryers, there's no efficiency to steal from somewhere else, because they use both the hot and cold coils - similar to (the same, really) dehumidifiers.
The tradeoff would also be running high-pressure refrigerant lines everywhere. That would require EPA certification (in the US, anyway) to connect/disconnect an appliance, and it would probably be less reliable. These sealed-system units are generally pretty reliable, because the refrigerant is installed at the factory under ideal conditions, and there's no connections that are made later that may be done poorly.
That is an interesting thought, but I assume that the working ranges of the different appliances are different so there would be some complexities and inefficiencies getting them all connected to a common circulation loop. If there was a thermal equivalent of a transformer used for alternating current, that would be amazing.
As far as I know all the common commercially available heat pump appliances all use the same refrigerants, so it doesn't seem like it would be that challenging.
In cars that have unified heat management the refrigerant cycle is handled as a separate element, with a manifold controlling individual coolant loops to each component. I'm picturing something similar for the home, with a coolant moving heat to and from each appliance using standardized communication to the manifold. There would probably need to be heat buffer tanks, but air to water heat pump systems for radiant heat already need this anyway.
A few years ago I was planning to build a velomobile that I would live out of for a year while circumnavigating Australia, and potentially indefinitely. (My plans changed.) I was disappointed at how hard refrigeration information was to come by (maybe I should have sought a traditional paper book), but I was kinda looking forward to figuring out if I could use one compressor to cool a small fridge, cool and perhaps heat the cabin, heat water, and heat a slow cooker (target 80°C). The bits I could work out suggested you might want a different refrigerant for the cooling and heating applications, or different back pressures; but I was rather hampered in my reckoning by my lack of domain knowledge—I was definitely going to have to talk to professionals! If so, and combined with the limited power collection available (<1m² usable solar panels on the vehicle, could lay out more while parked), butane was probably going to make more sense for cooking.
I even ran some naive numbers on the amount of water that would condense in expected conditions, concluding it could be handy but I’d probably still need to source more water.
It's worth noting that the very earliest electric refrigerators had a separate condensing unit outside; see this interesting 1920s Frigidaire training video for an example of what that was like: https://www.youtube.com/watch?v=W-t7DqOAMME
There were also centralised systems for apartments where one condensing unit supplied many evaporators in the refrigerator in each suite.
That would've been easily invalidated by prior art from nearly a century ago; as I mentioned in a sibling comment, this was a common arrangement in the early days of domestic refrigeration.
I used one for a couple years as my primary fridge. It was expensive, like $2k, didn't have very good temperature control and broke after 2 years and couldn't be repaired.
The latest wave of appliances is really fucking loud for some reason.
I think they're using different kinds of motor windings, bearings, insulation, etc. it's not related to the refrigerant or other system parameters. I've had older r600a fridges that were dead silent compared to anything sitting in a Best Buy showroom right now.
Likely high speed compressors --- the oldest hermetic systems used an induction motor running at 1800 RPM, then later they went to 3600 RPM, and now they're running on a VFD that possibly goes much faster. By making it pump faster, they can use a smaller compressor and reduce costs, at the expense of longevity and noise.
They use lighter lubricant oils than were historically used, which allows more vibration. It’s also why compressors burn out far more quickly than they used to.
I'm not the same person, but I do not (and never have) noticed any noise when my fridge is running. Whether that means we have different fridges or different tolerances for noise, I'm not sure.
More likely you have a noisier house/neighbourhood. I used to think my PC was silent until I moved to a quiet place and then I could suddenly hear it very clearly.
My wine fridge uses Peltier and is super quiet. It's the perfect application for this because wine doesn't need to be as cold as a normal fridge, and noise is a consideration.
It's not completely silent though, there's a small PC-like fan but it's way less loud than a compressor.
A hotel I was staying at had a small bar fridge that used a Peltier. I only know because it stopped working so I checked it and realized it was only a Peltier plus a heat exchanged (a cyclopropane loop).
I presume a full size fridge is outside of reach at this point.
Move it outside a cabinet, let it free stand. I found out that my nice kitchen niche for the refrigerator acted like a nice resonance chamber for the frequencies the compressor generated.
Not OP but it's a massive nuisance if you live in a studio. People don't realize how noisy a fridge is until there's one in the room that they sleep in.
New appliances are far better than old ones here. Especially old ones that (I assume) haven't been maintained and so are working far harder than they used to. I've lived in places with old ones that were fine and old ones that were awful, both. I've had much more consistently good results in places with newer ones.
Skill issue on the manufacturer's part. I live in a studio and never hear the fridge. This is part of a fitted kitchen, though, but I doubt the panel hiding the fridge makes that big of a difference.
Just today I ordered a 32dB Liebherr; the previous one had 35dB and could be heard all around the studio (I measured the noise using a dedicated sound meter).
I remember back in 2006ish, my PC modding days, buying some peltier modules from eBay and (attempting) to get subzero temps out of my CPU cooling.
If I recall correctly I got the setup powered but stopped short of actually putting it on my CPU when I couldn't mount it all in a way that would let me contain the condensation with what I had.
Maybe it was pccooling or pccasemods dot com? There was a really strong community forum back then where it was all going down, people were nitrogen cooling their PCs, watercooling was a big deal, and CPU temps of 60c were considered unsustainable.
I still overclock my computers but usually my aim is a silent computer under 60% load, so my goals have changed. Peltiers are not something I see taking over PC cooling even now. You still need the same radiator capacity, the peltier just moves the heat away faster and can get below ambient temps at the CPU.
Given that almost all CPUs these days are thermo-throttled, maybe a controllable Peltier element will be the next Turbo button? I remember that Peltier based cooling setups were electricity hungry.
I thought Peltiers can't lower the temperature by more than 40 degrees F, in practice less than that. This is not cold enough for a refrigerator on a warm day.
FWIW, in a typical apartment or single-family home, refrigeration uses a fraction of the energy that space cooling (also via a refrigeration/vapor compression cycle) requires on a warm day (and probably year round too unless in very mild climates). The psychrometric chart path is different so there are of course differences in the amount of energy required for the sensible and latent components, but the real difference is just the volume of air that needs to be dealt with.
My point being that at least from an energy and carbon perspective, lowering the space cooling demand via more effective building envelopes or increasing the space cooling supply efficiency - eg via membrane or dessicant dehumidification, better heat pumps etc) is far more impactful on a macro scale than better refrigeration.
Granted refrigeration in a warehouse eg is really also space cooling, but I’m just making the distinction between the dT=0-25F context and the dT>25F context. If I could only choose one technology to arrive at scale to improve the efficiency, it would be for the former context.
Also the area that needs insulating, and in the extremes the amount of air that needs to be exchanged with the outside to make the house livable, and the heat generated by the people living in it (stick a 100W lightbulb in a fridge and see how cold it can get).
The insulation is actually solvable, and for heating can basically remove the power requirements: a house heated and using heat exchange on air leaving vs entering can be heated a lot just by having people inside it, let alone the other energy they use for other purposes. It's just more expensive to build this way, and with cheap energy it can a long time to pay back. Cooling you can't push down past the heat generated inside the house divided by the COP of your cooler, though.
Yep, PassiveHouse standards which typically include an extremely tight envelope which forces installation of outdoor air supply famously can get away with just a few hundred watts of heating capacity because of heat exchange on the incoming and outgoing airstreams!
Sure I was playing a little fast and loose there, but (a) the large surface area of the home (and resulting conductive transfer through the walls + convection transfer via infiltration through gaps) is directly a result of the fact that you need a significantly larger volume for humans to move around in and live in than you do to store food and (b) even if we do look directly at the volume of air, the difference is significant since at the end of the day, since for any given constant deltaT, your energy spent is still linear with mass or volume. And we are talking about roughly 2-3 orders of magnitude difference in air volume between a house and a refrigerator.
Anyways, if you write out all of the heat balance equations, you get a few W/m2 of flux on the inside wall of the home and a few W/m2 of flux on the inside faces of the fridge, assuming a typical wood frame construction in summer time and steady states all around.
So yes, of course multiplying the flux through the home’s wall by the surface area of the home results in a massive heat gain value compared to the heat gain conducted through the surface of the refrigerator, but that’s arguably precisely because of the two different volume requirements.
There are no inherent max temperature deltas for Peltiers, just coefficient of performance, or watts moved per watts wasted, is atrociously low compared to just about everything else.
Peltier coolers have non-zero thermal conductivity and are pretty thin. At some point, heat leakage from the hot side to the cool side catches up to the rate of heat removal the cooler is capable of. What that point is will vary between devices, so there's no single or fundamental limit.
You can also stack them, but you need ever-increasing areas of TEC and corresponding power consumption. I think a triple-stack is sometimes practical if you want to get something to well below ambient and it doesn't really generate much heat itself (and you don't really want to use a traditional cooler due to size or vibration or something similar)
I worked on a scientific instrument a while ago, it had a Peltier heater on it to raise the sample to 60c from whatever residual environment temp was (approx. 20c in a lab). It was pretty amazing to see my old overclocking cooling solution come back around and be used in my professional career some 20+ years later
The TEC is more than 100% 'efficient' - you get more heat out than power you put in. Though I would guess it's mainly there to get heat control that can span above, around, and slightly below ambient.
If you are reasonably above ambient and the control loop is well tuned things work well, unless you are significantly perturbing the system frequently.
Honestly, I didn’t ever ask. As the SWE on the project I just had to switch a DIO on/off based on the reading from an AIO. I just trusted that the EE and ME folks were making smart decisions around BOM.
AI has literally nothing to do with this. Why do they feel the need to sprinkle the phrase everywhere? AI inverter/compressor? Come on, have some sense of shame, please.
It's the name of the game, alas. In the area I work in, I've seen many companies get many multiples of the investment we've gotten because it's "quantum", even though they're trying to do the same thing we're doing and doing less well at it.
The head of the board of a company I work with, who uses ChatGPT to get information on how to prune his plants and barely knows how to use Windows Explorer, decided that the whole company needs to go full AI. Everything has to be AI, and everyone has to learn AI. I thought that HN has been overdoing it with AI articles recently, but parts of the outside world seem to be much, much worse.
Actually AI has a lot to do with material design. I can't do a proper explanation, so please bear with this information dump:
- We don't have a strong physical theory for solid state physics. Quantum stuff doesn't scale well from 1 atom to a mole of atoms, because 10^23 goes into the exponent of number of energy levels in the system, and then we also have to model interaction of those levels.
- Physical properties of materials depend on their crystal structure, unevenness of that structure, spectrum of size of crystals, temperature, pressure, fields they're exposed to and current position of stars in the sky. Even the "wrong" solid state physics equations we have are highly non-linear.
- State-of-the-art effects are usually achieved with combinations of such materials.
- Exact parameters of the process used to put those materials together radically change the behavior of the system. Put that nanolayer with a different sort of vapor deposition or at different pressure, and the thing will stop working. Ever wondered why we don't produce all the neodymium magnets at N55 grade? Because even precise description of the process is not enough.
- AI doesn't care about the exact physics, but is sometimes very good at navigating in the large parameter space.
- Google have recently made AI that predicted thousands of novel material structures. They found more materials than were found by all human research over the whole humankind history.
I wouldn't expect AI to explain what's going on in solid-state physics anytime soon, but exploring crystal structures, doping, and process parameters automatically might actually get us new materials a couple hundred years faster.
What the article is talking about is using temperature probes and deciding if to run the compressor or peltier, or both depending on the temperature. That's what they are calling AI. Some if statements on a microcontroller.
In the article, the picture shows the peltier cooler is on top and the compressor is separate on the bottom. The "AI" is an adjective for the compressor only. There is little to no discrete logic in a compressor, so calling it that is nonsense.
This still feels like early days. The fact that they're sticking with hybrid systems for now suggests we’re not quite at the “compressor-free” future yet. Also, the environmental angle is compelling, eliminating refrigerants would be huge
Remember that Peltier coolers don't make heat disappear - they just move it from one side of the cooler to the other, and produce a lot of additional waste heat in the process. There are better ways of transferring heat from a hot IC to a heat sink nowadays - like liquid cooling for really high-performance systems, or capillary-action heat pipes for more typical needs.
> There are better ways of transferring heat from a hot IC to a heat sink nowadays
Peltiers were always a bad way to move the heat. What they offered was the ability to go below ambient, which at the time could improve overclocking. Peltiers of course lacked the capacity to actually take you there with any decent load, but in theory it could.
It never really made much sense for consumers, and once consumers realized that, the market went away.
TDP went up. TECs are extremely inefficient, and have limits on how much heat they can pump in a given area anyway. a 10W chip, you can maybe get away with, and it might be a meaningful improvement compared to the other cooling tech available at the time, if you don't mind the power usage. with a 100W chip, you have no chance.
(You can see a demo here where LTT try it, and they, after dumping 500W into the cooler, can get the CPU to a vaguely reasonable temperature, until they actually load it up: https://www.youtube.com/watch?v=sWrqyQWfhrs)
I'd consider performance processors something like a threadripper - 300W unlockable to 1000W. I haven't kept up with intel but I think they were indeed riding 200 to 300w.
They made both the problem of power consumption and removing waste heat infinitely worse. Instead, I used liquid cooling with a copper block on one side and a heater core from an RV on the other. It was so efficient, it didn't even need a fan so long as the radiator fins remained somewhat vertical. I recall it used a pink additive to distilled water. That was bank in the day when you had to make double sure the very large PVC tubing was kink free, clamped properly, and oh the joys of removing air from the system.
I remember one objection to Peltier coolers in that application is that could possibly cool the CPU below the dew point and cause water condensation, something a fan can't do.
A friend had a Peltier cooler on a late 90s/early 2000s CPU and had the condensation problem. The cooler (or maybe CPU, or both; I can't remember which) had an algae-like growth all over it.
But why? Cooling has been a solved problem for literally over a century. No Peltier cooler will ever even come close to normal refrigerant-based coolers in efficiency, both in cost and power consumed.
Peltier coolers can do precision better than a compressor. The also work at small sizes/small loads, where a compressor generally needs to be a certain minimum size to be feasible.
Do you need those things in a home refrigerator? I suspect not. But it might be handy for lab refrigerators.
You can actually get surprisingly small compressors now the last few year. There are a few compressor driven ice coolers available now that are likely worth the effort and price if you need a boost in cooling capacity or battery life over the cheaper peltier equivalents. And they are also used by some people for large bulky cosplay and costume outfits that are otherwise just a super insulated hotbox. Obviously they still have a minimum size limit, but I think many people would be surprised at how small you can get them now.
Not the poster but yes. Compressors use phase change which is fundamentally better way of transferring heat. Solids like peltiers will always leak more heat than liquid to gas phase change. Peltiers cannot get the same efficiency because of physics.
Have a look at any product marketing these days. Desktop motherboards are calling the fan curve settings AI cooling. Everything that a computer does now, even if it's done it for 30 years is now AI.
Seriously. When I see this kind of absurdity, I think of Humpty Dumpty in Through The Looking Glass:
> 'When _I_ use a word,' Humpty Dumpty said in rather a scornful tone, 'it means just what I choose it to mean--neither more nor less.'
> 'The question is,' said Alice, 'whether you CAN make words mean so many different things.'
> 'The question is,' said Humpty Dumpty, 'which is to be master-- that's all.'
> Alice was too much puzzled to say anything, so after a minute Humpty Dumpty began again. 'They've a temper, some of them-- particularly verbs, they're the proudest--adjectives you can do anything with, but not verbs--however, _I_ can manage the whole lot of them! Impenetrability! That's what _I_ say!'
> 'Would you tell me, please,' said Alice 'what that means?'
> 'Now you talk like a reasonable child,' said Humpty Dumpty, looking very much pleased. 'I meant by "impenetrability" that we've had enough of that subject, and it would be just as well if you'd mention what you mean to do next, as I suppose you don't mean to stop here all the rest of your life.'
> 'That's a great deal to make one word mean,' Alice said in a thoughtful tone.
being colorblind, specifically deuteranomaly, it worked, but i do not see the blue green, i see the weird off green i normally do and then basically brown. i would imagine complete red-green would actually impact it
A lot of people consider them gimmicks but the newer gen neck “air conditioners” do actually cool the air and exhaust the hot air so the peltier works. Would love to see it improved for more cooling capacity (and more airflow).
Not likely without other contributing technologies. These devices just move heat from one side to the other, at the cost of producing more heat. So you still have to dissipate even more heat on the hot end with a fan/heatsink, in addition to draining the battery running the Peltier.
My process control theory textbook has a chapter on neural networks and a lot of the language in control theory has an AI like tinge to it. I think this AI language is native to control theory so it might not be as overblown as it first sounds.
Huh, control theorists always try to rigorously prove the stability and performance of their algorithms. AI seems to be the opposite of that: just let the black box solve it and don't worry about any problems, we'll just add more training data if they happen!
Hating on Kubernetes in 2025 doesn't really dong the gong like it used to, if you want to control your workloads through a well integrated API and not be bound to a cloud vendor there aren't many realistic options.
Kubernetes is only hard because people make it hard and never bothered to understand the basics of their workload scheduler.
Kubernetes is NOT AI hype, it solves real problems for real people everywhere.
"Infrastructure projects" that are here to stay and only getting better: Linux, systemd, Postgres, Kubernetes etc...
I think a lot of the Kubernetes hate comes from non-technical people deciding to use it in cases where it either doesn't actually solve any problem, or the overhead is much greater than the benefit. There's a reason it originated from Google, and as much as it pains MBA factory graduates, the project or organizaton they manage is nowhere near Google scale.
I firmly believe control theory folks didn’t invent LLMs only because the idea of doing a big fit on everything sounds too much like a joke they were telling each other.
If you type ‘fuzzy logic’ in to google the autocomplete suggested search is ‘fuzzy logic rice cooker’. Control theory has been stealing ML terminology for a long time.
That's quite a lot of memory for a little PID controller, isn't it? But I guess these days, they already only cost pennies, so you wouldn't save much from going for even less memory?
It looks like joke today, but it's not a joke anymore.
Book can look at your search history youtube history, or reading history to have insight about your points of interest to create list of topics you like, then generate pages on the fly using LLM.
Yeah but I believe the internet rescued a lot of people from that existence, but the people running the internet now are actively working against that.
I wonder if there's a site that keeps track of these. It was crypto before, then ML, for a while a lot of job openings seasoned their text with IoT, etc.
So I'm not defending the use of the term AI in this context, however in this case it's part of the product name (for the most part, there's only a few other references). Does the product use anything that would be regarded as properly AI? No idea, but marketing definitely is going to want you to use the Full Product Name(TM) whenever possible.
Web 3 was blockchain nonsense "everywhere" and the buzz has died because the only people who want blockchains are speculators, money launderers, and niche boring users like nightly settlement of inter-bank accounts.
And it was sprinkled everywhere, like sesame seeds on burger, just like Web 2.0 in the 2010 and AI now :) . We just need a Web 3 blockchain Web 2.0 AI FinTech startup
Just a few years ago, the buzzwords were "germicidal", "immunity boosting", "anti-bacterial", etc. for every damn thing, especially in India, where marketing can get away with bizarre claims.
Is there anything that is proved to be "immunity boosting" apart from vaccines?
Well, and germs of course. Germs will cause an immune response, but that's probably not the kind of "boost" you are interested in.
As for AI, I guess it is sufficiently vague and also not a medical claim, so you can put it everywhere. But because it is just as vague for consumers, is it even worth something for marketing?
Depends on your definition of "boosting". Vitamin C is essential for plenty of immune functions and a deficiency has serious effects on immunity. Enormous doses of Vitamin C have been shown to be effective in various scenarios as well. https://pmc.ncbi.nlm.nih.gov/articles/PMC8239596/
While not wrong, that is very misleading - for the vast majority we already get plenty of vitamin C in our diet and we don't have the specific situations where enormous doses will be effective.
At least vitamin C is water soluble: your body will almost always just pee out any extra quick so it is very hard to overdose. Those who have a kidney problem should check with a doctor, for the rest of us massive doses just make our pee more expensive.
>for the vast majority we already get plenty of vitamin C
This is highly dependent on the population you're talking about. There are plenty of people around the world who don't have great nutrition and don't eat large amounts of fortified industrial food.
>for the rest of us massive doses just make our pee more expensive
The study I linked reported ridiculous doses of vitamin C "dramatically improved the clinical state and cardiovascular, pulmonary, hepatic and renal function" in ICU spesis patients. Like 200-400 grams of sodium ascorbate administered over 7 hours. The recommended daily nutritional consumption of vitamin C is around 80 milligrams.
There are other studies reporting other positive outcomes for large doses of vitamin C. This isn't a recommendation that you should take hundreds of grams every day yourself, but there is ample evidence in a clinical setting for "immune boosting" vitamin C in large doses. (and also criticisms of the daily value for being too low)
If your immune system was "boosted" and worked significantly better tomorrow than it does today, would there be large measurable outcomes? No, probably not. Unless you've got something wrong with you, day to day, it would be hard to detect a distinct improvement in an already working system. Getting a good signal-to-noise ratio for a healthy person getting slightly healthier is hard. The bias for not publishing negative results and various sorts of intentional and unintentional p-hacking means that small-effect-size research that showed say "1 gram of vitamin C per day improved x by 5% in healthy adults" wouldn't be that reliable anyway. "dramatic improvement" of ICU sepsis patients isn't something that suffers from p-hacking.
Someone dying of sepsis in an ICU though, there's a whole bunch of very clear signals whether something is meaningfully effective or not. What that result tells you is that megadoses of vitamin C are likely not bullshit and should continued to be studied in more and more difficult to detect cases.
What this tells me is that 1) because I know it's likely rather safe for me and 2) there's enough evidence to convince me it might help, I'll keep taking lots of vitamin C if I feel like I'm getting sick.
Mandatory Technology Connections video about Peltier cooling, "Why Thermoelectric Cooling Is Inefficient":
https://www.youtube.com/watch?v=CnMRePtHMZY
This video assumes that this technology will remain static over time.
Peltier cooling could have a higher utility local maximum than currently used refrigerants.
The video did not make predictions about the future, it just presented the current reality at the time. I think it is still a good introduction to set the right expectations when encountering this technology.
Maybe. However there is fundamental physics in play and so it is likely someone (not me!) can tell you the most efficiency we get from that system. I'd be curious what those people say.
Which is to say, this video isn't informative about the possible advantages.
Could you summarize the contents of this video so we don't have to watch it?
I haven't watched it recently, but here are the main takeaways I remember:
Peltier coolers are neat because they're very small and quiet - as opposed to vapor compression systems solutions. However, they are an order of magnitude less energy efficient.
Also Peltier coolers still have to obey the laws of thermodynamics, which means that to cool one side of the mechanism, you must heat the other side. In order to do any substantial cooling, you need a way to dispose of that heat on the other side. This usually involves the use of radiators and fans, which negate much of the size and noise benefits.
As a result, Peltier coolers are pretty niché. Your use case would have to require only a little bit of cooling. You'd have to need a form factor that cannot accomidate a vapor cooling solution. And you'd have to be willing to make the system very energy inefficient.
AFAIK, no one has tried to build a Peltier cell paired with a heat pump. I am not an expert, but I would imagine that it's a path that could bring higher efficiencies. Thoughts?
It seems to me that if you have a heat pump, you don't need the Peltier anymore.
The heat pump is already so efficient that, I assume, adding something less efficient just makes it worse.
Also not an expert, but I’m struggling to find a combination where one of those couldn’t be replaced with a passive thermoconductive element. It’s hard to beat the efficiency of “free”.
It would be more efficient to just make the heat pump slightly more capable of cooling to reach the same total performance
It's one of those cases where you can chase a very small relative gain by adding a lot of complexity.
Like horsepower.
Heat pumps have a COP of 3-4, add in an evaporative cooling tower for a COP of 7. Peltier coolers will always have a COP of less than 1 (.1-.5?)
Unless you want to spend more energy that you remove in heat, stick with heat pumps and cooling towers.
Thank you; finally some data about actual potential.
COP = Coefficient Of Performance, BTW. (Heat out) / (power in).
> Could you summarize the contents of this video so we don't have to watch it?
Thermoelectric cooling is not very good and takes a lot of energy to do.
They are very inefficient
I think there are some applications though. I remember PWMing a peltier element to cool something to more or less exactly 35°C. I didn't need to be efficient cooling, it just needed to be reliable under space constraints.
I am no hardware guy and I remember there was a giant heatsink despite the constraints. It was some kind of photosensor + lamp if I remember correctly.
I think there was also some software logic to reduce water condensing at something too cool compared to the ambient temperature.
Obviously there are applications. There's one on my CPU.
The question is: Are there large-scale applications?
> Could you summarize the contents of this video
Summary (courtesy Notebooklm):
Why Peltier Elements Are Inefficient
• Fundamental Design and Coefficient of Performance (COP): Peltier elements are technically heat pumps. However, their efficiency, measured by the Coefficient of Performance (COP), is inherently low, ranging from "somewhere between zero and very bad" in practical applications. In theory, with perfect conditions and very low current, a COP between 1 and 2 might be achieved, but this generates hardly any cooling. In contrast, vapor-compression heat pumps used in traditional refrigerators can achieve a COP of 3 or more, meaning they can move significantly more heat energy than they consume in power.
• Impact of Temperature Difference: The efficiency of a Peltier element varies wildly depending on the temperature difference it is working against. The materials used to construct a Peltier element conduct heat even when not running. The greater the temperature difference between the hot and cold sides, the more heat energy leaks back through the element itself, dramatically reducing its efficiency.
• Internal Heat Generation from Current: The more current you attempt to push through a Peltier element, the more heat it generates within itself, which further diminishes its cooling performance.
• Comparison to Traditional Refrigeration: A small personal refrigerator using a Peltier element was found to consume about 55 watts of power constantly. This is significantly more energy than a standard mini-fridge, which might average only 21.7 watts to keep items cold over three hours because its compressor doesn't run all the time due to a thermostat. Even a much larger refrigerator, holding over 50 times more items, uses slightly less energy annually than the Peltier-based toy fridge.
How much of that is because of fundamental limitations, and how much just the best we can do with current technology but more effort could solve?
"In theory, with perfect conditions and very low current, a COP between 1 and 2 might be achieved".
For those worried about tiny COPs from these gizmos, trawling through the actual paper -- as well as the PR from JHU APL -- in this HN post [1] shows claims of COPs of ~15 for Delta Ts of 1.3°C.
A compressor based cooler gets a COP of about 4 in the real world. I'm pretty sure this is an apples to oranges comparison to an expert (I am not one of those) but a factor of 3+ increase in COP is fairly noteworthy -- if it holds up.
[1] https://news.ycombinator.com/item?id=44424087
ugh, reading the paper, their methodology is kinda crap. They basically just guess what the thermal resistances in their system are, and use air temperature measurements to figure out the heat flow. This is not how to measure heat flow accurately. It might be OK as a comparison with whatever TEC they tested with, maybe, but it's not at all something I would trust to compare to another test setup. If their box is more insulative than they think it is, their results are gonna look better than reality. This can be validated at least approximately by just putting a heater in the box that's dissipating a known amount of heat and looking at the temperature rise, but it seems they didn't even do this. And in general the regime where you've got small temperature differences is where your systematic error in a system like this can become huge and distort the results by multiples.
(This is an area which is really hard and details matter. Heat is basically impossible to measure directly, and the indirect measurements are fraught with peril. Getting it wrong was a large part of why people thought they had demonstrated cold fusion)
For anyone wondering (like me) what COP is in this context, it’s Coefficient of Performance: https://en.m.wikipedia.org/wiki/Coefficient_of_performance
and COP 15 is measure at 1.3'C temperature difference towards outside of fridge, so if it is 80f in your home then good luck to your lettuce.
Why, what would happen to my lettuce if I set the thermostat near the cooling element to e.g. 3°C and have the cabbage be somewhere between that and 4.3°C?
Unless I'm very mistaken, this 1.5 degree delta is between the hot side (the condenser at almost your room temp) and the cold side (evaporator inside the fridge).
That is, if you want to chill your cabbage to 4.3 degrees C, your room temp should be not more than 5.6 degree C. Or... if it's 25 degrees outside, in an ideal case your cabbage is 23.7 degrees cool.
Btw, max theoretical COP increases rapidly as T'hot - T'cold shrinks to zero. Your practical AC or fridge delta T is much more than that, depending on the weather.
The delta of 1.3C is critical there - peltier cooling drops precipitously in efficiency as the delta increases, and struggles to hit a COP of even 1 in real world scenarios. Their figure works out at about 6.5% Carnot efficiency, whereas a normal heat pump is usually nearer 45% over a much broader range of temperatures, as you can separate the hot and cold sides completely. Not so with a peltier wafer.
What they’ve done here is add a point of failure, use additional materials as well as a traditional heat pump, and called it “AI” and “eco friendly”.
Never have I seen more prime VC bait.
COP has to be compared at the same delta T. The COP of a loss less refrigerator is infinite at Delta T = 0
And you dont get to stack Peltiers to increase COP, only to increase delta T.
Still, Peltiers are super cool and I have some ideas for their use od they get slightly better. Advances are super welcome.
They're already doing great. I have a portable neck AC unit that uses peltiers and it keeps my head and neck cool in otherwise nasty desert conditions when mining. Yes the radiative heating from the sun is still a bitch but a hat basically minimizes that, and also redirects the partially-chilled air more efficiently around my head and face.
The direct-contact neck cooling plates are an absolute lifesaver. Keep the sun off the back of your neck and chill one of the best heat sink locations exposed on your clothed body.
Sounds cool :-) I think we'd forgive you for providing a direct link to this product.
There's a bajillion of them around if you search "neck cooler" or similar. Very simple product sticking a few commodity items together. Some do only have fans though.
When I looked a while ago there wasn't really a clear winner or high quality unit. There is the "Coolify" series that are much more expensive but still somewhat middling reviews overall.
No link, I have the Ranvoo Aice Lite. Couple hundred bucks on Amazon.
The idea was just so astonishing that I ordered some from American Science and Surplus. I connected the leads to a battery and poof, one side got hot and the other got cold. Blew my mind.
I didn't actually have a use for it. It was just neat that it actually worked.
I understand the basic physics of it perfectly well. It's just one of those things where you expect basic physics to be overwhelmed by friction or something.
Had a similar experience myself.
The thinness of Peltier devices bothers me. Amazing temperature differences… without any room to insulate such!
I don't know how people can convince themselves that they can understand these effects.
Unusually effective heat insulation is exactly how they work. (This is tempered by eg radiative losses, so making them thicker doesn't work better.) placing poorer heat (vs electrical) insulators between peltier material is counterproductive, similar to using resistors to improve conduction between copper wires
Don't ask me for a better explanation :)
As to why COP or even Carnot efficiency hasn't been thrown out in favor of temp-difference independent efficiency metrics like exergy.
I can't tell you either
Indeed, but in practical use (like a mini refrigerator), one will want to insulate the hot and cold sides from each other.
If one makes the walls thick, then they end up with a hole for the Peltier device and somehow sandwich two heatsinks on the device while maintaining insulation around it.
Perhaps easier to deal with in CPU cooling and such since one side is simply smacked into the thing being cooled.
What is exergy? The one time a mechanical engineering colleagues tried to explain it to me, he reached the wrong conclusion on the problem we were working on.
I haven't seen it in any physics thermodynamics book, and only mech eng. seem to know what it is, and then only in the US.
Faires (undergrad MIT Thermo book from the 50s) makes no mention of it as far as I can tell.
@tlocke
But that isn't a mathematical expression. At best, it would appear to be energy * maximum_Carnot_efficiency (for heat engines anyway)
But it seems not to be adding very much, since Carnot efficiency depends on delta_(T). The OPs point that exergy doesn't depend on T is tautological since T has already been accounted for by the Carnot expression.
Exergy is the amount of work that a system can potentially do. It was on my physics course in the UK I'm fairly sure.
Careful! Wothout a heatsink they'll easily self destruct (solder on the PN junctions)
> Still, Peltiers are super cool
I'd say they are super hot, but it depends from which side you look at it.
One can only hope some kind of phonon diode material can exist that a slight voltage can overcome something so inescapable as entropy by providing it only lanes that suit us.
I believe the “apples to oranges” is the temperature gradient. AC units would routinely manage 15-20c and are rated for more than that. And some freezers manage up to 50c. The greater the gradient the worse the efficiency in general.
What about a ΔT of say 20°C? I'd reckon most refrigerators and air conditioners are around there (temp difference of refrigerant between evaporator and condenser).
Stacking a bunch of these Peltiers to give more temperature difference would give a pretty low CoP. Say, for a 13°C temperature difference you'd have to stack 10 of them and use 10x the power. It's even worse actually as the hotter ones have to also pump the waste heat from the cooler ones.
Note that a small temperature difference that is sustained very consistently over a long time using a tiny amount of electricity (let's say half of what the parent post cited, so like a COP of 8) could add up to a lot of nearly-free cooling. You'd chill your walls for weeks and when a heat wave comes with hot nights for a week, if you(r home automation) close(s) the blinds during the heat of the day, the more-powerful AC might barely have to do anything
Just an idea of course, but I'd not write new tech off as "ok but just 1.3 degrees who cares" when the claimed COP is so insanely good without first trying it out
Chill for weeks? What kind of thermal mass are you thinking of there?
A brick wall is on the heavy side and you'd be able to store a whopping 1/8 of a ton of AC in ten feet of brick wall. It'd save about 0.1kWh of AC power use. A whole house will have many lengths of wall but also multiple tons of cooling requirements, so that doesn't help much. And how are you going to distribute those small temperature differences without wasting a ton of power?
And that's after you figure out how to trap a temperature difference in your walls for more than a day.
The ΔT between evaporator (typically 2-6°C while cooling) and condenser (often 40-50°C in cooling mode) is much higher than 20°C. The condenser is often almost 20°C above ambient outside temperature.
The design ΔT of ~10°C is the typical return-to-supply air ΔT.
>in this HN post [1] shows claims of COPs of ~15 for Delta Ts of 1.3°C.
>A compressor based cooler gets a COP of about 4 in the real world.
Real life refrigeration usually isn't very interested in a difference of 1.3 C. The Carnot COP for this temperature drop near ambient conditions is, I believe, around 200. When you consider a cooling technology relative to the Carnot efficiency (or COP) you get a better idea of what the efficiency means in practice. For an AC unit blowing 10 C air on a 40 C day, the Carnot COP is about 10, while real units get less than half that. But I think that's still better than the Peltier effect getting less than 10% performance relative to Carnot limits.
The COP generally varies with Delta-T, a compressor-based heat pump's isn't any different. An AC unit is more efficient when it is cooler outside.
That said, 1.5C is tiny.
Delta Ts of 1.3°C.
Might as well not use a refrigerator if your ambient temperature is that low.
You missed the "delta", meaning change.
One side would have been ~23C and the other 24.3C.
That's exactly what I mean --- if your ambient temperature is ~5C then you don't need a refrigerator if it only cools down by another 1.3C.
That's still a lot hotter than I'd like my refrigerator, to be fair to OP.
15 is absurd, regular peltiers get like 1.05 at most don't they? That's like inventing the warp drive but for cooling.
You can get a COP of ~3-4 out of regular TECs, but only at pretty low temperature differences. That's the killer, fundamentally the TEC material itself is thermally conductive and heat really wants to flow back the other way, so no matter how well it moves the heat, it winds up fighting against the heat load generated by itself. A refrigerant based heat pump works much better because the heat basically only moves in the direction the refrigerant itself is moving.
What's a "low temperature difference"?
I want to build a wine cooler in my basement ~20-24c, and I want it at ~16c. Is that low enough to be reasonably efficient?
TECs are wonderful little devices with operating characteristics unlike comparable devices.
They can be designed to move a specific amount of heat or to cool at some delta-T below the hot side (and due to inefficiencies the hot side can climb above ambient temperatures too, raising the “cold side” above ambient!)
I ran through a design exercise with a high quality TEC and at 8°C delta-T for a wine cooler you could expect a COP of around 3.5–4 (theoretically). This is pretty good! But below the 2.5V max to do that you’re only able to exhaust up to around 40W. For a wine cooler this is not so bad. For a refrigerator it’s a harder challenge because the temperature drops when the door opens, and if someone sticks in a pot of hot soup, it’s important to eject that heat before it raises the temperature inside to levels where food safety becomes a problem. For a CPU it’s basically untenable under load because it’s too much heat entering the cold side thus temperatures will rise.
https://fluffyandflakey.blog/2019/08/29/cooling-a-cpu-with-t...
Things often overlooked:
- Most TECs are cheap and small and come without data sheets, so people tend to become disillusioned after running them too hot.
- You have to keep the hot side cool or else the delta-T doesn’t help you. For a wine cooler this is probably no big deal: you can add a sizable fan and heat sink. For CPU cooling it becomes a tighter problem. You basically can’t win by mounting on the CPU; they are best at mediating two independent water-cooling loops.
- Q ratings are useless without performance graphs. It’s meaningless to talk about a “100W” TEC other than to estimate that it has a higher capacity than a “20W” TEC.
- Ratings and data sheets are hypothetical best cases. Reality constrains the efficiency through a thousand cuts.
When I think about TECs I think more about heat transfer than temperature drops. If you open a well-insulated wine cooler once a week then once it cools it will only need to maintain its temperature, and that requires very little heat movement. Since nothing inside is generating heat you basically have zero watts as a first-order approximation. For the same device mentioned above, it stops working below 1V, and at 8° delta-T that’s a drop in COP to around zero but it’s also nearly zero waste. If you were to maintain a constant 2.5V, however, it would continue to try and pull 40W to the hot side. This would cause the internal temperature to drop and your COP would decrease even though the TEC is using constant power. The delta-T would in fact increase until the inefficiencies match the heat transfer and everything stabilizes. In this case that’s around a 20° drop from the hot side, assuming perfect insulation.
Unlike compressors, TECs have this convenient ability to scale up and down and maintain consistent temperatures; they just can’t respond quickly and dump a ton of heat in the same way.
edit: formatting of list
This is great, thank you! Now I’ve got a new project to fiddle with :)
Physics dictates that the energy required to cool an object by 1C increases linearly, with difference to ambient temps.
It's like trying to swim against the current - the faster the river flows, the harder it is to move forward at the same rate.
lol yeah they did not. it's right on the abstract.
> is ~15 for temperature differentials of 1.3 °C.
COP of peltier elements can be large only when the temperature difference is small, such as the measly 1.3 degrees you quoted. When do you want to cool something by only 1.3°C compared to the surrounding temperature?
AI hi guys, pretty great AI article about this AI Peltier AI cooling AI tech. Really looking AI forward for another AI Samsung new AI devices.
Best regards,
AI
Ps.: AI
I heard it was trained on stolen refrigerators that were for research purposes only!
Hello! I am a big respectable VC firm. Here is a blank check and a unicorn hat, congratulations on your new AI startup.
It's "if" statements.
Maybe a "switch" for when things get advanced.
It's that kind of AI.
I take three!!
AI comment
> AI comment
AI rebuttal about it not being AI
The Johns Hopkins press release is much better
https://www.jhuapl.edu/news/news-releases/250521-apl-thermoe...
Definitely worse article
I don't understand how they could quote him saying: “This thin-film technology has the potential to grow from powering small-scale refrigeration systems to supporting large building HVAC applications, similar to the way lithium-ion batteries have been scaled to power devices as small as mobile phones and as large as electric vehicles,”
Then the entire article foregoes comparing their peltier device to traditional compressor based heat pumps.
Oh nifty. These can also be used to generate electricity from body heat. Or presumably any waste heat.
Only if you have a cold environment into which to dump that heat into. You could maybe trickle charge your phone in the winter, if you don't mind a cold spot where the device sits on your body.
Assuming body is 38 degrees C, and ambient is 20 C, Delta T is 18 C, which "works" to whatever extent.
can this be applied to create devices that maintain a temperature range? ie more than x AND less than y? that’s often needed for medications…
I wish they had gone into detail about what makes this device different from Peltier coolers that are available today. You can already get mini fridges that use them, but they're not very good - the total amount of cooling you can get from them is not enough to maintain the temperatures that a regular refrigerator does.
Yep, it's hard to tell if this is a true leap or just marketing polish on incremental improvements
Yeah I have a cooler box with a peltier cooler, but it will only cool a few degrees, not sure I'd trust it as a standalone fridge. Plus its energy use is much higher.
The most impressive thing about this article is that they somehow managed to shoehorn AI in a fridge.
Reminiscent of the Long Island Blockchain Company (makers of Long Island Iced Tea (TM)).
Did they solve the inefficiency problem of peltiers?
They claim 75% better. I doubt this is better than compressors in general but they seem to be going for a niche where a high power compressor does the high load part but when only a litte cooling is needed the compressor is now very inefficient and so the peltier is better. Normal fridges just let the temperature have a wider swing which is good enough for most needs.
> Normal fridges just let the temperature have a wider swing which is good enough for most needs.
These wide swings annoy me. You hear that you shouldn't let your fridge go above 4°C, because that's dangerous. And you obviously don't want your fridge to go below 0°C. But finding a setting where the hottest part of the fridge doesn't go above 4°C (or even 5°C or 6°C) during the hottest part of the cycle and the coldest part of the fridge doesn't go near 0°C during the coldest part of the cycle is really pretty difficult, in my experience.
Put a thermometer in a plastic bag (to keep it dry) in a jar of water in your fridge, you'll likely find the temperature of items inside the fridge much more stable than the air temperature.
The yogurts at the back of my fridge still freeze up.
There's going to be a temperature gradient in a typical fridge, if only because one side is getting opened every now and then, and the other is separated by all the products which are inside.
I suppose what they're trying to do here is even out that gradient without running the compressor.
The time it stays above 4C unless you're opening the door of your refrigerator all the fucking time is mostly negligible in terms of the difference it will make for bacterial growth.
The foodstuff itself is loaded with water, it won't have an excursion dangerously above 4C just because you opened the door and the air temperature raised a few degrees.
If you are really worried about it (you shouldn't), and you don't keep your refrigerator full, add a few water bottles for thermal mass.
Why should that be dangerous? I have never heard that.
I have always had my fridge at 8°C and never had something dangerous happen to me. I have never come across fridges that were way cooler, apart from fridges of friends in Canada and the US. What's the reasoning?
The recommendation I've always heard in the Netherlands is 7°C, it's more recent that I've been seeing 4°C on meat packaging in Germany (where I live now). I doubt anyone's fridge is consistently at or below 4° without freezing things constantly, so I've been assuming this is wishful thinking and/or ass-covering on the manufacturer's part and not what anybody actually does. Your 8°C is close enough that it probably makes little difference, though afaik this is an exponential curve (at 14°C it would last far less than half as long) so I'd not be surprised if things spoil a bit sooner than they otherwise would
Even if your products generally meet their "should be safe at least until" date (Mindesthaltbarkeitsdatum, idk if it's the same as "best before"), you might exceed that longevity more often than you do now and thus have less food waste by setting the fridge colder - if food waste is a thing you have in the first place (I'm the type of person that is hungry all the time, opens the fridge when hungry, and isn't super selective (among what I've bought anyway), so food I buy ~always gets eaten before it spoils, but then when I see food waste numbers, apparently that's not the case for everyone so I'm just throwing this out there)
Edit: trying to fact-check myself, I can't find any trustworthy source in Dutch saying your foodstuffs fridge should be more than 4°C. I measure new fridges when moving in and again at least once during the first summer to make sure they stay at or below 7°C when we had the door open a normal amount of times, so I know they're that (and not much cooler, to not freeze items or waste energy). So far, products meet their minimum shelf life date thingy and almost always exceed it. Strange. Maybe this recommendation I heard predates the internet (showing my age here), or maybe every page on the internet assumes that nobody actually measures it properly and so they recommend a value that's half of what's actually safe?
Thanks for that explanation!
The Mindesthaltbarkeitsdatum is the deadline to which the item may still be sold by the vendor, iirc.
> I doubt anyone's fridge is consistently at or below 4° without freezing things constantly.
My refrigerator is typically between 3-4°C, never had any problems with things freezing.
Crazy, I'd think the fluctuations would be way too big. How long does it normally run for when it does a cooling cycle? Is that temperature throughout the fridge or only on certain shelves?
Ours (well, our landlord's) runs very consistently for 40 +/- 3 minutes, with just over 1h30min in between during normal use, or 1h54min if it was closed the whole time (like at night). The temperature of products, as measured with a cheap infrared thermometer that's probably off like 10%, varies between 2 and 7 degrees, but it's not very consistent between shelves (top seems warmer but then the very bottom one, that is only half deep, is as warm as the top again). The products I checked have all been in there for days; reflectivity may be part of the difference, not sure. I don't know what the air temperature difference is at the beginning and end of a cooling cycle though
[dead]
You should buy a fridge with a fan in the cooling chamber. My Samsung fridge has a nice fan and associated ducting to circulate air and keep temperatures ~uniform.
I may consider that the next time I buy a fridge, but the reality is that this isn't something I care deeply enough about to buy a new fridge over.
Maybe you can retro a fan into your current fridge. Nice little project!
I think the 4C recommendation is an average. There is not magic in ~1% temperature difference from 3 to 6. Just slightly different rates. Which again slow, but do not stop when it is at low end compared to high.
> you shouldn't let your fridge go above 4°C, because that's dangerous.
Dangerous how?
If you buy something good like a Sub-Zero it will stay within 1 degree of the set point.
> You hear that you shouldn't let your fridge go above 4°C
Really? My fridge says 8°C, I think?
Between 3°C and 5°C is the recommended refrigeration temperature
https://www.which.co.uk/reviews/fridges/article/how-to-store...
Christ.
I also run my fridge at 8C, which I think was the default setting when I bought it.
Gonna go change that right now.
8 degrees or just 8 on the 1-10 scale that many of them use?
(I always remember the recommended fridge temperature as 40F, which avoids the confusion.)
I suspect it's that. Most inexpensive refrigerators don't have thermostats. Which seems insane; it cannot add that much to the price.
8C is a perfectly fine temperature for a wine fridge. And they usually have thermostats because a wine fridge is a luxury item. As opposed to keeping people from getting foodborne illness.
> Which seems insane; it cannot add that much to the price.
Be careful what you wish for.
When buying my current fridge, I specifically tried to go out of my way to avoid complexity, but the opening for my fridge is an odd size so my choices were limited. The only fridge I could buy that didn’t have a bunch of crap that’s was guaranteed to break (ice maker, water dispenser [seriously? aren’t most fridges right next to a faucet?], LCD-covered glass panel, etc) that also had a thermostat had a digitally-controlled thermostat. No knob or physical buttons, just a capacitive surface for temperature adjustment and some LCD screens showing the fridge and freezer set temps. (Not the actual measured temps, that would be too useful, just the set temps.)
In hindsight, I probably should’ve just gotten one with a regular dial, but I was a bit fixated on the “real” thermostat. So now I’ve got that to look forward to breaking in 4-5 years and figuring out where the hell to source a discontinued fridge LCD panel from.
> water dispenser [seriously? aren’t most fridges right next to a faucet?
Same reason you'd keep a container of filtered water in your fridge.
...why would I keep a container of filtered water in my fridge? My water comes out of the tap already filtered at my municipal water supplier?
Maybe if I was on a well I would need to filter my own water, but then I definitely wouldn’t trust that job to a fridge.
Many people prefer a cold glass of water compared to the cool to room-temperature water that comes out of the tap (Hence the refrigerator for cooling). Filtering at the individual level is typically done for flavor, because depending on where you live, while the water is probably safe to drink (assuming HN readership demographics), it may or may not be particularly pleasant. Even here in the Bay Area, where we have that sweet, sweet Hetch-Hetchy water, I know people who live in buildings with pipes bad enough that they filter their water.
75% better than "utterly awful" is an improvement, but you're still going to be searching for applicable niches with a microscope.
There are many applications for peltier cooling in industrial settings, i.e. sensor devices (or, actually, microscopes with CMOS sensors). Those have to be temperature stabilized to minimize noise while also be IP rated, i.e. no air cooling. 75% efficiency gains would be awesome here.
What exactly does "75% better" mean? 75% more cooling per watt? 75% less distance from theoretical ideal?
Would a peltier element allow for a more constant temperature? Or can you turn a peltier up and down easily? With a compressor based system it's always been "on or off", something that can ease up and down would be nice.
Yet I'’d love to see real-world power consumption numbers compared to traditional fridges
Traditional Peltier devices operate at ~10% efficiency (COP of 0.5-0.7) compared to vapor-compression systems (COP of 2-4), but recent advances in thermoelectric materials like bismuth telluride alloys and segmented elements have pushed lab efficiencies to ~15-20%.
It's clearer to think in terms of "efficiency relative to ideal Carnot efficiency".
Compressor systems use twice as much energy as an ideal system, while Peltier systems use about 10x as much.
The article says they have made peltiers 75% more efficient than existing ones.
OK, a typical Peltier device has 3.5% coefficient of performance, that is, it produces 35 W of cooling per 1 kW consumed.
Fine, let's expect that the new tech doubles the efficiency, to 7%. Still, to my mind, pretty wasteful, on par with a steam railway engine. A Peltier element is good in cases where you can afford a large heat removal device, but need precise temperature control and no moving parts. For a home fridge, I'll take the sound of the compressor and the temperature fluctuations of a 400% efficient compressor-based heat pump over a Peltier element any day.
While Peltier cooling have low efficiency wouldn't it be ideal in some cases like:
- energy source is solar, DC already and abundant.
- cold climate so the fridge can contribute to heating the room
Anyway good to know those small electric cooler with Peltier effect must be consuming a lot electricity.
A common use case: coolers. You don't want a whole compressor, but you do want to keep your hot dogs safe and your beer drinkable. It's not very efficient but it's enough for short periods.
There are mini-compressor based coolers available now if you look for them. They cost a little bit more, and obviously have more weight than a peltier, but I think are worth it if a bag of ice isn't viable by itself because they will run for way longer than a peltier setup.
I considered putting some in our under-sink reverse osmosis tank, to cool the water. Couldn’t come up with a way to exhaust the heat well enough to even look at how much electricity it would have cost me though - probably too much to make it worth it.
"that is, it produces 35 W of cooling per 1 kW consumed."
No, they don't. First, without defining a delta T, efficiency is meaningless (unless its a Carnot cycle).
Second, the efficiency is (depending on op. point) higher than 100%. See [1]. You can pump 20 W of thermal power with 2 A @ 4 V = 8 W
20 W of cooling for 8 W of work, or an efficiency of > 200%. This is common to all refrigeration cycles, and frankly for a puny 10C, it sucks.
[1] https://www.datasheethub.com/wp-content/uploads/2022/10/data...
Oh, indeed I was wrong; COP of Peltier elements is much higher than I had gleaned from online charts, not the order of 0.03 (3%) but can easily reach values above 1.0 (100%) and be e.g. 0.5 (50%) at ΔT = 30C, enough for a home fridge.
Still a bit far from compressor-based designs, but not negligible, and almost doubling the efficiency is indeed a serious advance.
That is wrong, COP is expressed as a ratio, not a percentage (efficiency is expressed as a percentage, which is COP * 100). And as others have said, both efficiency and COP are dependent on ΔT both in refrigeration and thermoelectric cooling.
A steam railway engine is a lot better than you would think. They were more efficient than diesel engines when diesel took over - the diesel engine needs much less human labor and so was cheaper overall, but for efficiency steam was better. (Note that diesel technology has improved since the 1950s, so I don't know how they compare now)
> In the Bespoke AI Hybrid Refrigerator Samsung launched in 2024, the compressor operates under normal conditions such as routine storage and retrieval, while the Peltier device activates alongside the compressor during high-load situations — like when storing large amounts of groceries or placing hot food inside — thereby enhancing both cooling performance and energy efficiency
I'd choose a fridge with a larger compressor.
Larger compressor would probably mean short cycling though. That has its own downsides.
What I want is a silent refrigerator, will this bring that? pray
I've been thinking about what it would look like to convert my GE fridge into a minisplit (i.e., move the compressor, condensing coil & fan outside).
You can buy R600a on Amazon right now. One $60 can will charge the system ~5 times.
With home HVAC, fridges, water heaters, and dryers all using now able to use of dependent on heat pumps I wonder how long it be before we see modular appliances that connect to coolant lines where the temperature differential is supplied by a central high efficiency heat pump.
Cars already have heat scavenging that can move heat from where it's being created through losses to places where it's valuable, like the cabin or battery pre-heating. Especially in cold climates it feels like homes should be next.
There's some commercial options for this, but it's not common. Usually, these devices just have their own compressors, because they all pale in comparison to the heat pump(s) used for climate control. For example, I have a HP water heater, and its heat pump is about 1/3 of a ton, whereas most homes need 3+ tons for climate control. Fridges are a fraction of that.
For HP clothes dryers, there's no efficiency to steal from somewhere else, because they use both the hot and cold coils - similar to (the same, really) dehumidifiers.
The tradeoff would also be running high-pressure refrigerant lines everywhere. That would require EPA certification (in the US, anyway) to connect/disconnect an appliance, and it would probably be less reliable. These sealed-system units are generally pretty reliable, because the refrigerant is installed at the factory under ideal conditions, and there's no connections that are made later that may be done poorly.
That is an interesting thought, but I assume that the working ranges of the different appliances are different so there would be some complexities and inefficiencies getting them all connected to a common circulation loop. If there was a thermal equivalent of a transformer used for alternating current, that would be amazing.
As far as I know all the common commercially available heat pump appliances all use the same refrigerants, so it doesn't seem like it would be that challenging.
In cars that have unified heat management the refrigerant cycle is handled as a separate element, with a manifold controlling individual coolant loops to each component. I'm picturing something similar for the home, with a coolant moving heat to and from each appliance using standardized communication to the manifold. There would probably need to be heat buffer tanks, but air to water heat pump systems for radiant heat already need this anyway.
A few years ago I was planning to build a velomobile that I would live out of for a year while circumnavigating Australia, and potentially indefinitely. (My plans changed.) I was disappointed at how hard refrigeration information was to come by (maybe I should have sought a traditional paper book), but I was kinda looking forward to figuring out if I could use one compressor to cool a small fridge, cool and perhaps heat the cabin, heat water, and heat a slow cooker (target 80°C). The bits I could work out suggested you might want a different refrigerant for the cooling and heating applications, or different back pressures; but I was rather hampered in my reckoning by my lack of domain knowledge—I was definitely going to have to talk to professionals! If so, and combined with the limited power collection available (<1m² usable solar panels on the vehicle, could lay out more while parked), butane was probably going to make more sense for cooking.
I even ran some naive numbers on the amount of water that would condense in expected conditions, concluding it could be handy but I’d probably still need to source more water.
It's worth noting that the very earliest electric refrigerators had a separate condensing unit outside; see this interesting 1920s Frigidaire training video for an example of what that was like: https://www.youtube.com/watch?v=W-t7DqOAMME
There were also centralised systems for apartments where one condensing unit supplied many evaporators in the refrigerator in each suite.
https://patents.google.com/patent/US20120192586A1/en
(abandoned)
That would've been easily invalidated by prior art from nearly a century ago; as I mentioned in a sibling comment, this was a common arrangement in the early days of domestic refrigeration.
Here's a video from someone who managed to salvage some of the components of such a system: https://www.youtube.com/watch?v=h1tXIYl20jA
Absorption refrigerators have been around for approximately a century, are silent, and a little more efficient than peltiers.
I used one for a couple years as my primary fridge. It was expensive, like $2k, didn't have very good temperature control and broke after 2 years and couldn't be repaired.
So... A little better than Peltier cooling in every other dimension too.
You can hear your refrigerator???
The latest wave of appliances is really fucking loud for some reason.
I think they're using different kinds of motor windings, bearings, insulation, etc. it's not related to the refrigerant or other system parameters. I've had older r600a fridges that were dead silent compared to anything sitting in a Best Buy showroom right now.
Likely high speed compressors --- the oldest hermetic systems used an induction motor running at 1800 RPM, then later they went to 3600 RPM, and now they're running on a VFD that possibly goes much faster. By making it pump faster, they can use a smaller compressor and reduce costs, at the expense of longevity and noise.
Ew.
They use lighter lubricant oils than were historically used, which allows more vibration. It’s also why compressors burn out far more quickly than they used to.
That too. https://news.ycombinator.com/item?id=32462954
You can't? Refrigerators have always made a noticeable background noise as they cycle on and off.
The advantage of the newer variable speed scroll compressors in some high end fridges is that they can run continuously at a slower speed.
I'm not the same person, but I do not (and never have) noticed any noise when my fridge is running. Whether that means we have different fridges or different tolerances for noise, I'm not sure.
It's tolerance. I've bought several of the quietest fridges and they all bother me. Old fridges sound nicer but newer ones are actually quieter.
Is it the fridge noise in particular or all noises that you find you are sensitive to?
In practice the fridge is the only common thing. Most stuff in a house is quiet or, like fans or furnaces, make simple sounds that are easy to ignore.
More likely you have a noisier house/neighbourhood. I used to think my PC was silent until I moved to a quiet place and then I could suddenly hear it very clearly.
In an "open home" concept I guess it might make sense, but I've never lived in a place like that.
I guess all of the places I've lived the kitchen was always its own room, maybe adjacent to the dining room if anything.
No new appliances (>10y now I think about it, they came with the house.)
Yes, absolutely. In particular, I find this obnoxious when staying in hotel rooms that have a minifridge.
My wine fridge uses Peltier and is super quiet. It's the perfect application for this because wine doesn't need to be as cold as a normal fridge, and noise is a consideration.
It's not completely silent though, there's a small PC-like fan but it's way less loud than a compressor.
You might be able to swap that fan out for a higher quality one like a noctua.
oh thanks for the tip! I'll look into it.
They already exist! Well, small ones.
A hotel I was staying at had a small bar fridge that used a Peltier. I only know because it stopped working so I checked it and realized it was only a Peltier plus a heat exchanged (a cyclopropane loop).
I presume a full size fridge is outside of reach at this point.
Move it outside a cabinet, let it free stand. I found out that my nice kitchen niche for the refrigerator acted like a nice resonance chamber for the frequencies the compressor generated.
I can barely hear it now.
If the cabinet is poorly designed (or ventilation is otherwise obstructed), it will also retain heat, making the fridge have to work harder.
Simplest solution would be to have a compressor that is only active e.g. during the day (when the user is not at home).
But the user doesn't open the door when they aren't at home.
There will be greater fluctuations of temperature but that doesn't need to be a problem.
Peltier modules have no moving parts, so they're inherently silent
Why?
Not OP but it's a massive nuisance if you live in a studio. People don't realize how noisy a fridge is until there's one in the room that they sleep in.
New appliances are far better than old ones here. Especially old ones that (I assume) haven't been maintained and so are working far harder than they used to. I've lived in places with old ones that were fine and old ones that were awful, both. I've had much more consistently good results in places with newer ones.
Skill issue on the manufacturer's part. I live in a studio and never hear the fridge. This is part of a fitted kitchen, though, but I doubt the panel hiding the fridge makes that big of a difference.
Just today I ordered a 32dB Liebherr; the previous one had 35dB and could be heard all around the studio (I measured the noise using a dedicated sound meter).
Here is an article from Johns Hopkins with a little more information on the newly created material. https://www.jhuapl.edu/news/news-releases/250521-apl-thermoe...
I remember back in 2006ish, my PC modding days, buying some peltier modules from eBay and (attempting) to get subzero temps out of my CPU cooling.
If I recall correctly I got the setup powered but stopped short of actually putting it on my CPU when I couldn't mount it all in a way that would let me contain the condensation with what I had.
Maybe it was pccooling or pccasemods dot com? There was a really strong community forum back then where it was all going down, people were nitrogen cooling their PCs, watercooling was a big deal, and CPU temps of 60c were considered unsustainable.
I still overclock my computers but usually my aim is a silent computer under 60% load, so my goals have changed. Peltiers are not something I see taking over PC cooling even now. You still need the same radiator capacity, the peltier just moves the heat away faster and can get below ambient temps at the CPU.
Given that almost all CPUs these days are thermo-throttled, maybe a controllable Peltier element will be the next Turbo button? I remember that Peltier based cooling setups were electricity hungry.
And electrically noisy, at least in my experience. Not a property of the Peltier junction, just crappy PWM control afaict.
Bit-tech.net was a good one
Oh yeah! That might have been the one actually.
Killed an AMD K6 with my diy peltier cooler when I was 14 good times.
I thought Peltiers can't lower the temperature by more than 40 degrees F, in practice less than that. This is not cold enough for a refrigerator on a warm day.
FWIW, in a typical apartment or single-family home, refrigeration uses a fraction of the energy that space cooling (also via a refrigeration/vapor compression cycle) requires on a warm day (and probably year round too unless in very mild climates). The psychrometric chart path is different so there are of course differences in the amount of energy required for the sensible and latent components, but the real difference is just the volume of air that needs to be dealt with.
My point being that at least from an energy and carbon perspective, lowering the space cooling demand via more effective building envelopes or increasing the space cooling supply efficiency - eg via membrane or dessicant dehumidification, better heat pumps etc) is far more impactful on a macro scale than better refrigeration.
Granted refrigeration in a warehouse eg is really also space cooling, but I’m just making the distinction between the dT=0-25F context and the dT>25F context. If I could only choose one technology to arrive at scale to improve the efficiency, it would be for the former context.
Definitely not the volume of air. The thermal mass of air is tiny.
The difference is in the thermal mass of the building and the surface area exposed to the sun.
Also the area that needs insulating, and in the extremes the amount of air that needs to be exchanged with the outside to make the house livable, and the heat generated by the people living in it (stick a 100W lightbulb in a fridge and see how cold it can get).
The insulation is actually solvable, and for heating can basically remove the power requirements: a house heated and using heat exchange on air leaving vs entering can be heated a lot just by having people inside it, let alone the other energy they use for other purposes. It's just more expensive to build this way, and with cheap energy it can a long time to pay back. Cooling you can't push down past the heat generated inside the house divided by the COP of your cooler, though.
Yep, PassiveHouse standards which typically include an extremely tight envelope which forces installation of outdoor air supply famously can get away with just a few hundred watts of heating capacity because of heat exchange on the incoming and outgoing airstreams!
Sure I was playing a little fast and loose there, but (a) the large surface area of the home (and resulting conductive transfer through the walls + convection transfer via infiltration through gaps) is directly a result of the fact that you need a significantly larger volume for humans to move around in and live in than you do to store food and (b) even if we do look directly at the volume of air, the difference is significant since at the end of the day, since for any given constant deltaT, your energy spent is still linear with mass or volume. And we are talking about roughly 2-3 orders of magnitude difference in air volume between a house and a refrigerator.
Anyways, if you write out all of the heat balance equations, you get a few W/m2 of flux on the inside wall of the home and a few W/m2 of flux on the inside faces of the fridge, assuming a typical wood frame construction in summer time and steady states all around.
So yes, of course multiplying the flux through the home’s wall by the surface area of the home results in a massive heat gain value compared to the heat gain conducted through the surface of the refrigerator, but that’s arguably precisely because of the two different volume requirements.
That's a question of mass and power.
There are no inherent max temperature deltas for Peltiers, just coefficient of performance, or watts moved per watts wasted, is atrociously low compared to just about everything else.
Peltier coolers have non-zero thermal conductivity and are pretty thin. At some point, heat leakage from the hot side to the cool side catches up to the rate of heat removal the cooler is capable of. What that point is will vary between devices, so there's no single or fundamental limit.
You can also stack them, but you need ever-increasing areas of TEC and corresponding power consumption. I think a triple-stack is sometimes practical if you want to get something to well below ambient and it doesn't really generate much heat itself (and you don't really want to use a traditional cooler due to size or vibration or something similar)
Article author really wants those AI hits on search engines!
I worked on a scientific instrument a while ago, it had a Peltier heater on it to raise the sample to 60c from whatever residual environment temp was (approx. 20c in a lab). It was pretty amazing to see my old overclocking cooling solution come back around and be used in my professional career some 20+ years later
And they couldn’t use a 100% efficient resistive element because…?
The TEC is more than 100% 'efficient' - you get more heat out than power you put in. Though I would guess it's mainly there to get heat control that can span above, around, and slightly below ambient.
Because Peltiers are, at their most inefficient 100% efficient?
At a reasonable delta T you can get 200% efficiencies.
typically only used if you need to stabilize the temp very close to room temperature or cool it below, unusual choice otherwise compared to a heater.
Even above ambient, being able to reverse direction allows limiting any overshoot.
Yeah, it occurred to me after asking that that’s likely why - with resistive heating you’re bound to have a fiddly time getting a stable temperature.
If you are reasonably above ambient and the control loop is well tuned things work well, unless you are significantly perturbing the system frequently.
Honestly, I didn’t ever ask. As the SWE on the project I just had to switch a DIO on/off based on the reading from an AIO. I just trusted that the EE and ME folks were making smart decisions around BOM.
AI has literally nothing to do with this. Why do they feel the need to sprinkle the phrase everywhere? AI inverter/compressor? Come on, have some sense of shame, please.
It's the name of the game, alas. In the area I work in, I've seen many companies get many multiples of the investment we've gotten because it's "quantum", even though they're trying to do the same thing we're doing and doing less well at it.
I assume that if you have "quantum AI", VCs break into your house and force bags of money on you.
The head of the board of a company I work with, who uses ChatGPT to get information on how to prune his plants and barely knows how to use Windows Explorer, decided that the whole company needs to go full AI. Everything has to be AI, and everyone has to learn AI. I thought that HN has been overdoing it with AI articles recently, but parts of the outside world seem to be much, much worse.
Actually AI has a lot to do with material design. I can't do a proper explanation, so please bear with this information dump:
- We don't have a strong physical theory for solid state physics. Quantum stuff doesn't scale well from 1 atom to a mole of atoms, because 10^23 goes into the exponent of number of energy levels in the system, and then we also have to model interaction of those levels.
- Physical properties of materials depend on their crystal structure, unevenness of that structure, spectrum of size of crystals, temperature, pressure, fields they're exposed to and current position of stars in the sky. Even the "wrong" solid state physics equations we have are highly non-linear.
- State-of-the-art effects are usually achieved with combinations of such materials.
- Exact parameters of the process used to put those materials together radically change the behavior of the system. Put that nanolayer with a different sort of vapor deposition or at different pressure, and the thing will stop working. Ever wondered why we don't produce all the neodymium magnets at N55 grade? Because even precise description of the process is not enough.
- AI doesn't care about the exact physics, but is sometimes very good at navigating in the large parameter space.
- Google have recently made AI that predicted thousands of novel material structures. They found more materials than were found by all human research over the whole humankind history.
I wouldn't expect AI to explain what's going on in solid-state physics anytime soon, but exploring crystal structures, doping, and process parameters automatically might actually get us new materials a couple hundred years faster.
What the article is talking about is using temperature probes and deciding if to run the compressor or peltier, or both depending on the temperature. That's what they are calling AI. Some if statements on a microcontroller.
In the article, the picture shows the peltier cooler is on top and the compressor is separate on the bottom. The "AI" is an adjective for the compressor only. There is little to no discrete logic in a compressor, so calling it that is nonsense.
Oh wow. I've seen so many "AI" that apparently I started reading around them.
Maybe it’s hooked in to ChatGPT
“You are a refrigerator, examine this photo of the temperature reading and decide (y/n) if the compressor should turn on”
Because it works wonders on non-technical people
Sounds like they’re using much thinner and thus more efficient bonding on each end of the Peltier device?
75% gains off that seems impressive. Must be something really fancy - thinking of a heat sink just using better therm paste barely moves the needle
Much better British tech is the smearing of yogurt on windows https://fortune.com/2025/07/14/yogurt-on-windows-cool-home-b...
I'm really curious how it occurred to them to try this in the first place. Also I thought Brits spelled it "yoghurt"?
So PID controllers are AI now? Cool!
This still feels like early days. The fact that they're sticking with hybrid systems for now suggests we’re not quite at the “compressor-free” future yet. Also, the environmental angle is compelling, eliminating refrigerants would be huge
I used peltier coolers on K6 processors 1M computer years ago. They worked great and lasted about as long as the PCs did.
Not long after I bought mine, they disappeared from cooler offerings. I've wondered what became of the tech.
> I've wondered what became of the tech.
Remember that Peltier coolers don't make heat disappear - they just move it from one side of the cooler to the other, and produce a lot of additional waste heat in the process. There are better ways of transferring heat from a hot IC to a heat sink nowadays - like liquid cooling for really high-performance systems, or capillary-action heat pipes for more typical needs.
> There are better ways of transferring heat from a hot IC to a heat sink nowadays
Peltiers were always a bad way to move the heat. What they offered was the ability to go below ambient, which at the time could improve overclocking. Peltiers of course lacked the capacity to actually take you there with any decent load, but in theory it could.
It never really made much sense for consumers, and once consumers realized that, the market went away.
peltier on the liquid reservoir might be worth trying...
They are inefficient and dump a bunch of additional heat into the heatsink that you need to get rid of.
TDP went up. TECs are extremely inefficient, and have limits on how much heat they can pump in a given area anyway. a 10W chip, you can maybe get away with, and it might be a meaningful improvement compared to the other cooling tech available at the time, if you don't mind the power usage. with a 100W chip, you have no chance.
(You can see a demo here where LTT try it, and they, after dumping 500W into the cooler, can get the CPU to a vaguely reasonable temperature, until they actually load it up: https://www.youtube.com/watch?v=sWrqyQWfhrs)
Your K6 processor has a TDP somewhere around 20-30W depending on the model.
Modern performance CPUs have TDPs in the 100-200W range.
Peltier cooling generates a lot of additional heat. It doesn’t scale well to the higher loads. That’s why you don’t see them any more.
I'd consider performance processors something like a threadripper - 300W unlockable to 1000W. I haven't kept up with intel but I think they were indeed riding 200 to 300w.
They made both the problem of power consumption and removing waste heat infinitely worse. Instead, I used liquid cooling with a copper block on one side and a heater core from an RV on the other. It was so efficient, it didn't even need a fan so long as the radiator fins remained somewhat vertical. I recall it used a pink additive to distilled water. That was bank in the day when you had to make double sure the very large PVC tubing was kink free, clamped properly, and oh the joys of removing air from the system.
I remember one objection to Peltier coolers in that application is that could possibly cool the CPU below the dew point and cause water condensation, something a fan can't do.
A friend had a Peltier cooler on a late 90s/early 2000s CPU and had the condensation problem. The cooler (or maybe CPU, or both; I can't remember which) had an algae-like growth all over it.
For this we calculate the dew point with 2 sensors and react accordingly in my temperature controller.
I'd really like to try out these better peltiers, our current ones suck. And the fans to remove the heat are huge and loud.
There are other ways of building refrigerant-free refrigerators too
https://en.wikipedia.org/wiki/Magnetocaloric_effect
and
https://en.wikipedia.org/wiki/Elastocaloric_materials
of course they also have the dew point problem, but so do ordinary refrigerators and freezers.
I need active CPU and light cooling. Fridges are easy.
Anybody who points to dew point issues would be fascinated by
https://youtu.be/qAZ-q3KmDHM
They weren't worth it. Their only advantage is they can cool below room T, but then you have condensation problems.
But why? Cooling has been a solved problem for literally over a century. No Peltier cooler will ever even come close to normal refrigerant-based coolers in efficiency, both in cost and power consumed.
Peltier coolers can do precision better than a compressor. The also work at small sizes/small loads, where a compressor generally needs to be a certain minimum size to be feasible.
Do you need those things in a home refrigerator? I suspect not. But it might be handy for lab refrigerators.
You can actually get surprisingly small compressors now the last few year. There are a few compressor driven ice coolers available now that are likely worth the effort and price if you need a boost in cooling capacity or battery life over the cheaper peltier equivalents. And they are also used by some people for large bulky cosplay and costume outfits that are otherwise just a super insulated hotbox. Obviously they still have a minimum size limit, but I think many people would be surprised at how small you can get them now.
Are you implying that "normal refrigerant-based coolers" are somehow always the most appropriate solution for any cooling task?
Not the poster but yes. Compressors use phase change which is fundamentally better way of transferring heat. Solids like peltiers will always leak more heat than liquid to gas phase change. Peltiers cannot get the same efficiency because of physics.
So a compressor that's too big to fit and/or too noisy/vibrating for a particular application is still somehow the most appropriate solution?
So this is basically a one-way heat transfer film?
Can it be made multi-layer?
And can two plys be glued back to back so both are trying to transfer heat from center outwards and act as an insulator?
> and act as an insulator?
As a heater in this case.
> Can it be made multi-layer?
Yes, but each layer adds inefficiency and it's own energy.
Seems to be blocking visitors from Vietnam?
https://archive.li/6aT7Y
AI compressor.
It's full on clown world.
What's even AI about it?
The part that's AI about it is the part where it's much easier to get funding if you say it has AI in it.
Much like a hybrid vehicle, this system intelligently switches between the two cooling methods depending on what best suits the situation.
So...it has a microcontroller. Impressive! /s
Have a look at any product marketing these days. Desktop motherboards are calling the fan curve settings AI cooling. Everything that a computer does now, even if it's done it for 30 years is now AI.
It's somehow worse than Blockchain ever was.
It may be hard to believe, but software used to be a fun, intellectually satisfying field.
> Bespoke AI Hybrid Refrigerator
Come the fuck on.
Seriously. When I see this kind of absurdity, I think of Humpty Dumpty in Through The Looking Glass:
> 'When _I_ use a word,' Humpty Dumpty said in rather a scornful tone, 'it means just what I choose it to mean--neither more nor less.'
> 'The question is,' said Alice, 'whether you CAN make words mean so many different things.'
> 'The question is,' said Humpty Dumpty, 'which is to be master-- that's all.'
> Alice was too much puzzled to say anything, so after a minute Humpty Dumpty began again. 'They've a temper, some of them-- particularly verbs, they're the proudest--adjectives you can do anything with, but not verbs--however, _I_ can manage the whole lot of them! Impenetrability! That's what _I_ say!'
> 'Would you tell me, please,' said Alice 'what that means?'
> 'Now you talk like a reasonable child,' said Humpty Dumpty, looking very much pleased. 'I meant by "impenetrability" that we've had enough of that subject, and it would be just as well if you'd mention what you mean to do next, as I suppose you don't mean to stop here all the rest of your life.'
> 'That's a great deal to make one word mean,' Alice said in a thoughtful tone.
If Mr. Dumpty were alive today, he'd be enjoying a brilliant career as a corporate executive.
being colorblind, specifically deuteranomaly, it worked, but i do not see the blue green, i see the weird off green i normally do and then basically brown. i would imagine complete red-green would actually impact it
A lot of people consider them gimmicks but the newer gen neck “air conditioners” do actually cool the air and exhaust the hot air so the peltier works. Would love to see it improved for more cooling capacity (and more airflow).
I wonder if this will eventually lead to battery temperature conditioning on cell phones and laptops similar to electric cars.
Not likely without other contributing technologies. These devices just move heat from one side to the other, at the cost of producing more heat. So you still have to dissipate even more heat on the hot end with a fan/heatsink, in addition to draining the battery running the Peltier.
Can peltier devices stack?
Yes but each stack layer has to remove the waste heat from the inefficiency of all the previous layers, so usually you see them laid out in a pyramid
This article uses the term AI 5 times to talk about basic microcontroller sensor readings.
My process control theory textbook has a chapter on neural networks and a lot of the language in control theory has an AI like tinge to it. I think this AI language is native to control theory so it might not be as overblown as it first sounds.
Huh, control theorists always try to rigorously prove the stability and performance of their algorithms. AI seems to be the opposite of that: just let the black box solve it and don't worry about any problems, we'll just add more training data if they happen!
Yes! Let's spend 100x the resources on AI to do what a PID controller or Kalman filter can do.
I’m less scared about 100x the resources (in most applications) in exchange for decreasing programmer time / need.
What worries me far more is the lack of formalism around risk / boundary cases by undertrained teams using modern AI solutions.
Anyone building on top of a thing should either understand (a) how it’s built in detail or (b) its specifications and behavior in detail.
Most of these teams understand neither about LLMs.
>Most of these teams understand neither about LLMs.
That's where my 100x comes from, not from the dev effort but from the debugging of issues of an unknown black box.
A thermostat is a cybernetic device.
So we need both AI and Kubernetes.
Hating on Kubernetes in 2025 doesn't really dong the gong like it used to, if you want to control your workloads through a well integrated API and not be bound to a cloud vendor there aren't many realistic options.
Kubernetes is only hard because people make it hard and never bothered to understand the basics of their workload scheduler.
Kubernetes is NOT AI hype, it solves real problems for real people everywhere.
"Infrastructure projects" that are here to stay and only getting better: Linux, systemd, Postgres, Kubernetes etc...
I think a lot of the Kubernetes hate comes from non-technical people deciding to use it in cases where it either doesn't actually solve any problem, or the overhead is much greater than the benefit. There's a reason it originated from Google, and as much as it pains MBA factory graduates, the project or organizaton they manage is nowhere near Google scale.
> and as much as it pains [your average software developer], the project or organizaton they manage is nowhere near Google scale.
Fixed.
Independently scale and deploy fridges at edge. Fridge@Edge ™
Edge devices are problematic. I want to outsource the risk and decrease capex (and just hope opex works out).
CloudFridge.
We're going to need decrease load and off-shift the munchies so the fridge doesnt crash. Can we use a CDN?
Comestible distribution network.
Cloud involves losing sovereignty of data. And rabbits.
Local Automated Refrigeration Devices Eat Rabbits.
LARDER.
sorry, us-west-1 went down for 6 hours and my milk spoiled
Definitely has "Unauthorized Bread" vibes.
Suck it, Jian Yang
And Cornelis Drebbel invented the first thermostat and the first chemical air conditioning system in the early 1600s. Cybernetic alchemist.
I firmly believe control theory folks didn’t invent LLMs only because the idea of doing a big fit on everything sounds too much like a joke they were telling each other.
If you type ‘fuzzy logic’ in to google the autocomplete suggested search is ‘fuzzy logic rice cooker’. Control theory has been stealing ML terminology for a long time.
:D Rice cooker marketing is weirdly synchronized on that.
Speaking of, what does it actually mean? That the cooker isn’t using a timer?
Do most of them run off weight + time + heat response logic?
I mean yes, that's what reinforcement learning is.
I always thought my software PID controller with 512kb main memory is quite smart. Pure AI in my opinion.
That's quite a lot of memory for a little PID controller, isn't it? But I guess these days, they already only cost pennies, so you wouldn't save much from going for even less memory?
It is a serious PID controller using an intelligent AI ringbuffer. But yeah, 512kb was probably vastly exaggerated... 512 bytes is probably enough :)
I purchased a book the other day, and when I turn the pages, it automatically shows me the text that I want to read. It must be powered by AI.
Ah, you've discovered the Built-in Orderly Organised Knowledge device [1]. Juse wait until you dicover PENCILS!
[1] https://cool.culturalheritage.org/byorg/abbey/an/an21/an21-8...
It looks like joke today, but it's not a joke anymore.
Book can look at your search history youtube history, or reading history to have insight about your points of interest to create list of topics you like, then generate pages on the fly using LLM.
I imagine theres going to be an entire new class of intelligent humans powered by LLMs that
learn this way, and they are going to be at such a disadvantage to the people that do it the old fashioned way and dont take shortcuts.
The vast majority of humans are going to offload all their critical think skills to LLMs. I dont want to be friends with those people.
I want to start a blog about these people called "Artifically Intelligent"
They exist already, and are the people who watch a lot of cable TV. They are called NPCs
Yeah but I believe the internet rescued a lot of people from that existence, but the people running the internet now are actively working against that.
> I dont want to be friends with those people.
That really limits the friendship selection pool :(.
Quality over quantity.
I like this reply so much.
"AI" has become the seasoning you sprinkle on anything to make it sound futuristic
I wonder if there's a site that keeps track of these. It was crypto before, then ML, for a while a lot of job openings seasoned their text with IoT, etc.
so long as it isn't about ai based crypto running on IoT, not interested
"AI" has become the seasoning you sprinkle on anything to make it sound contemporary.
AI is to software services what avocado is in the food industry.
It is the cheese abd soy sauce of computer science now days
That my friend is the Korean Way (^tm).
If they don’t meet the minimum AI namedrop quota, Seocho Samsung HQ rejects the proposal.
So I'm not defending the use of the term AI in this context, however in this case it's part of the product name (for the most part, there's only a few other references). Does the product use anything that would be regarded as properly AI? No idea, but marketing definitely is going to want you to use the Full Product Name(TM) whenever possible.
> marketing definitely is going to want you to use the Full Product Name(TM) whenever possible
Counterpoint: fuck marketing.
But zero web 2.0 references so we can rejoice.
Nah the new buzz was the Web 3, Web 2.0 is sooo 2010
Web 3 was blockchain nonsense "everywhere" and the buzz has died because the only people who want blockchains are speculators, money launderers, and niche boring users like nightly settlement of inter-bank accounts.
And it was sprinkled everywhere, like sesame seeds on burger, just like Web 2.0 in the 2010 and AI now :) . We just need a Web 3 blockchain Web 2.0 AI FinTech startup
You still see them in the numeric art field along with NFT.
My parent's new fridge has the word AI printed on at least 5 of the stickers that were plastered all over the thing.
And I bet those stickers ripped when trying to remove them, and left a sticky residue on the surface.
The fridge has Ai not the stickers.
If someone has a few exa-flops of compute handy, the sticker thing could do with some attention.
Just a few years ago, the buzzwords were "germicidal", "immunity boosting", "anti-bacterial", etc. for every damn thing, especially in India, where marketing can get away with bizarre claims.
Is there anything that is proved to be "immunity boosting" apart from vaccines?
Well, and germs of course. Germs will cause an immune response, but that's probably not the kind of "boost" you are interested in.
As for AI, I guess it is sufficiently vague and also not a medical claim, so you can put it everywhere. But because it is just as vague for consumers, is it even worth something for marketing?
Depends on your definition of "boosting". Vitamin C is essential for plenty of immune functions and a deficiency has serious effects on immunity. Enormous doses of Vitamin C have been shown to be effective in various scenarios as well. https://pmc.ncbi.nlm.nih.gov/articles/PMC8239596/
While not wrong, that is very misleading - for the vast majority we already get plenty of vitamin C in our diet and we don't have the specific situations where enormous doses will be effective.
At least vitamin C is water soluble: your body will almost always just pee out any extra quick so it is very hard to overdose. Those who have a kidney problem should check with a doctor, for the rest of us massive doses just make our pee more expensive.
>for the vast majority we already get plenty of vitamin C
This is highly dependent on the population you're talking about. There are plenty of people around the world who don't have great nutrition and don't eat large amounts of fortified industrial food.
>for the rest of us massive doses just make our pee more expensive
The study I linked reported ridiculous doses of vitamin C "dramatically improved the clinical state and cardiovascular, pulmonary, hepatic and renal function" in ICU spesis patients. Like 200-400 grams of sodium ascorbate administered over 7 hours. The recommended daily nutritional consumption of vitamin C is around 80 milligrams.
There are other studies reporting other positive outcomes for large doses of vitamin C. This isn't a recommendation that you should take hundreds of grams every day yourself, but there is ample evidence in a clinical setting for "immune boosting" vitamin C in large doses. (and also criticisms of the daily value for being too low)
The vast majority of people are not currently in an ICU though. That situation doesn't apply to us.
If your immune system was "boosted" and worked significantly better tomorrow than it does today, would there be large measurable outcomes? No, probably not. Unless you've got something wrong with you, day to day, it would be hard to detect a distinct improvement in an already working system. Getting a good signal-to-noise ratio for a healthy person getting slightly healthier is hard. The bias for not publishing negative results and various sorts of intentional and unintentional p-hacking means that small-effect-size research that showed say "1 gram of vitamin C per day improved x by 5% in healthy adults" wouldn't be that reliable anyway. "dramatic improvement" of ICU sepsis patients isn't something that suffers from p-hacking.
Someone dying of sepsis in an ICU though, there's a whole bunch of very clear signals whether something is meaningfully effective or not. What that result tells you is that megadoses of vitamin C are likely not bullshit and should continued to be studied in more and more difficult to detect cases.
What this tells me is that 1) because I know it's likely rather safe for me and 2) there's enough evidence to convince me it might help, I'll keep taking lots of vitamin C if I feel like I'm getting sick.
We used to call any sensor with a digital output a smart sensor, although there was nothing smart about it. Now that smart became AI.
I kinda of gave up after the second mention.
How else would we know they used AI???
"Sorry, your subscription has expired and now all your food is mouldy. If you have any complaints, call 1-800-FUCK-YOU and we'll just laugh at you."
Even worse: If you have any complaints, call a number we won't publish anywhere on our site and a robotic voice will just laugh at you.
"We used copilot to hack the code together."