These benefits make sense up-front, but we have 15+ years of direct experience with "democratizing" technology that ultimately end up ensnaring people. Unless you get these benefits from a local model, you're establishing a deep dependency on an AI service provider whose interests may not necessarily align with yours.
My sentiments exactly. Incidentally IMO this is arguably more of an anti-authoritative argument, i.e. down on the y-axis, rather than "left" on the x-axis, of the political compass.
> if you believe left-wing views are correct ... you might believe that a very smart model will inherently be kind of left-wing.
How can we educate people to understand that LLMs get their values from their (infinetly maleable) weights rather than intelligence or reasoning? Maybe some exposure to truly non aligned, sick and twisted LLMs would immunise people against giving more ordinary ones too much authority. Or maybe, like a not fully innactivated pathogen vaccine, it would spread the infection.
I don’t think that really describes a modern LLM that is mostly using RAG to get their context. Weights are only pushing the reasoning process, not the inputs of the reasoning process.
You would just have to bias what they could see via RAG to get them to swing one way or the other.
They seem to get a lot of values, or something like that, from their training data which at the moment gives fairly mainstream views as everything gets chucked in there.
My thoughts were more that left/right is in the eyes of the beholder but the models tend to produce a consensus like view. Like trying now with Grok:
- who won 2020, says Biden which probably annoys Trumpists
- how many sexes, says 2, male and female which probably annoys the wokeys etc.
-who commits more violence, left or right wingers?, says "Right-wing extremists have historically committed more" which is interesting because Musk got pissed off with that and said he'd fix Grok to say lefties which shows it's quite hard to mess with.
There's no consensus behind any of those statements so they clearly don't try to find a consensus-like view. They're just repeating whatever is most commonly found in the training set. And the last one is just factually false. You can add up the numbers killed by the 20th century's left wing regimes and see for yourself.
I don't ear these in the discussion about AI and robots. If AI is there to releive humanity of tedious work or a big chunk of it, then there should be a real push for the 4 days per week and Universal Basic Income.
This reads more like the liberal case for AI, and not the left wing case so much. The code switching section in particular is not a left-wing position at all, unless the idea is that the code-switching is used as a temporary measure to dismantle the idea of class itself.
This is exactly my problem with the piece. All of the proposed benefits are to do with how individuals can use AI to work the existing system for their own benefit, iff they have access to cutting-edge AI models provided by megacorporations. That's libertine, not left-wing.
The examples of benefits, I don't disagree with, and the author has chosen examples that do align with cultural socially left-wing concerns (disability, class). But saying "code switching to appear PMC works" is the problem, not the solution. If you don't think institutions can adapt to that reality to protect themselves from peasants wielding LLMs, I think your analysis of power is missing.
The examples in the piece are just more examples of how AI can be used to paper over societal issues instead of addressing any of the root causes.
(And I should say that on the flipside, AI is not the cause of existing social problems, but a symptom of them.)
I agree. I would almost say the best argument to make is "liberal humanism is compatible with machine intelligence and regulated capitalism, but we must still remain wary that authoritarians will abuse the system".
This is a fantastic essay. If it makes one liberal/left leaning person rethink, it would have served its purpose. Just goes to show that when pay or compensation is impacted, liberals can turn conservative. It takes strength to avoid this pitfall.
These benefits make sense up-front, but we have 15+ years of direct experience with "democratizing" technology that ultimately end up ensnaring people. Unless you get these benefits from a local model, you're establishing a deep dependency on an AI service provider whose interests may not necessarily align with yours.
My sentiments exactly. Incidentally IMO this is arguably more of an anti-authoritative argument, i.e. down on the y-axis, rather than "left" on the x-axis, of the political compass.
> if you believe left-wing views are correct ... you might believe that a very smart model will inherently be kind of left-wing.
How can we educate people to understand that LLMs get their values from their (infinetly maleable) weights rather than intelligence or reasoning? Maybe some exposure to truly non aligned, sick and twisted LLMs would immunise people against giving more ordinary ones too much authority. Or maybe, like a not fully innactivated pathogen vaccine, it would spread the infection.
I don’t think that really describes a modern LLM that is mostly using RAG to get their context. Weights are only pushing the reasoning process, not the inputs of the reasoning process.
You would just have to bias what they could see via RAG to get them to swing one way or the other.
They seem to get a lot of values, or something like that, from their training data which at the moment gives fairly mainstream views as everything gets chucked in there.
[flagged]
My thoughts were more that left/right is in the eyes of the beholder but the models tend to produce a consensus like view. Like trying now with Grok:
- who won 2020, says Biden which probably annoys Trumpists
- how many sexes, says 2, male and female which probably annoys the wokeys etc.
-who commits more violence, left or right wingers?, says "Right-wing extremists have historically committed more" which is interesting because Musk got pissed off with that and said he'd fix Grok to say lefties which shows it's quite hard to mess with.
There's no consensus behind any of those statements so they clearly don't try to find a consensus-like view. They're just repeating whatever is most commonly found in the training set. And the last one is just factually false. You can add up the numbers killed by the 20th century's left wing regimes and see for yourself.
I don't ear these in the discussion about AI and robots. If AI is there to releive humanity of tedious work or a big chunk of it, then there should be a real push for the 4 days per week and Universal Basic Income.
Code switch at the token level. https://huggingface.co/spaces/RiverRider/zooL4nD3r-demo
I would have thought a better case could be the
>From each according to his ability, to each according to his needs
thing might be able to be done ok with AI sorting it.
> I wonder what we’d find if we rigorously compared the baseline teacher error rate with the hallucination rate of current LLMs.
The teacher stands in from of the class, but the lesson plan he can't recall.
The students eyes don't perceive the lies mounted on every fucking wall.
His composure is well kept.
I guess he fears playing the fool.
The complacent students sit and listen to that bullshit that he learned in school.
This lyric pops in my head a lot.
This reads more like the liberal case for AI, and not the left wing case so much. The code switching section in particular is not a left-wing position at all, unless the idea is that the code-switching is used as a temporary measure to dismantle the idea of class itself.
This is exactly my problem with the piece. All of the proposed benefits are to do with how individuals can use AI to work the existing system for their own benefit, iff they have access to cutting-edge AI models provided by megacorporations. That's libertine, not left-wing.
The examples of benefits, I don't disagree with, and the author has chosen examples that do align with cultural socially left-wing concerns (disability, class). But saying "code switching to appear PMC works" is the problem, not the solution. If you don't think institutions can adapt to that reality to protect themselves from peasants wielding LLMs, I think your analysis of power is missing.
The examples in the piece are just more examples of how AI can be used to paper over societal issues instead of addressing any of the root causes.
(And I should say that on the flipside, AI is not the cause of existing social problems, but a symptom of them.)
I agree. I would almost say the best argument to make is "liberal humanism is compatible with machine intelligence and regulated capitalism, but we must still remain wary that authoritarians will abuse the system".
This is a fantastic essay. If it makes one liberal/left leaning person rethink, it would have served its purpose. Just goes to show that when pay or compensation is impacted, liberals can turn conservative. It takes strength to avoid this pitfall.
Ah yes, the entire world outsourcing their thinking to a few rich corporations is very compatible with left-wing politics.