I just built an app for a problem I have: newsletter overload. I subscribe to 15+ newsletters but only read 2-3, never at the right time, and most content isn't relevant to me.
So I built Upletter (https://upletter.app), you describe what you want to read about, set your preferred tone/style, choose when to receive it (e.g., "Tuesday & Thursday, 9 AM"), and AI generates a personalized newsletter just for you.
Current state: Extremely early MVP. You can create one newsletter with custom settings. That's it. I'm validating whether this is something people would actually want before building out the full system.
Questions I have:
1. Would you use something like this?
2. What's missing or wrong with the concept?
3. How is this better/worse than just asking ChatGPT daily?
4. What would make you actually pay for this?
I know it's rough, but I'd love honest feedback before I invest more time building the wrong thing. Roast it, please!
The idea isn't bad, but the execution of the idea will make it or break it. If you code up a simple LLM wrapper, it will suck, because it will just hallucinate. At the other extreme, if it just spits back URLs that look interesting, you aren't offering anything that can't be solved by an RSS reader. Finding the correct balance of applying some LLM assistance, but not giving it enough rope to turn into AI slop would be the secret sauce.
Getting people to pay for it is a whole other question. We've built an online culture where people don't pay for content unless it is a streaming service of some kind. I think there is potential to change that, but you'd be fighting that battle at the same time as trying to nail the correct product. Not impossible, but this is probably a harder path to success that it would seem at first glance.
Hey HN
I just built an app for a problem I have: newsletter overload. I subscribe to 15+ newsletters but only read 2-3, never at the right time, and most content isn't relevant to me.
So I built Upletter (https://upletter.app), you describe what you want to read about, set your preferred tone/style, choose when to receive it (e.g., "Tuesday & Thursday, 9 AM"), and AI generates a personalized newsletter just for you.
Current state: Extremely early MVP. You can create one newsletter with custom settings. That's it. I'm validating whether this is something people would actually want before building out the full system.
Questions I have:
1. Would you use something like this? 2. What's missing or wrong with the concept? 3. How is this better/worse than just asking ChatGPT daily? 4. What would make you actually pay for this?
I know it's rough, but I'd love honest feedback before I invest more time building the wrong thing. Roast it, please!
The idea isn't bad, but the execution of the idea will make it or break it. If you code up a simple LLM wrapper, it will suck, because it will just hallucinate. At the other extreme, if it just spits back URLs that look interesting, you aren't offering anything that can't be solved by an RSS reader. Finding the correct balance of applying some LLM assistance, but not giving it enough rope to turn into AI slop would be the secret sauce.
Getting people to pay for it is a whole other question. We've built an online culture where people don't pay for content unless it is a streaming service of some kind. I think there is potential to change that, but you'd be fighting that battle at the same time as trying to nail the correct product. Not impossible, but this is probably a harder path to success that it would seem at first glance.