benjaminsky2

I’ve validated Jev’s confidence score. Accuracy scales linearly with confidence for the 3 use cases I tested. >.9 it matched a human labeler. I immediately discovered a user behavior I didn’t expect for ~$3. I can now mitigate in real-time due to low cost and latency. This may have a major positive financial impact for all our customers.

Not sure why anyone would feel the need to dunk on this thing without showing a real failure example.

Omg, did anyone else have to solve like 15 captchas while trying to signup for alerts? I gave up.

If you ever end up incarcerated this will be invaluable:)

I read the whole thing. Seems pretty fair to me, including the section prohibiting discrimination against conservative ideas. With all the coverage I expected something crazy.

Vlad the Implier

If your project is well organized and individual files are small and the dependency graph isn’t too crazy, Claude code does an amazing job building only the context it needs even as the project grows. You just have to be aggressive about refactoring for maintainability. The bonus here is that it’s easier for humans to work on too.

This is awesome! I’m work with a team of analysts and data engineers who own a pretty big snowflake data warehouse. We write a ton of dbt models and have a range of sql skill levels on the team. This would be the perfect way to allow more junior devs to build their skills quickly and support more complex models.

I would recommend you target data warehouses like snowflake and bigquery where the query complexity and thus value prop for a tool like this is potentially much higher.

Not to mention that they all use or amplify the exact same talking points at the exact same time.

I’ve always built analytics daily by default, and certain metrics real-time - via a different system - only when a compelling and permanent use case exists (except operational analytics). I mean something like YouTube view counter/likes or delivering order count/total as a b2b2c so your customers can watch their promotions in real-time. Major bonus, you have the super complicated, hard to maintain, real-time system to point to when someone wants everything real-time but invested the minimum possible.

Maybe the star is a bigger publication than I’m aware of but it seems pretty hilarious for them to have a paywall.

To extend the lossy compression hypothesis, human generated text is lossy compression of our sensory experience of reality while LLMs are lossy compression of that.

I’m willing to give this a try if you can clean it up a bit. UX is rough on mobile. I also prefer subs to hashtags but whatever, I’m easy. I bailed on Reddit like 2yrs ago and there’s really no alternative that I can find except HN.

I did that the other day. I cooked a kadai chicken recipe verbatim from ChatGPT. It wasn’t bad, but it wasn’t amazing either.

> artist whose domain has not yet been disrupted by AI fires artist in favor of AI

“Person I don’t like acts in rationally self interested manner, legally outperforming less informed parties in zero sum game”

Should we feel bad when we sell a stock and it goes down?