Hacker Newsnew | past | comments | ask | show | jobs | submit | az226's commentslogin

It is not crazy.

Each engineer is very valuable. LLM tokens are cheap. You scale up inference compute, and your engineers can focus on higher order stuff, not reviewing incorrect responses, validating bugs, and what not.

It’s shocking to me that there isn’t a $2,000 / $20,000 per month subscription tier for coding assistants. I’ve always in my mind called this ExecGPT since around 2021, but the notion was that executives have teams that support them to be high functioning and high leverage, responsible for quality of thinking and decision making, not quantity of work output.

And the value/prop existed and continues to exist even as the models get smarter, even Opus 4.6.


> It’s shocking to me that there isn’t a $2,000 / $20,000 per month subscription tier for coding assistants.

What would be the benefit for the providers in offering this over just having those people use the API? I don't think it makes any sense for them.


Most people don’t know how to build an effective harness that is well optimized and battle tested to work for a wide range of scenarios. But they can happily and easily use a finished service if they have the money for it.

My questions asks about the benefit for the providers. The harnesses (Claude Code, Codex) already work with the metered API which is what all those potential "$20k subscription" customers are already using.

You have a different agent write the tests and another run the tests. You tell them each that they aren’t checking their own work, they’re checking someone else’s. You can tell them to be skeptical. Then you can also tell them that don’t fail the code for no reason, because a third agent will be checking your tests and you will be penalized for inaccurate testing.

This approach balances out and maximizes accuracy.


> just use psychological tricks on the LLM, bro, you'll cajole it into not hallucinating

Can't help but chuckle at that.


If it sounds stupid but it works, it's not stupid.

> but it works

Proofs?


Feels very much like a Flappingbird with a dash of AI grift.

It’s possible that AWS is a Wiz customer, which would allow them to do more stuff.


I’d guess that we would not have had the pleasure of reading this article if wiz was payed by AWS. There were multiple high impact bug in 2025 that we read about here, where security researchers had to turn down small six figure bounties to avoid NDAs…


100% this.

They could have made their service better or more competitive like with price but instead they chose this route. SMH.



This is very impressive. Google really is ahead


They are definitely ahead in multi modality and I'd argue they have been for a long time. Their image understanding was already great, when their core LLM was still terrible.


This is genuinly impressive. The OpenAI equivalent is less detailed AND less correct.


When OpenAI Marketing Material is actually showing how far Gemini3 is ahead...



Widex Moment 440. They’re expensive but really good.


I had the chance to test Moment while my Widex Evoke 440 had been in repair. They are indeed a lot better as the Evoke, sounding wise, while having only mild hearing loss. But they are still only MFI (can only connect to an iphone) and since I am migrating away from Apples ecosystem this will be a blocker.


I have the same and really like them other than the BTE still require an induction loop necklace for Bluetooh.


How many times have you needed to reset the optimizer during the RL training cycles?


Oh no! They lowballed their own valuation. Capped profit shares are much less valuable. Who is going to allow this self-dealing to stand?


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: