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

I’ve found they are able to compose well, let it build small components and stitch them together

That is a good approach, bottom up, manage complexity. But the general picture is - you set the direction and hold the model responsible, it does the actual work. Think of it as your work is the negative of the AI work, it writes the code, you ensure it tests that code. The better test harness you create, the better the AI works. The real task is to constrain the AI into a narrow channel of valid work.

I hear this argument all the time but it seems to leave out code reviews

In teams of high performers who have built a lot of mutual trust, code reviews are mostly a formality and a stop gap against the big, obvious accidental blunders. "LGTM!"

I do not know or trust the agents that are putting out all this code, and the code review process is very different.

Watching the Copilot code review plugin complain about Agent code on top of it all has been quite an experience.


like a vibe coder knows what a code review is, or an LLM knows how to take feedback

> "Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model"

I believe NVidia’s ProRL showed otherwise right?


I feel like vibe coding has completely changed this paradigm. Now all I care about is how fast and reliable the language is.

Probably throw in a dash of “is my model trained on this language sufficiently to perform this task well”.

Relax

IMO this is mostly just an ego thing. I often see staff+ engineers make up reasons why AI is bad, when really it’s just a prompting skill issue.

When something threatens a thing that gives you value, people tend to hate it


Agree, Google put a ton of work into making TPUs usable with the ecosystem. Given Amazon’s track record I can’t imagine they would ever do that.

There might be enough market pressure right now to make them think about it, but the stock price went up enough from just announcing it so whatever

Amazon has no interest in making their platform interoperable.

Didn’t we just see big pretraining gains from Google and likely Anthropic?

I like Dario’s view on this, we’ve seen this story before with deep learning. Then we progressively got better regularization, initialization, and activations.

I’m sure this will follow the same suit, the graph of improvement is still linear up and to the right


The gains were on benchmarks. Ilya describes why this is a red herring here: https://youtu.be/aR20FWCCjAs?t=286


Gemini 3 is a huge jump. I can't imagine how anyone who uses the models all the time wouldn't feel this.


What does it do that Opus doesn't do?

I like Ilya's points but its also clearly progress, and we can't just write it off because we like another narrative


How do? Isn’t the bitter lesson about more search and compute? As opposed to clever algorithms


The subtext especially around shoving more effort into foreseeable dead-ends is the apt aspect here.


We can't possibly know they are dead ends

Take it with a grain of salt, this is one man’s opinion, even though he is a very smart man.

People have been screaming about an AI winter since 2010 and it never happened, it certainly won’t happen now that we are close to AGI which is a necessity for national defense.

I prefer Dario’s perspective here, which is that we’ve seen this story before in deep learning. We hit walls and then found ways around them with better activation functions, regularization and initialization.

This stuff is always a progression in which we hit roadblocks and find ways around them. The chart of improvement is still linearly up and to the right. Those gains are the cumulation of small improvements adding up.


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

Search: