Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AlphaEvolve: Gemini-powered coding agent scaling impact across fields

deepmind.google

Post reply on HN
Print view View original

121–130 of 159 posts

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#121

Post by DeathArrow »

danpalmer wrote: post #108
alecco wrote: post #9

Are Googlers themselves happy using Gemini coding agent instead of Claude Code or Codex? (no snark, I'm really asking)

Yes. The models are good, the models are fast , and the internal tooling has caught up at this point too. There's a lot of UI/UX/tooling stuff that's still being worked through, integrations with VCS, and solving deeper problems that I probably can't talk about, but I'd say the frustrations of most are about the rate of change much more than the actual abilities. One thing that's interesting is a bunch of internal th…

>One thing that's interesting is a bunch of internal thought leaders who swear by the Flash models over the Pro models.

I've seen people outside Google favoring flash Gemini models over the Pro.

There are also some benchmarks where flash models have higher scores, so yes, apparently speed does matter.

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#122

Post by DeathArrow »

Groxx wrote: post #52
dekhn wrote: post #30

Earlier quoted context omitted.

If you mean specifically the Gemini VS Code Extension: it's terrible compared to Claude Code or Codex. I don't know how they can get away with it. Just constant timeouts, weird failure modes, have to start a new chat to switch modes... but I don't think any of that is specific to gemini the model- it seems to be the extension. As for actual solutions to problems ignoring the VS Code extension aspect, I find all three…

The overall quality of LLM coding tools is shockingly bad. I haven't found a single one without major issues, and many have the same problems reappear every few months, sometimes bad enough to almost break the entire thing (e.g. 100% failure rate in editing files, broken for weeks, with the same cause each time, multiple times in a year). I'd say I'm surprised by it, but uh

>The overall quality of LLM coding tools is shockingly bad

Most of them were vibecoded in days, so what do you expect? And new versions just add features, they never fix the old cruft.

Probably there would be some money to be made if someone actually takes the time to write a good agent harness.

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#123

Post by DeathArrow »

nine_k wrote: post #16
alecco wrote: post #9

Are Googlers themselves happy using Gemini coding agent instead of Claude Code or Codex? (no snark, I'm really asking)

The point of dogfooding is exactly that: if we're unhappy, we're the ones to improve.

Are you in the Gemini team?

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#124

Post by 01100011 »

vonneumannstan wrote: post #40
momojo wrote: post #25

This reminds me of Antirez's "Don't fall into the anti-AI hype" [0] In a sentence: These foundation models are really good at optimizing these extremely high level, extremely well defined problem spaces (ie multiply matrices faster). In Antirez's case, it's "make Redis faster". There have been two reactions: "Oh it would never work for me" and "I have seen months of my life accomplished in an hour", and I think they'…

>I think the rest of us should rest easy knowing that LLM's can't (and maybe were never meant to) tackle the tacit-knowledge-filled, human-system-centric, ambiguously-defined-problem-space jobs most mortals work. A Statement all but guaranteed to look incredibly short sighted by 2030.

The past few years has seen a great rise in casuals reminding us of AIs limitations only to be proven wrong in 6 months. I don't think we're close to AGI, but in 2 years I've gone from AI doubter to AI convert. It's not perfect, but I don't need it to be.

The real question to me is if the system can pay for itself. Economics are racing against efficiency gains and it's anyone's guess which wins.

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#125

Post by Davidzheng »

HarHarVeryFunny wrote: post #72
pingou wrote: post #6

AI improving itself (or at least the architecture it runs on), the singularity is near as they say. Do we have other examples of AI being used to improve the LLMs, apart for the creation of synthetic data and the testing of the models?

There is an apples and oranges difference between AI improving itself (becoming more capable) and AI optimizing software that happens to be used for AI training or inference. A more efficient transformer just costs less to run. "AI improving AI" would be if one generation of AI designed a next-gen AI that was fundamentally more capable (not just faster/cheaper) than itself. A reptilian brain that could autonomously d…

I don't think there is a fundamental divide between implementation speedups and optimization and algorithmic/architecture optimizations

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#126

Post by fractorial »

danpalmer wrote: post #108
alecco wrote: post #9

Are Googlers themselves happy using Gemini coding agent instead of Claude Code or Codex? (no snark, I'm really asking)

Yes. The models are good, the models are fast , and the internal tooling has caught up at this point too. There's a lot of UI/UX/tooling stuff that's still being worked through, integrations with VCS, and solving deeper problems that I probably can't talk about, but I'd say the frustrations of most are about the rate of change much more than the actual abilities. One thing that's interesting is a bunch of internal th…

You’re absolutely kidding yourself if you genuinely believe that.

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#127

Post by exfalso »

risyachka wrote: post #49
poisonfountain wrote: post #34

Earlier quoted context omitted.

>I think the rest of us should rest easy knowing that LLM's can't (and maybe were never meant to) tackle the tacit-knowledge-filled, human-system-centric, ambiguously-defined-problem-space jobs most mortals work I don't believe that anymore, to be honest. Models are starting to get good at ambiguity. Claude Code now asks me when something is ambiguous. Soon, all meetings will be recorded, transcribed and stored in a…

In coding the ambiguity is very, very limited and constrained compared to any non dev job that involves any decision making

That's.. not even close to being the case. It's literally a series of ambiguous questions and strategic decisions.

Non-ambiguous is like a first semester algorithms class in university.

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#128

Post by cpard »

All the *Evolve publications have very impressive results but from the time I’ve spent on the information published I feel that the attention goes to the LLMs and the AI side of things, although the outcomes reported are in almost all cases the result of very well designed environments for both the LLM and the evolutionary algorithm to work well.

This paper here is a great example of that and it’s worth a reading.

Magellan: Autonomous Discovery of Novel Compiler Optimization Heuristics with AlphaEvolve https://arxiv.org/abs/2601.21096

Re: AlphaEvolve: Gemini-powered coding agent scaling impact across fields

#130

Post by nl »

mrdomino- wrote: post #119
nl wrote: post #109

Earlier quoted context omitted.

The labs are spending hundreds of millions of dollars hiring people doing many fairly random (but economically valuable) tasks to collect this tacit knowledge for RL. It works really well.

It ceases to become tacit as soon as it is collected. Maybe this rephrase will help: the proposed solution is to render all knowledge explicit.

> It ceases to become tacit as soon as it is collected.

I'm not sure.

It it is collected via preferences then it isn't necessarily something that can be communicated (except in the LLM's latent space).

That still feels tacit to me.

To simplify that argument, the relationship between King and Queen in the Word2Vec latent space can be easily explicitly labelled.

But the relationship between Napoleon and Tsar Alexander I also exists and encodes much of the tacit knowledge about their relationship but isn't as easily labelled (eg, Google AI Mode says "Napoleon I and Tsar Alexander I had a volatile "bromance" that shifted from mutual admiration to deep animosity, acting as a defining conflict of the Napoleonic Wars".)

Word2Vec is a very simple model. In a more complex LLM that deeper knowledge can be queried by asking questions but you can never capture it all. Isn't that what "tacit knowledge" is?

Post reply on HN