11–20 of 39 posts
It's like reading a self-written obituary of someone who stopped thinking for themselves. They aren't excited about anything. They aren't in awe. They haven't done any hard work. They're just here to ooze lukewarm sludge
A lot of negative comments on here, which seems to always be the case with HN and vibe coding. The reality is that it’s actually starting to work, quite well.
I’ve been putting questions into LLM research functions, including Claude’s research mode, and letting them churn until a report appears. I’ve been starting with topics where I’m already familiar with the answer but want a refreshed. So far, I’m not impressed. Some times the info will be correct. Most of the time it strings together a lot of words from the material it finds but it reads like an undergrad trying to pa…
1) A response originating from LLM pre-training, in a domain where there has not been any (successful) Rl-for-reasoning post-training. In this case the amount of reasoning around the raw facts "recalled" by the LLM is going to be limited by any reasoning present in the training data.
2) A non-agentic response in a domain like Math Olmypiad problems where the LLM was post-trained with RL to encourage reasoning mirroring this RL training set. This type of domain-specific reasoning training seems to have little benefit to other domains (although in the early LLM days it was said that training on computer code did provide some general benefit).
3) An agentic response, such as from one of these research systems, where it seems the agent is following some sort of generic research / summarization template with proscribed steps. I've never tried these myself, but it seems they can be quite successful in deep diving and gathering relevant source material, but then the ability to reason over this retrieved material is going to come down to the reasoning capability of the underlying model per 1) and 2) above.
Bottom line would seem to be that with today's systems domain specific reasoning capability largely comes down to RL post-training for reasoning in that specific domain, resulting in what some call "jagged" performance - excellent in some areas and very poor in others. Demis Hassabis, for one, seems to be saying that this will not be fixed until architectural changes/additions are made to bring us closer to AGI.
> OpenAI has released GPT-5-codex, and supposedly uses it ... to automate a lot of their ... AI research If I was the owner of an AI company that was forever trying to juice its valuation and raise money, you can bet I'd be telling people I had built a magic self-improving AI.
And even with pure model development, making incremental changes to try different strategies in notebooks etc are probably not that hard to write, when given clear instructions by a data scientist, etc. (I’m not saying these disciplines are easy, I’m saying that a data scientist could more easily describe what they want)
Or the devops stuff. Or the RL UIs.
All that stuff is run of the mill software in service of building the models. And it can be vibe coded.
Same for tasks you know how to do but AI does them faster, there is also value there. (Claude Code used by a senior goes here)
The interesting thing is when AI is lifting up the ceiling everywhere, but maybe then is when we are almost on AGI territory.
The title makes it sound like we have reached the singularity. The real insight here is that amateurs may have a difficult time competing with AI
The author wasn't doing "AI research" before and neither was GPT5. This is not at the frontier of anything, it is just an already solved problem in training that GPT5 found. Had the author been willing to actually do a Google and GitHub search, or just twiddle the training knob parameters enough on their own, they would have found a better solution than working alone. Also this footnote: > Alone” here is relative - I…
>I built significant pieces of the Copilot onboarding, purchasing, billing and settings flow. For eight months I headed up the Copilot anti-abuse effort. I then led the launch of GitHub Models, and am now working on other Copilot projects.
As an aside I had a look at GitHub Models and it was quite interesting - you can try the API for a number of models for free using your GitHub login.
It's like reading a self-written obituary of someone who stopped thinking for themselves. They aren't excited about anything. They aren't in awe. They haven't done any hard work. They're just here to ooze lukewarm sludge
Now do the same for people who buy their food at the grocery store instead of growing it themselves like our ancestors did.
> OpenAI has released GPT-5-codex, and supposedly uses it ... to automate a lot of their ... AI research If I was the owner of an AI company that was forever trying to juice its valuation and raise money, you can bet I'd be telling people I had built a magic self-improving AI.
Nah I think the more obvious thing is that most of the code they’re writing (by sheer volume) is the same stuff most programmers write (code that does telemetry, APIs, usage models, billing, product features, react widgets, etc etc.) stuff that the AI coding models do really well. And even with pure model development, making incremental changes to try different strategies in notebooks etc are probably not that hard t…
I’ve had the exact same experience. I’ve been vibe coding most of my research now, previously was an MLE handcrafting model code. A lot of negative comments on here, which seems to always be the case with HN and vibe coding. The reality is that it’s actually starting to work, quite well.
What happened to the MLE? Are they all going to end up that way?
> A lot of negative comments on here, which seems to always be the case with HN and vibe coding. The reality is that it’s actually starting to work, quite well.
It's hard to be positive about the idea of your skills getting devalued and getting kicked to the curb.