191–200 of 1001 posts
Earlier quoted context omitted.
Even that, we don’t know what got updated and what didn’t. Can we assume everything that can be updated is updated?
You might be able to ask it what it knows.
https://claude.ai/share/59818e6c-804b-4597-826a-c0ca2eccdc46
>This is a topic that would have developed after my knowledge cutoff of January 2025, so I should search for information [...]
> Users requiring raw chains of thought for advanced prompt engineering can contact sales So it seems like all 3 of the LLM providers are now hiding the CoT - which is a shame, because it helped to see when it was going to go down the wrong track, and allowing to quickly refine the prompt to ensure it didn't. In addition to openAI, Google also just recently started summarizing the CoT, replacing it with an, in my opi…
Is this really worthy of a claude 4 label? Was there a new pre-training run? Cause this feels like 3.8... only swe went up significantly, and that as we all understand by now is done by cramming on specific post training data and doesn't generalize to intelligence. The agentic tooluse didn't improve and this says to me that it's not really smarter.
So I decided to try Claude 4 Sonnet against my "Given a list of 1 million random integers between 1 and 100,000, find the difference between the smallest and the largest numbers whose digits sum up to 30." benchmark I tested against Claude 3.5 Sonnet: https://news.ycombinator.com/item?id=42584400 The results are here ( https://gist.github.com/minimaxir/1bad26f0f000562b1418754d67... ) and it utterly crushed the proble…
Almost guaranteed, especially since HN tends to be popular in tech circles, and also trivial to scrape the entire thing in a couple of hours via the Algolia API.
Recommendation for the future: keep your benchmarks/evaluations private, as otherwise they're basically useless as more models get published that are trained on your data. This is what I do, and usually I don't see the "huge improvements" as other public benchmarks seems to indicate when new models appear.
Earlier quoted context omitted.
Gemini has beat it already, but using a different and notably more helpful harness. The creator has said they think harness design is the most important factor right now, and that the results don't mean much for comparing Claude to Gemini.
Way offtopic to TFA now, but isn't using an improved harness a bit like saying "I'm going to hardcore as many priors as possible into this thing so it succeeds regardless of its ability to strategize, plan and execute?
The naming scheme used to be "Claude [number] [size]", but now it is "Claude [size] [number]". The new models should have been named Claude 4 Opus and Claude 4 Sonnet, but they changed it, and even retconned Claude 3.7 Sonnet into Claude Sonnet 3.7. Annoying.
It seems like investors have bought into the idea that llms has to improve no matter what. I see it in the company I'm currently at. No matter what we have to work with whatever bullshit these models can output. I am however looking at more responsible companies for new employment.
I'm a (minor) investor, and I see this a lot: People integrate LLMs for some use case, lately increasingly agentic (i.e. in a loop), and then when I scrutinise the results, the excuse is that models will improve, and _then_ they'll have a viable product.
I currently don't bet on that. Show me you're using LLMs smart and have solid solutions for _todays_ limitations, different story.
It’s been hard to keep up with the evolution in LLMs. SOTA models basically change every other week, and each of them has its own quirks. Differences in features, personality, output formatting, UI, safety filters… make it nearly impossible to migrate workflows between distinct LLMs. Even models of the same family exhibit strikingly different behaviors in response to the same prompt. Still, having to find each model’…
How important is it to be using SOTA? Or even jump on it already? Feels a bit like when it was a new frontend framework every week. Didn't jump on any then. Sure, when React was the winner, I had a few months less experience than those who bet on the correct horse. But nothing I couldn't quickly catch up to.
I believe in using the best model for each use case. Since I’m paying for it, I like to find out which model is the best bang for my buck.
The problem is that, even when comparing models according to different use cases, better models eventually appear, and the models one uses eventually change as well — for better or worse. This means that using the same model over and over doesn’t seem like a good decision.
Have they documented the context window changes for Claude 4 anywhere? My (barely informed) understanding was one of the reasons Gemini 2.5 has been so useful is that it can handle huge amounts of context --- 50-70kloc?
Context window is unchanged for Sonnet. (200k in/64k out): https://docs.anthropic.com/en/docs/about-claude/models/overv... In practice, the 1M context of Gemini 2.5 isn't that much of a differentiator because larger context has diminishing returns on adherence to later tokens.
I think it also crushes most of the benchmarks for long context performance. I believe on MRCR (multi round coreference resolution) it beats pretty much any other model's performance at 128k at 1M tokens (o3 may have changed this).
Earlier quoted context omitted.
Even that, we don’t know what got updated and what didn’t. Can we assume everything that can be updated is updated?
You might be able to ask it what it knows.
Earlier quoted context omitted.
It feels like the days of Claude 2 -> 3 or GPT 2->3 level changes for the leading models are over and you're either going to end up with really awkward version numbers or just embrace it and increment the number. Nobody cares a Chrome update gives a major version change of 136->137 instead of 12.4.2.33 -> 12.4.3.0 for similar kinds of "the version number doesn't always have to represent the amount of work/improvement…
It feels like LLM progress in general has kinda stalled and we're only getting small incremental improvements from here. I think we've reached peak LLM - if AGI is a thing, it won't be through this architecture.
That said, whether or not being a provider of these services is a profitable endeavor is still unknown. There's a lot of subsidizing going on and some of the lower value uses might fall to the wayside as companies eventually need to make money off this stuff.