yanis_t

Simon, is your pelican test really captures differences among models or should you at least try like 10 times or something to average the random effects

If google stops driving traffic to the websites, website owners will have zero incentive put out new content, and then what?

Do we know which model was used for the rewrite?

> the most talented developers I know amplify what they can do with AI

Not the most talented developer, but this has been pretty much my experience as well. Just keep it under control, know what and why its doing at every step, read the code, and then it will boost your productivity.

Good for him. His learning materials are unmatched, but I don’t think there was a viable path with his educational company.

That's really cool. A nice touch is how the help is implement with every widget having a little question mark you can click, and it shows a tutorial on the left panel. Also thank you for letting us in without sign up.

Are there any publicly verifiable sources that Mythos is that much more intelligent than Opus, so to be considered much more dangerous (as it is presented in the public discourse by Anthropic)

So I assume it's because it's not statically build, but requires a DB connection all the time?

My posts* got to the first spot on hackernews couple of times. Never once it broke down like that. And why would it, it's just a bunch of html and css files served through (free) vercel (don't think it matters). I wonder what do people run their blogs these days, so they fail under the pressure so easily.

* https://news.ycombinator.com/from?site=yanist.com

Just my curiosity. Is (insert country) sovereign X is an efficient marketing strategy these days?

This is totally fine. Just don't expect that it will help the industry in EU. If we want to help, we need to push the politians.

Sorry to bring it to you folks, but this is not how you build a competitive industry. You put properly though-through legislation in place if you want EU to be competitive with US and China. Not regulations, for f sake.

Well, I use Vi (NeoVim) and love it. In fact I use it with language servers and Copilot (LLM) and couldn't be happier. You can't get any more modern than that.

It’s funny how many forks aiming to keep it free from LLM-generated code. The luddites are present even in the most progressive parts of the population.

I mean couldn’t they already do that? Isn’t it the whole per pose of agents to do whatever any person can do on a machine?

Anecdotal, but having used Claude Code exclusively for the last several months, I was pleasantly surprised by how capable Pi + Kimi K2.6 is. It's also much faster (via OpenRouter), and at a fraction of the cost.

Here are the listings this thread that allow remote work https://www.whoishiringjobs.com/?workMode=remote (salary and other filters also included)

That reminds me of an old Fry and Laurie sketch.

Well of course too much is bad for you, that's what "too much" means you blithering twat. If you had too much water it would be bad for you, wouldn't it? "Too much" precisely means that quantity which is excessive, that's what it means. Could you ever say "too much water is good for you"? I mean if it's too much it's too much. Too much of anything is too much. Obviously. Jesus.

Came here to write something like that his. It’s so sad we still have these stupid conversations with people thinking black and white in 2026.

Assuming it is almost as good as Opus 4.6 (which benchmarks seem to give evidence for), and assuming we are having a good enough harness (PI, OpenCode), it's is now more than 5x cheaper.

I just want to remind you that this is happening at the same time as Anthropic A/B tests removal of Code from Pro Plan, and as OpenAI releases gpt-5.5 2x more expensive than gpt-5.4...

Is there a harness that is as good as cloud code that can be used with open weight models?

Already on Openrouter. Pro version is $1.74/m/input, $3.48/m/output, while flash $0.14/m/input, 0.28/m/output.

The benchmarks of Opus 4.6 they compare to MUST be retaken the day of the new model release. If it was nerfed we need to know how much.

> In Claude Code, we’ve raised the default effort level to xhigh for all plans.

Does it also mean faster to getting our of credits?

> where previous models interpreted instructions loosely or skipped parts entirely, Opus 4.7 takes the instructions literally. Users should re-tune their prompts and harnesses accordingly.

interesting

I read couple of articles in the series and I still couldn't get what was the point author is trying to make. Reads like, "let me give you 100 arguments why I think this is bad".

Do LLMs lie? Of course not, they are just programs. Do the make mistakes or get the facts wrong? Of course they do, not more often then a human does. So what is the point of that article? Why my future is particularly bad now because of LLMs?

Spec-driven approach is fun. I wonder at which point of anytime at all we are going to commit only specs into the got repo, while the actual code can be generated.

Obviously we’re not here yet because of price, context, and non-determinism, but it’s nice area to experiment with.

I wish we didn't call this AI as the term is crazily overloaded.

Those are programs. The only difference is how we write them. Not with "if"s and "for"s. We take a bunch of bits that do nothing. Then we organize them in a way so that it outputs whatever it is we want.

Also could be a bit both, with harness constructing context in a way that model misinterprets it.

I suppose you can just compare it to France https://app.electricitymaps.com/map/3mo/daily?signal=electri...