(T) @testererr

Interesting. This validates my feeling that on a completely OOD SWE benchmark all open models flop. DeepSeek is way ahead of the pack and still merely on par with Gemini 3.1 (lol). Well at least it's not a psychotic reward hacker. …We'll have to deal with that part somehow.

Lisan al Gaib @scaling01

4 days ago

we are getting spoiled with long-horizon coding benchmarks recently here's another one: SWE-Marathon Website: swe-marathon.org Paper: swe-marathon.org/swe-marathon-p…

keyboard_arrow_left Previous keyboard_arrow_right Next

11 5 225 63K 64

It's satisfying to finally nail down the math behind the EoS plots we’ve been assaulted with for years. I suspect this is the definitive origin story our field has been waiting for. Paper: arxiv.org/abs/2604.20446

Wait, how will *any* enterprise use Fable or Mythos if this is the case?

bruh wat

Cognition @cognition

2 days ago

Introducing FrontierCode: a coding eval that raises the bar for difficulty & quality. Each task took 40+ hrs of work by leading open-source maintainers. Models write sloppy code that works but isn’t maintainable. Our eval is first to measure: would you actually merge this code?

229 312 4K 2.4M 2K

It's official, we have AGI.

Claude @claudeai

15 hours ago

Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for general use. Its capabilities exceed those of any model we’ve ever made generally available.

4K 13K 91K 31.4M 18K

It'll take a while for people to understand how absurdly good this is. we're looking at 710 m^2 or panels + radiators, 120 kW power at 100% capacity factor, and likely ≈2kg per m^2, 12 kg/kW. On Earth, that's at least *270* kg of panels, plus batteries, plus support.

keyboard_arrow_left Previous keyboard_arrow_right Next

Sawyer Merritt @SawyerMerritt

a day ago

SpaceX has just officially unveiled its AI1 satellite, the first generation of its AI satellite. Overall Specs: • 150 kW peak compute payload • 120 kW average compute payload • 70 kW per ton • Compute provider interchangeable Dimensions: • Wingspan: 70 meters • Deployed

416 2K 14K 2.3M 3K

Actually it's fine guys! I figured out a way, see below. Claude Fable 5 is a great model afterall, and I also finally appreciate the difference between CLAUDE.md and AGENTS.md. It's all good.

elie @eliebakouch

14 hours ago

mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community also the fact that this is un purpose not visible to the user is crazy

297 516 5K 1.8M 1K

So I was totally sold on GPT 5.5. Solely on the basis of what I saw tonight with Fable, I will start to use again Claude extensively, probably as the main model as long as Fable is accessible. It's not just very capable, it performs organized steps and have deep understanding.

this is my personal singularity moment this post may sound like a paid ad. I only wish. I'm concerned, more so than happy. the world is changing, and, among the scenarios where AI goes terribly wrong, inequality is the most realistic, yet, the one Anthropic seems to be the least concerned about. I'm glad OpenAI is taking the opposite stance: *personal AGI for everyone*. I think this is a commendable position in the times we live. but who am I in the queue of the bread? anyway, Fable is here, so I'll just report my first-hour experience first of all, all my pet prompts are solved. → λ-calculus puzzles → bug questions → one-shot apps all are trivial to it. I don't have anything harder other than my ongoing work so, in the last several days, I've been toying with HVM5, a new interaction net evaluator with a faster loop. after writing the first version, I left 32 GPT-5 agents working for ~20 hours each. this resulted in up to 2x speedups, but the file size increased by 2-fold and quality decreased significantly. I then simplified the whole thing into an even simpler core, and left Opus 4.8 and GPT 5.5 optimizing it for 8 hours. Opus got a legit 6% - 34% speedup in most benches. GPT got better results, but, sadly, an unusable file. I then asked Fable to optimize it. 2 hours later, it landed a 1770% speedup in one case, 100%+ in other 4, and 22% in average. yes, in 2 hours it outperformed me, opus 4.8 and a swarm of gpt 5.5 agents, by one order of magnitude. that could not possibly be legit. "it must be hardcoding the benchmarks" (GPT trauma). so I read its explanation and what it did was, indeed, the most high impact optimization one could try first. seems like HVM5 was wasting a lot of time garbage-collecting unused branches of pattern-match nodes. I had optimized that for static mats, but not for dynamic mats. skill issue. Fable figured how to do it for these, resulting in a massive speedup in some benches but wait, is that *correct*? I'm not sure yet, it is credible, but this is the kind of thing that is very easy to get wrong on interaction nets. the problem is, when I was ready to start auditing Fable's solution so I could tell whether it was buggy or legit, it interrupted me to tell me it had found a massive bug on the code *I* had written. ... wait, what? so... for garbage collection purposes, I stored a bit on lambda term pointers that meant "the variable bound by this lambda has been freed, so, its lambda must free whatever argument it is applied to". that's fine. yet, on duplicator nodes, I also used the same bit to mean "one of the duplicated variables was freed, so, treat this dup as a passthrough no-op". so, if a lambda entered a duplicator, it would mistake the lambda's collection bit for its own, resulting in corrupted interaction! that's a mouthful, why I'm writing this? just so you can appreciate the sheer absurdity of what just happened. I didn't ask it to find bugs. I asked it for an optimization. and even if I did ask it to find bugs, this bug is so astonishingly subtle and specific, identifying it takes mastering the domain to an extent that it beyond even me. I'd easily need hours or days to fix it, *if* I ever came across it. chances are it would just go unnoticed. and Fable found it and fixed it like it was nothing, while it was busy adding a 17x speedup to a file that neither I, nor Opus 4.8, nor a fleet of GPT 5.5 managed to barely make 2x faster. oh and there is also another tab where it is also ripping through Bend's codebase and finishing everything I had to do I don't know what to say anymore this isn't about Anthropic or OpenAI, this is about our collective future as a species. the world is changing, and we need to be aware of it, and discuss how to handle this change. receipt below . . .

keyboard_arrow_left Previous keyboard_arrow_right Next

When Fable 5 is used for frontier LLM development, it does not notify the user and instead limits the model’s capabilities through methods such as prompt modification, steering vectors, and PEFT. Anthropic estimated that this would affect approximately 0.03% of traffic.

You're not even allowed to ask Fable about basic biology questions, let alone anything that could potentially be dangerous.

@claudeai I get the goal to stimulate interest and maintain an air of exclusivity, but this constant angle shooting is going to piss people off more than excite them. We know what you are doing; this is not some slick routine. Just put the fries in the bag guys.

Fable 5 is state-of-the-art on nearly all tested benchmarks, with exceptional performance in software engineering, knowledge work, scientific research, and vision. The longer and more complex the task, the larger Fable 5’s lead over our other models.

Apple are saying they have rewritten parts of macos kernel in Swift!

Ivan Fioravanti ᯅ @ivanfioravanti

20 hours ago

macOS 27 Beta 1 Apple Foundation Models context benchmark test! 🔥 System model tested here, I have no access to pcc (Private Cloud Compute), yet 😎 I recorded a video to show the very minimal load on the system, sounds strange to me. Hardware: Apple M5 Max, 128.0GB RAM, 18

keyboard_arrow_left Previous keyboard_arrow_right Next

16 11 137 24K 41

"White privilege" is a racist lie. DEI is blatant systemic racism. Woke is a form of social Marxism. Reject it all. No more discrimination.

I admire Fabrice Bellard. He is almost certainly a better overall programmer than I am.

Spencer Baggins @bigaiguy

2 days ago

A French engineer who lives quietly in Paris has spent 30 years writing software that the entire internet now runs on without knowing his name. He wrote the code that streams every YouTube video, every Netflix show, every TikTok clip. He wrote the code that runs the virtual

356 4K 24K 2.9M 11K

A simply superb paragraph by Sam Harris.

Dear Muslims who believe that "Israel is only 76 years old," The word Israel is mentioned 43 times in the Quran. Israel is not younger than your grandma. Israel is older than Islam.

Japan and Germany were devastated during WW2. Their children didn’t become terrorists. Stop making excuses for your terrorists.

Multiculturalism means embracing every other culture over the West’s. Under the guise of “inclusion,” it acts as a Trojan horse that delivers the total erasure of Western culture in practice. Every person and organization promoting it is an enemy of the West.