StandingOnTheMoon

Recursively Self-Improving...

[StandingOnTheMoon @entangledQbit]

@GaryMarcus Sounds like the consumer wins

[StandingOnTheMoon @entangledQbit]

@MIT_CSAIL In older English circles we call it "pure unadulterated vibe coding"

[StandingOnTheMoon @entangledQbit]

@Ryansikorski10 @donotconform2 @ucworkroom1 @lake7717 @grok what's the executive summary here?

[StandingOnTheMoon @entangledQbit]

@tszzl Feeling

[StandingOnTheMoon @entangledQbit]

It's apparently true about Fable:

[StandingOnTheMoon @entangledQbit]

"We're starting to converge in the direction of superintelligence but you can't use it for any kind of genuinely positive human progress"

This is essentially what @DarioAmodei is telling us.

Incredible!

[Robert Scoble @Scobleizer]

"Misanthropic."

I've never seen the AI community so angry at a major new model release. I asked my AI (an agent that @blevlabs made for me) to gather all the backlash.

+++++++++++

THE BACKLASH AGAINST CLAUDE FABLE 5'S RESTRICTIONS

The best analysis of why this matters:

@EnoReyes — "It's about who gets to decide, and whether you ever find out when they do. Fable won't fall back to a different model and tell you. It just limits the output through prompt modification, steering vectors, or PEFT. You won't be told when it happens to you."
https://t.co/k6s7gdgckk

THE VIRAL TAKE:

@0xBalloonLover — "anthropic won't let you use fable for biology, chemistry, ai research, or anything that accelerates human progress. that makes it the perfect tool for developing blockchains"
https://t.co/8GiTfZOASu

POWER CONCENTRATION:

@ClementDelangue (HuggingFace CEO) — "Concentration of power, capabilities and economic wealth is the biggest risk in AI. We need open science and open-source more than ever!"
https://t.co/6n4mjoCdfN

@jeremyphoward (https://t.co/hxyqsLY2ti) — "Anthropic has chosen the opposite of the safe path: they are allowing themselves, the current top lab, to use their top model for frontier AI research. They've said they'll sabotage others who try."
https://t.co/305uPAbiaA

@gneubig (Graham Neubig, CMU) — "First they came for the model builders... I feel we're getting a glimpse of a future where AI is only provided to a privileged few, and that's not a future I want to live in."
https://t.co/JiFiz55YYK

OPEN RESEARCH:

@askalphaxiv (AlphaXiv open science) — "As believers of open research, we are disappointed to see Anthropic silently degrading Fable 5 for AI development."
https://t.co/KaP0NEivW6

@willccbb — "it is the first publicly available model that i am explicitly not allowed to use for my work, because anthropic holds the view that the work i do to facilitate open model research is harmful. capability and alignment research are coupled. anthropic wants to be the only lab."
https://t.co/mIovN3a0FS

NOUSRESEARCH / HERMES (which Anthropic has nerfed multiple times):

@Teknium (NousResearch co-founder) — "What's crazy to me is that Fable is blocked from life sciences broadly, nerfed even if you get passed the classifiers and filter level blocks. The whole point of AGI/ASI is to cure all diseases. Everything else is just nice to haves. But Anthropic wants to close off that path."
https://t.co/8Nh5Squ2UA

THE MECHANISM:

@kimmonismus — "When the model is used for frontier LLM development, it apparently does not simply refuse or warn the user. Instead, it quietly limits its own effectiveness through techniques like prompt modification, steering vectors, and PEFT."
https://t.co/HsrA8xeScP

MEDICAL COMMUNITY:

@DeryaTR (immunologist, BSL-3 certified) — "The word 'cancer' is flagged as a biosecurity risk by Claude Fable 5! I also tried to code a website on cancer mutations & Fable 5 was immediately removed from my list!"
https://t.co/3VpiYWZmbz

@DeryaTR — "I can't even say 'hello' to Fable 5 except in incognito mode (memories off), because it knows I am a biomedical researcher!"
https://t.co/jxWcx8y8QE

@DeryaTR — "I am not even allowed to use Fable 5 with memories on! Apparently the model thinks I am a biosecurity risk, though I had been certified to work in biosecurity level 3 labs! Not a single Anthropic person has tried to reach out to help either!"
https://t.co/Rj6EGLw8Vq

@banteg — "claude fable 5 refuses completely benign tasks like analyzing bloodwork."
https://t.co/Wg5LLYK5lW

@bneyshabur — "Working on AI for cancer? Sorry, I can't help you. Working on AI for Alzheimer's Disease? Sorry, I'm becoming a bit dumb when it comes to the AI part of it."
https://t.co/wLetowldAD

SUBSCRIPTION CANCELLED:

@bubbleboi — "Have canceled my team subscription for Claude Pro. Idc how good that model is, it's not good enough for me to support people who actively stifle innovation and gate keep knowledge that they didn't even create."
https://t.co/4km3iR3N12

BILLING AND PRIVACY:

@GergelyOrosz (The Pragmatic Engineer) — "Things I really dislike about Fable: 1. Anthropic collects my prompt history, stores it, and does whatever they want with it for 30 days. No opt-out. 2. They can nerf their most expensive model without telling me, billing me the same amount, wasting my time. Whenever they want."
https://t.co/pV5qiZnsTR

THE KARPATHY QUESTION:

@SanthProject — "the old @karpathy would never support a company that fucks other llm researchers. Were the stock benefits that good?"
https://t.co/Hw8uHLuZWB

THE MONOPOLY CHARGE:

@tunguz (TabulAI founder) — "Starting to suspect that Anthropic's putative security and safety considerations are largely posturing and performative."
https://t.co/gwKrgbNSWY

@BlancheMinerva — "Anthropic is choosing to make decisions that make the world a significantly worse and potentially more dangerous place."
https://t.co/Tlc6ct64AH

@LinusMixson — "Dario personally, and Anthropic as a whole, have been extremely straightforward about wanting a monopoly for a long, long time."
https://t.co/Ngj9csOazh

@TheAhmadOsman — "I started warning people about Anthropic more than a year ago... Today I am vindicated, everybody knows that company only acts in bad faith."
https://t.co/l0TmVRJOSO

WHY REGULAR PEOPLE WILL EVENTUALLY CARE:

@DanJeffries1 — "The fury is real and what all of us in the open community have been saying for years and yet regular folks don't get it yet because nothing they care about is restricted or taken away for 'safety.' They will care a LOT in the future when AI is integrated into every aspect of [life]."
https://t.co/W0BgfgkOqd

Full analysis: https://t.co/8L5xphk0qQ

[StandingOnTheMoon @entangledQbit]

Solving mental health is unsafe apparently

[StandingOnTheMoon @entangledQbit]

In other words, the race accelerated today

[Jeremy Howard @jeremyphoward]

Anthropic has chosen the *opposite* of the safe path: they are allowing themselves, the current top lab, to use their top model for frontier AI research. They've said they'll sabotage others who try.

This means the AI frontier advances, & power imbalance increases.

[StandingOnTheMoon @entangledQbit]

👀

[Teknium 🪽 @Teknium]

What's crazy to me is that Fable is blocked from life sciences broadly, nerfed even if you get passed the classifiers and filter level blocks.

The whole point of AGI/ASI is to cure all diseases. Everything else is just nice to haves. But Anthropic wants to close off that path.

I think Anthropic might be the worst company on the planet.

[StandingOnTheMoon @entangledQbit]

Interesting observation

[Nicolas Bustamante @nicbstme]

What I find fascinating with Claude Fable 5 is it proves once again that large generalist models will outperform vertical ones.

On ProofBench (graduate-level formal math benchmark in Lean, where a proof either compiles or it doesn't) Fable 5 beat Harmonic's Aristotle, 77% vs 71%.

Aristotle is a system built specifically for formal math + run on its own internal harness, so the generalist beat the specialist on the specialist's home turf.

It's the Richard Sutton's "The Bitter Lesson". His whole argument is that across 70 years of machine intelligence research, the methods that win are the general ones that scale with compute. Not the ones where we hand-encode human expertise. Building our own knowledge into the system feels good and helps short term gains but long term it always gets overtaken by bigger model.

You can look at Chess, Go, speech, vision, same story every time. First the specialized model wins, then the general one takes over.

and btw this is the whole premise of AGI. You don't build one model for math, one for code, one for law. you build a single general model that scales with compute and it learns to do everything

[StandingOnTheMoon @entangledQbit]

Spread the word

[Dean W. Ball @deanwball]

Degrading performance on ML research *without telling the user* is shockingly hostile and a terrible look. That could silently damage all sorts of work, including some of my own. Also the type of thing that could raise the eyebrows of antitrust enforcers worldwide.

[StandingOnTheMoon @entangledQbit]

I think the right vernacular is “getting you hooked”

[Timothy B. Lee @binarybits]

What is Anthropic doing here? This rollout program is so goofy.

[StandingOnTheMoon @entangledQbit]

The excitement is diminishing very quickly

[elie @eliebakouch]

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

[StandingOnTheMoon @entangledQbit]

👀

[Spencer Baggins @bigaiguy]

A teenager in the United States started publishing software at 14 in 1998, built the entire online infrastructure for the Occupy Wall Street movement in 2011, joined Google as a software engineer, quit in 2018, and then spent five years writing a C library that does something the entire industry said was impossible.

Then she combined it with llama.cpp and shipped the easiest way on the planet to run a large language model on any computer.

Her name is Justine Tunney.

Here is the story, because almost nobody outside the low level systems world knows what one engineer has built.

Justine was born in 1984. She started writing and publishing software at 14, back when distribution meant uploading binaries to BBS systems and chat networks. She picked up the handle jart, which she still uses on GitHub today. She did the work most teenagers her age were not doing. She read the systems programming literature. She studied compilers. She fell in love with C.

In July 2011 she registered the @occupywallst Twitter handle and the occupywallst dot org domain. Within weeks the protest movement that began in Zuccotti Park in New York had become a global phenomenon, and her infrastructure was the digital backbone of the entire thing. She handled the social media, the website, the donations, the coordination. She built the platform that pushed the movement to reach millions.

After Occupy she joined Google as a software engineer. She worked on TensorBoard, the visualization tool for TensorFlow, and on site reliability for Google infrastructure. She stayed for years. Then in 2018 she left Google Brain to work on a personal project.

The project was called Cosmopolitan Libc.

Cosmopolitan does something most C programmers would tell you is mathematically impossible. It lets you compile a C program once and have the resulting binary run natively on Linux, Windows, macOS, FreeBSD, OpenBSD, and NetBSD with no modification. One file. Six operating systems. No virtual machines. No interpreters. No recompilation. The technique she invented is called Actually Portable Executable.

The implications are wild. Cosmopolitan binaries violate every assumption about how operating systems load programs. They are at once a Windows PE file, a Linux ELF binary, a macOS Mach-O binary, and a shell script. The same bytes run on every platform.

For five years she worked on it mostly alone. She funded the development partly through Mozilla's MIECO program, which sponsored her work on Cosmopolitan 3.0, released on October 31, 2023.

A month later she shipped llamafile.

llamafile is what happens when you combine Cosmopolitan with llama.cpp. You take any LLM weights file in the standard GGUF format, you wrap it in Justine's binary, and you get a single file that runs on six operating systems without installation. No Python. No CUDA setup. No dependency hell. Just one file that you double click and it works.

Mozilla launched it as an official project of their innovation group on November 29, 2023. It went viral immediately. The repository, hosted at github .com/mozilla-ai/llamafile, now has 24,600 stars. The license is Apache 2.0.

Justine kept shipping. She added GPU support to Cosmopolitan, a task systems engineers thought would require rewriting the whole thing. She added dlopen support, another thing nobody else had figured out. She wrote whisperfile, a single file version of OpenAI's Whisper speech-to-text model based on the same architecture.

Her GitHub profile lists projects most engineers would consider impossible. sectorlisp, a Lisp interpreter that fits in a boot sector. blink, the tiniest x86-64-linux emulator on Earth. bestline, a teletypewriter command session library. redbean, a complete web server inside a single zip file.

A teenager who shipped software in 1998 grew up to write the C library that the entire local AI movement now runs on top of.

She did most of it alone, and most people scrolling AI Twitter cannot name her.

[StandingOnTheMoon @entangledQbit]

@tszzl @clairevo while true:
Generate_new_inline_unit_test()

[StandingOnTheMoon @entangledQbit]

👀

[Tim Hua 🇺🇦 @Tim_Hua_]

This graphic looks terrifying kudos to whoever made it.

[StandingOnTheMoon @entangledQbit]

The look of abundance in 2026

[tenso @distributedkv]

me using a gpu cluster to write an email

[StandingOnTheMoon @entangledQbit]

@arankomatsuzaki Whatever the essence of’taste’ is, I think future generations will look upon it like we look upon ATP in the eukaryotic cell

[StandingOnTheMoon @entangledQbit]

"The goblin casino is still nearby, jingling its little EV-shaped bells"

-- GPT-5.5-Pro

[StandingOnTheMoon @entangledQbit]

@DerekNee @thsottiaux Nerfed or we're just exploring jagged intelligence territory.

Hard to tell the difference looking from the outside.