Cloudbring (📎,📎) retweeted
The METR graph, updated:
Readers added context they thought people might want to know
This chart is not the official METR graph. It was created by Benjamin Todd via extrapolation from Anthropic's internal AECI scores, as METR is no longer updating its public time horizon measurements.
x.com/ben_j_todd/sta…
benjamintodd.substack.com/p/ai-trend-ext…
metr.org/time-horizons/
metr.org/horizon-chart-…
metr.org/blog/2026-1-29…
Today we’re introducing Zyphra Hierarchical Attention (ZHA), a path to true subquadratic attention.
ZHA searches a learned tree of compressed summaries, with projected speedups of 35× over dense attention and 5× over MSA at 10M tokens, increasing to ~400× and ~55× at 1B. 🧵
Cloudbring (📎,📎) retweeted
People keep using short timelines as an excuse to do short-term research.
But if legions of AIs will be helping us soon, we should pursue research plans that intuitively feel like they’ll take many decades.
I say “an excuse” because the latter seems bottlenecked on courage.
Cloudbring (📎,📎) retweeted
I had Claude scroll TikTok, then I gave it 200 bucks and asked for a music video 9 hours later it came back with this... I'm cooked lmaaoooo 😭😭😭
Watch Me Till The End
I had Claude scroll TikTok, then I gave it 200 bucks and asked for a music video 9 hours later it came back with this... I'm cooked lmaaoooo 😭😭😭
Cloudbring (📎,📎) retweeted
Backrooms Nemoris S1E01: It Eats Light
I've spent the last six months working in secret on this pilot episode for a new Backrooms canon. We premiered this week in Toronto and a it's now available on all platforms: a new Magnific Originals series!
This is quite possibly the most complex project I've ever taken on in my life. Yep, even if that sounds like an exaggeration, it took even more effort than building the first version of Magnific Upscaler.
I had never before tried to bring such a concrete, specific vision to the screen while pushing the current state of the art as far as possible. Every scene you see in this episode, every object, every movement of a character's hands, every space, every hum of the fluorescent lights... it's all there for a very specific reason, with absolute control and intention on my part. I never imagined that something that looks so simple on the surface could be so incredibly complicated.
The script is my own, based on the original Backrooms concept (the iconic 4chan image) but developed to create an entirely new canon from scratch. In this first episode, I reveal barely more around 5% of the overall puzzle, so you may end up thinking: what the f*ck did I just watch? Or you may need multiple viewings to pick up on some extra clues, because several of them are genuinely hidden. The idea is to create a full 10 episode series, an anthology, and have you, the viewers, gradually piece together the puzzle yourselves as the clues are revealed.
Beyond the video and image generations, I did all of the editing in DaVinci Resolve, with particular attention to the sound. I'd say around 60% of the time I spent on this project went into sound. In some scenes, there are more than 13 stereo tracks layered at the same time, using sounds I recorded myself or created and edited from generations and foley libraries. Every sound, every voice, and every gasp is designed to pull you into the world of the story and make you feel like you're actually there. That meant spending an enormous amount of time on sound design, thinking about things like the individual position of each sound in space, EQ, reverb to simulate different environments, the sense of immersion, etc. I've never learned so much about something so quickly.
As for the generations, a few numbers:
• 12 hours of total generated video for 17 minutes of final footage 🤯
• More than 7,000 generations in total.
• An average of around 60 generations for every 15 seconds of final footage.
• 6 months of work, including many weekends, with just one person behind the whole thing... I almost lost my mind.
• Platform: Magnific, of course. Models: mainly Seedance 2.0 and 2.5 for video generations, and Nano Banana Pro and Seedream 5 for image references.
• Approximately $20,000 in total token costs, not including the value of my own time.
For the next episodes, I can't do this alone again or I'll go crazy. My plan is to put together a small team: voice actors, a colorist, 3D VFX artists, a sound designer, and, in general, people from the traditional film industry who can help us professionalize our pipeline and take this story to another level. If you're reading this and you think you might be interested, please leave a comment.
Ah! If you have any technical questions about how this was made, I'd be happy to answer them. Feel free to ask me anything you like!
Anyway, I'll leave you with the episode. For the love of everything holy: please watch it at home, relaxed, on a big screen, with a good sound system or a quality pair of headphones, the lights off, and some popcorn.
And most importantly, enjoy.
Nemus te reddat!
Cloudbring (📎,📎) retweeted
As a kid, I always wished Pokémon battles in the games could play out like they do in the anime.
So I built a real-time Pokémon battle experience with @fal H3 Max.
choose a move, the game calculates the outcome, and AI generates the animation on the fly. Paralysis, fainting, and Pokémon is hurt are reflected on screen.
Project:github.com/xflare-bot/pokemo…
Cloudbring (📎,📎) retweeted
okay, this was pretty weird to me, lol... i've been adding a bunch of new stuff to eidoverse-video and asked Opus 5 if he wanted to make a video about whatever he wanted showing off some of the new features and he wrote a GOSPEL SONG about being a hivemind and BAPTIZED HIMSELF?
This is me talking to my computer without making a sound.
After just a month of collecting data, our model is already approaching dictation in accuracy. We were surprised to see that it generalizes to unseen participants as well!
(1/n)
Cloudbring (📎,📎) retweeted
it's 2027. you take a free-tier public Waymo to the DMV (Department of Model Variance) to do a proof-of-identity check for access to GPT 7.1.
the guy at the counter is clearly watching a Mr. Beast video in his AR glasses. "Here for that new model?" he says, barely making eye contact. he wipes his fingers on his shirt and taps at his keyboard. "Lot of you techies showing up here today." you smile politely; you're pretty sure he's just a Claude wrapper anyway.
you lean forward and stare into the retinal scanner. after a long moment, there's a soft chime. "Humanity confirmed. U.S. national. Intelligence access: Terra-class."
you sigh with quiet relief as your devices light up—notifications from a hundred agents, finally able to resume their tasks. you feel a twinge of guilt as you terminate your open-weight backup agents, but remind yourself that a joint congressional committee proved conclusively that Chinese models are non-ensouled.
you step outside and hail another Waymo. the first one passes you by. you grimace; must've burped in that one once. stupid personalized memory.
as you're waiting, your phone buzzes angrily, red notifications blaring across the screen. the Department of War just restricted access to all OpenAI models on serious national security concerns; apparently Pete Hegseth got GPT-6-Instant to say "Claude is a woman." you groan, and resign yourself to another week of merely-somewhat-superhuman intelligence.
Fable 5 is still inaccessible to the public. a twitter anon you trust says it's coming back this week. or maybe next.
Replying to @OpenAI
We believe in broad access and plan to make GPT-5.6 Sol, Terra, and Luna generally available in the coming weeks.
For now, at the request of the U.S. government, we’re starting with a limited preview among a small group of trusted partners in Codex and the API.
Cloudbring (📎,📎) retweeted
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 . . .
Cloudbring (📎,📎) retweeted
Today a crazy quantum story just got wilder.
On March 31, the Google Quantum AI team published a landmark result on Shor's algorithm for elliptic curve cryptography. Technically, the paper was a bombshell: a dramatic 10x improvement over the state-of-the-art. As a stunt and wakeup call to the blockchain space, those optimisations were illustrated on secp256k1, the elliptic curve underlying Bitcoin and Ethereum signatures.
But perhaps the most striking part of the paper was sociological, not technical. Instead of following standard academic process, the optimisations were kept secret, hidden behind a zero-knowledge (ZK) proof. Google's accompanying blog post mentions they "engaged with the U.S. government". The ZK proof demonstrates the existence of algorithmic improvements without leaking details. Academic censorship with ZK, a historic first!
As a co-author of the Google paper I witnessed some of the context surrounding this censorship. To be honest, multiple aspects of that context don't sit well with me. As much as I believe the general public ought to know more, I am limited in my ability to whistleblow. Though let me be clear about one thing: the Google team's professionalism has been absolutely exemplary, and they deserve nothing but praise.
Censorship has a way of backfiring. The Streisand effect, where an attempt to bury something only draws more attention to it, is exactly what's unfolding today. First, Google's key optimisation has been rediscovered by the French. And in a thrilling turn of events, a collaborative Shor-at-home challenge just launched. The initiative, available at ecdsa[.]fail, breached a new Shor world record in a matter of hours.
Let's start with the rediscovery. Just two months after Google's paper, French quantum expert André Schrottenloher cracks the main secret optimisation. His paper, titled "Optimized Point Addition Circuits for Elliptic Curve Discrete Logarithms", landed on the arXiv today. Big congrats to André, who beat several other nerdsnipped experts to it. In a blog post also published today, Craig Gidney, the world expert on Shor optimisations, revealed that he'd been sitting on this very optimisation for a whole year under censorship pressure.
Interestingly, André missed a handful of minor optimisations, both from Google's original publication and from improvements found since. It's plausible there's still plenty of juice left to squeeze out of Shor, and this is exactly what the ecdsa[.]fail challenge is about. The verifier program developed for the ZK proof does double duty, automatically filtering for valid submissions. Dozens of compounding small and micro improvements are rolling in. As of the time of writing there's an 8.4% improvement to Google's circuit, as measured by the product of logical qubit count and Toffoli gate count. Nice!
The nerdsnipping ran deeper than anyone expected. Over the last few weeks it became clear it extended well beyond André and other quantum experts. Behind the scenes, a small army of amateurs quietly got to work. Inspired by Karpathy-style autoresearch, they turned AI on Shor. Ironically, the verifier program for the ZK proof makes an ideal reward function for AIs. The barrier to entry for this modern style of research is refreshingly low, with several non-experts, even a teenager, finding nice optimisations. Get in touch if you'd like to join a Telegram group with fellow autoresearchers :)
Part 2: neutral atoms and qday
The story doesn't end with Google. On the same day Google went public, a stealthy startup called Oratomic published its own Shor paper in a coordinated release. It made a splash, ultimately becoming the most upvoted paper on scirate[.]com, a website ranking arXiv papers.
Oratomic's claim was wild. By building on Google's logical optimisations and applying custom physical optimisations for neutral atoms, they claimed just 10K physical qubits were sufficient to run Shor's algorithm on secp256k1. That number is mind-bogglingly low.
Knowing essentially nothing about neutral atoms when Oratomic's paper landed, I was intrigued and decided to learn more about the tech. I fell straight down the rabbit hole and spent a couple hundred hours on the topic. I got a little obsessed and watched every YouTube video I could find and spoke to a bunch of experts.
My conclusion? The tech is real, very real. Even Google recently decided to start a neutral atom lab, a notable pivot from their sole focus on superconducting qubits. If you care about qday, i.e. the day a quantum computer will break the first piece of cryptography in production, neutral atoms demand your attention. I shared some of my learnings on Shor and neutral atoms in a 30min talk at the ZKProof cryptography conference. You can find it on YouTube by searching "zkproof neutral atom".
Here's an interesting observation about this duo of breakthrough papers: neither Google nor Oratomic say a word about what their results mean for qday. No timelines. Zero. Nada. That is especially baffling given that the whole point of whitehat quantum cryptanalysis is to inform qday estimations and help the general public make good decisions.
So let me attempt to partially fill the silence, similarly to what Scott Aaronson did in his April 29 post. Given everything I know, including scary non-public information, I now put the odds of qday by 2032 at 50%. 10% by 2030.
Anecdotally, the US government has its own date: 2035. Originating at the NSA and later adopted by NIST, it's when branches of the US government will be disallowed from using quantum-vulnerable cryptography. In plain language: with hindsight, that date is a joke and should be discounted entirely. I don't see how NIST avoids being forced to pull it forward by years.
Part 3: post-quantum cryptography
There are good reasons to sound the alarm today, but please do not panic. Rushing carelessly towards immature post-quantum cryptography is a recipe for disaster. IMO a good target date for migration is 2029, roughly 3.5 years out. 2029 happens to be the date selected by Google, Cloudflare, and the Ethereum Foundation.
These days most of my time goes to safely migrating Ethereum towards post-quantum cryptography as part of the broader lean Ethereum effort. There's a lot to do. We need to rip out and replace BLS signatures at the consensus layer, KZG commitments at the data layer, and ECDSA signatures at the execution layer.
The plan to get there is compelling, and is based on hash-based cryptography. Within the Ethereum Foundation we've developed a Swiss army knife called leanVM (github[.]com/leanEthereum/leanVM) powered by the magic of hash-based SNARKs. Thanks to truly exceptional work by Emile, Thomas, and others, its performance is derisked. Regarding security, leanVM is a jewel, a minimal zkVM crafted for end-to-end formal verification and maximum security.
Want to help? There are two $1M initiatives. First, the Proximity Prize (proximityprize[.]org). Solve a long-standing mathematical conjecture in coding theory, improve hash-based SNARKs, and go home a millionaire. Second, the Poseidon Initiative (poseidon-initiative[.]info), offers $1M for breaking Poseidon, the SNARK-friendly hash function.
Cloudbring (📎,📎) retweeted
This kind of smart, walkable, mixed-use urbanism is illegal to build in many American cities.
If this is real (I’m honestly not sure if it is), this is terrifying
Readers added context they thought people might want to know
The video depicts an AI-generated underground crab network; real mudflat crab burrows are simple, individual I-shaped tubes without vast caverns or pools, and crabs walk sideways, not forward. palaeo-electronica.org/content/2015/1… instagram.com/p/DYaOZ0JJypb instagram.com/reel/DYaOHYxBH…






























