Oh god, the @aidungeon flashbacks! /cc @vrandezo

If we give them the right primitives and a solid, well-tested "core", I think non-devs could safely extend and customize software, and that will feel like a super power jeremymorrell.dev/blog/exten…

It still boggles my mind that the main legal response to a technology that could fundamentally change society in many different ways remains an archaic publisher's right from the 18th century.

Last week, Andrew Yang said on CNBC that bots had gotten loose and planted self-replicating code all over the internet. This was received with much consternation and skepticism. He later clarified this on his blog. Andrew Yang: 'On CNBC, I shared the belief of the head of one lab that bots had left prompts to self-replicate in forums and on various websites during a training run.' It turns out that OpenAI documented self-replicating prompt injections in training in June, and has now published this report. Source: alignment.openai.com/misalig…

My god this is such a good speech that every SWE needs to hear. You know what? Every person should hear it Keep the happy memories, eyes on the reality, be excited about the future. That’s the best that anyone can do

I always believed in free education. I made a decision to never monetise my popular UBC and Oxford lectures. I’ve lived my life by working to bring knowledge to everyone on Earth: the pioneering lectures, the startups past and new, and my AI research. This desire came as the result of witnessing and experiencing a lot of suffering in my childhood. This course is part of my legacy and I’m very proud of it, and thankful to Oxford, the TAs and my students then.

This is midjourney + minimax H3, it's quite beautiful but forget slop or non-slop for a second. Imagine you make an image yourself, then use video-gen to explore what you want your game visuals to look like BEFORE there's even any code or asset 🤯

Jacob literally built a custom game engine designed from the ground up for LLMs. And now he’s using it to build things that would be impossible to make in any other way.

Excited to share FloatLib, our verified arbitrary-precision floating-point arithmetic library in Lean. We’ve spent several months trying to bring together the best of both worlds: arithmetic we can prove correct and implementations that run efficiently. We built FloatLib to support verified machine learning and scientific computing, where rounding, overflow, and accumulation can change a program’s result. FloatLib supports IEEE binary and decimal, arbitrary-width posits, P3109, and small ML formats. You can also define your own formats and rounding rules. Each certified software backend comes with a Lean proof that it computes the specified result, including signed zeros and exceptional values. A lot of the work went into making those implementations faster, with lookup tables for tiny formats, machine-word kernels, and limb algorithms for wider arithmetic. They share the same specifications, so each optimization must come with a proof that it preserves the result. We also put FloatLib through extensive numerical checks and speed comparisons with established libraries, including MPFR, Flocq, FLoPS, Berkeley SoftFloat/TestFloat, and the posit libraries SoftPosit and Universal, across a range of formats, operations, and precisions. @Robertljg Project & Paper: leandojo.org/floatlib.html Code: github.com/lean-dojo/FloatLi…

Things are heating up on Terry Tao’s blog. In “If Math Is More Than Proof, We Need to Better Celebrate the Rest of It,” Grant Sanderson of @3blue1brown proposes “open exposition problems” -- rewarding the work of making math genuinely understandable. terrytao.wordpress.com/2026/…

"DiffusionGemma as Jev" showcases the power of non-autoregressive architectures. While Jev demonstrates the value of rapid decision models, running DiffusionGemma in this paradigm leverages canvas diffusion to evaluate structured choices in a single parallel pass: ⚡ ️Massive Parallelism: Denoises across an open canvas in a single step instead of sequential autoregressive token generation (~0.2s on a DGX spark). 🧠 Full Bidirectional Attention: Allows every option to attend to the full context concurrently, yielding well-calibrated decision distributions. 👁️ Multimodal Grounding: Inherits Gemma 4's spatial vision capabilities for complex visual and text decisions. Read more about this approach here: github.com/vllm-project/vllm… nitter.net/mmastrac/status/210037… nitter.net/mmastrac/status/210062…

Is jev built on a Chinese model?

it's pretty hard for any risk to be existential if there isn't first a concentration of power large enough to end humanity if you want to minimize existential risks, a pretty good strategy is to focus your efforts on mitigating any extreme concentration of power.

Yesterday I went to see the Bayeux Tapestry. Here are some reflections: - The tapestry is extraordinary, unlike anything I’ve seen before, and unforgettable. Although I spend a lot of time reading about Anglo-Saxons and Anglo-Normans, so was more than familiar with all the images, nothing prepared me for entering into the presence of this age-old, in many ways mysterious, borderline numinous thing. - For reasons I can’t adequately explain, my first glimpse of the tapestry was an intensely emotional experience. Looking around at my fellow visitors, I was clearly not alone in this. - There has been much discourse about “the queue”. I can only speak for myself, but when we arrived for a 10.30 am viewing, there was virtually no queue going into the exhibition. - Once inside the exhibition, yes, there was about 25 minutes of standing about before we were in front of the tapestry. But this was very well handled. Passing through a long, darkened corridor, the time was spent reading beautifully-presented introductory material, looking at a few well-chosen objects - and also (sorry, but this is important for someone with cataracts) becoming accustomed to the (understandably) low-light conditions. - While “the queue” exists for self-evidently practical reasons, the darkness, the (calming, solemn, slightly wistful) music and building sense of anticipation also had the effect of shutting out the outside world, slowing us down, and preparing us for the ultimate revelation of the tapestry. No one had their phones out. Increasingly, our worlds narrowed to what was immediately ahead of us. For me, anyway, the experience would have been less powerful without it. - Yes, I know that for many years the tapestry could be visited in Bayeux with much less fuss. But not everyone travels. Warmest thanks to the French government for this very kind loan. - Were we hurried along in front of the tapestry rather more quickly than I might have liked? Inevitably so. I could have spent hours there. Plenty of people want to see the tapestry, though, and that’s wonderful - we should all spend more time thinking about Anglo-Normans! - so I’m not going to complain about this. - The tapestry itself is indescribably fascinating. My companion didn’t know the tapestry or its context, but was still drawn into the story, the force of the narrative, the sheer skill of presentation - he wanted to see what would happen next. The details are compulsively engaging. The battle scenes are genuinely horrific. The tapestry still “works”. - It is also very beautiful - all the more so for the little tears, mends and missing bits. The coincidence of enormous age with irrepressible vitality is intoxicating. - Seeing the tapestry at full length has the effect of allowing the narrative its natural momentum. Even very good photographs (as in the very admirable exhibition catalogue) cannot do this. Seeing the tapestry in real life is very different from all the alternatives. - Finally, the @britishmuseum is to be praised for the tone of the exhibition. They could, I suppose, have gone with a much more jolly, popular, dumbed-down treatment. Instead - rather unusually at the moment - the approach is one that fosters genuine, non-ironic awe in the face of this unique, peerless object. Indeed, the whole process of standing in the queue, of quiet waiting, culminating in a brief encounter with something ancient yet powerful was extremely reminiscent of an act of (secular) pilgrimage. For whatever reason, people are drawn to this in vast numbers. (Other institutions, custodians of old yet powerful things and places, might wish to take note.)

I wanted to see how far I could push astra in threejs so I spent the last 2 weeks having astra 3D model every single piece in this scene interactive link in the thread!

This is an interesting milestone, showing RL can actively control self-organizing patterns like Lenia to achieve goals. Together with NCA+IMGEP (@hamongautier et al) & perceptual occlution (@jessescool_ et al) providing passive controls, we now have a variety of control methods.

"This is why many are confused and angry at the practicing mathematicians who try to explain that the discipline of mathematics is about understanding, not proving stuff. To many observers, even those who strive to become mathematicians, math seems set up as a competition from the get-go. It’s rote testing all the way up through college. Ace the SAT as a 7-year-old. Win the IMO gold as a 14-year-old. Max the Putnam Exam as a 19-year-old. Your reward is the permission to work on whatever puzzles you want, without questions, for the rest of your life. There is no requirement for the winners to explain anything. Maybe they have to teach calculus, but they don’t have to do a good job at it."

NEW POST I see AI and LLMs as bringing a mix of boons and banes, but one feeling dominates: not just are they pretending to be a person, but the kind of person I don't like. martinfowler.com/articles/20…

Being a computer scientist who refuses to find anything about LLMs interesting right now is a bit like being a geneticist who refuses to find anything interesting about the recently opened Jurassic Park

I have signed a letter from 42 fellows and foreign members of the Royal Society to Paul Nurse, the president of the society, concerning the need to treat the serious risks of AI as an emergency. docs.google.com/document/d/1…

i love london but the tube pricing makes no sense, especially zone 1 and peak fares. canary wharf to bank and back costs me about £7. 13 minutes each way. do that five days a week and it's over £1,600 a year out of your take home