Live footage of a successful paper review rebuttal 😅

There is no AGI until they can cook this.

12 months of progress on the jagged frontier: 1) AI got spikier, not rounder. 2) Maths, fluid reasoning, expert work: past human level. 3) There is almost no middle ground. Once a capability crosses human level, it often quickly becomes 100x+ cheaper than a human. 4) Cost keeps decreasing after capability increases stop. Language and knowledge look saturated but price keeps falling. 5) The gaps are not a budget issue. Human-level memory, self-monitoring and long-run planning are not for sale at any price. AI is now a superhuman specialist that is sub-human at running itself.

What a nice finding.

A general principle: when you interact with AIs, you should know that you are interacting with AIs. Tech like this that helps disguise AI makes us all worse off. Working on camouflaging AI use should be viewed as unethical. This doesn't help anyone.

water is transparent only within a very narrow band of the electromagnetic spectrum, so living organisms evolved sensitivity to that band, and that's what we now call "visible light".

Time is scarce. Focus scarcer.

“Beacon of excellence” Always preying on the narcissists.

Papers should evolve to have 6 pages written for Human Readers, and unlimited appendix space with all details needed for Agent Readers.

Third question on AI. A question that also remains unasked is whether the AI boom can continue without leading to a massive increase in inequality. A recent paper by Stijn Van Nieuwerburgh runs the numbers on how much revenue the AI industry needs to generate to recover its massive investment (summary and a link to the paper can be found here: brookings.edu/articles/finan…). Van Nieuwerburgh’s arithmetic should make us more concerned. AI investments will average about 3.6% of GDP annually between 2025 and 2032. Van Nieuwerburgh calculates that, using a 10% rate of return, the industry would need to generate annual revenues of about $3.7 trillion by 2032 to recover these costs (growing from its current levels of about $200 billion or so). That is significantly more than 10% of current US national income, and will likely remain around 10% of national income by 2032, even if GDP growth rose from its current level. A large fraction of this revenue will go to capital income. That means a massive increase in the share of capital in national income, which has already risen substantially over the last 25 years or so – now standing at an all-time high of about 47% (bls.gov/news.release/pdf/pro…). Capital income is much more unequally distributed than labor income, so a massive increase in the capital share of national income will translate into a very sizable surge in inequality. The rise in inequality may not stop with the capital share. My work with Pascual Restrepo documents that (automation-driven) increases in the capital share of national income are typically associated with rising labor income inequality as well (see, for example, economics.mit.edu/sites/defa…). The same may happen in the next several years, boosting inequality further. What is missing from our current debate is any discussion of a fundamental dilemma these numbers pose: can the AI boom avoid both an economically costly crash and a huge increase in inequality? If the industry reaches these revenues, inequality surges. If the industry does not become profitable, a crash, with substantial costs in terms of lost output and jobs, becomes likely. My assessment would be that the industry is unlikely to reach levels of revenue Van Nieuwerburgh calculates. First, diffusion has been and will likely continue to be slow. Second, competition from open-weight models, which are getting better, will limit how much proprietary models can charge. Third, despite important advances, I still believe that AI models will not be able to automate entire occupations anytime soon, thus limiting their value to businesses as cost-saving devices. Whether this leads to a crash or not is more complicated and will depend on whether various AI companies are bailed out and what kind of support they receive. Nevertheless, even if revenues fall short of these gargantuan amounts and we avoid a dramatic surge in inequality, I expect that the diffusion of AI will push up inequality between capital and labor and within labor. If inequality does surge, a further question becomes central: can our democracy survive such astronomical levels of inequality?

This is the most interesting recent development in bitcoin

Using arXiv publication trends, we could estimate when AI might provide as much uplift to a given field as it does to mathematics today. Quantum physics could reach that point within a year, and physics as a whole by late 2029. Biology, of course, hasn’t budged yet.

For folks not understanding where we are at with AI now - if the task is verifiable, it is now solved by AI. Yes, earlier days you could argue that 'AI isn't that good' (e.g. see first attached chart here, where GPT-4o underperformed average accountants on tasks) But now, it's a totally different story (only ~2 years later). The second chart is astounding. It took me a while to even understand it because it looks so odd. It compares manual accounting tasks to tasks complete with Opus 5.5. Basically Opus 5.5. solves all tasks almost instantly at a 100% accuracy rate, whereas a person takes a lot more time and gets a lot of things wrong. It makes the chart look entirely silly because the axes aren't even comparable. So yeah, this is where we're at. If it is verifiable, it is solved. This is just how these model architectures work now.

We'll be spending a lot more time trying to understand the outputs of language models. A few thoughts, tips & tricks: Writing. Something I've had success with: Ask your LLM to explain something in ASD-STE100, it's a controlled language specification originally developed for aerospace maintenance documentation. LLMs well-versed in this language and it comes with heavy constraints on clean writing style that I often find a lot more readable. Sometimes I've tried to soften it a bit e.g. ask for "80% of the way to ASD-STE100" because the spec is quite stringent. But even better: Diagrams / images. Instead of writing, ask your LLM to create a diagram. These can be a lot easier to process, parse, and understand. But even better: Web pages. Ask for output "in HTML" to get a beautiful, interactive webpage. LLMs are getting really good at frontend and can create beautiful experiences, animations, etc. But even better: Explainer videos. The output format I am most bullish on is fully custom / bespoke explainer videos generated on any arbitrary topic. Experiment with things like "Create a 3b1b style video explainer on X. Use my ElevenLabs API key for audio narration". (you'd need an API key for the latter or you can ask your LLM to find you decent free alternatives that use your local compute). This is actually starting to work! In summary: - As LLMs get better, they will do more and more of the legwork autonomously, and a lot more of our work will rise up the abstractions into oversight and understanding. - Luckily, LLMs can help here too because as intelligence and code are increasingly abundant, you can ask for large, custom, discardable software artifacts (e.g. web apps, video explainers) that would have never made sense to create before. Push the boundaries here and you'll be surprised.

Many people have opined that AI is ruining open source because people fling half baked slop patches at a higher volume than maintainers can handle. Actually the opposite is true, it’s going to make open source vastly better. Most open source projects have been dramatically underfunded. Projects with massive usage have hardly any contributors. They have had to scrape by on a skeleton crew of part time people that can’t be counted on to finish things. I think very small teams are going to be able to run very large code bases with much higher quality now. They won’t want your patch, they’ll want really precise descriptions of the problem. Given that the patch will be easy enough. The right model isn’t some diffuse committee of people mostly not paying attention, it’s 1-3 high context people with high commitment and great taste who care about the project. AI is fundamentally commoditizing software, and open source has historically been the end point of commoditization. There is going to be a lot of awesome open source coming our way.

RIP all-remote video interview hiring processes, hello onsite rounds It was a matter of time, wasn't it?

In other words, for the target audience: A five hour limit of 0.0137 papers, with a weekly limit of 0.46 papers.

arXiv has updated our policy on rate limiting for all submitters. This update was made to fairly distribute moderator time & support the arXiv community of staff, volunteers, readers & authors. Please read our announcement to learn more: blog.arxiv.org/2026/10/01/up…