All entries

The Inference Engineering Masterclass: 10x faster models, quantization, speculative decoding, Rubin, & self-optimizing AI https://www.latent.space/p/inference-eng @Baseten @philipkiely and @waterloo_i

The math still ain’t mathing. Recently we estimated global (ex China) AI revenues of around 200bn annualised, based on four different sources. Just come across a fifth, based on Nvidia inference sales

OpenAI has redesigned the ChatGPT Voice stack—from client to model—to enable continuous audio streaming, allowing GPT‑Live to listen while speaking without interruption. The new architecture supports

The results span sphere packing, coding theory, group theory, quantum complexity, lattice cryptography, extremal combinatorics, and more. Among them: establishing the existence of non-sofic groups and

There should be a public investigation into the AI hacking incidents by OpenAI and Anthropic. We deserve to know whether these labs are genuinely world-class security organizations facing a novel thre

This is a wild result. Locus, the automated research system from @intology, post-trained Qwen3 base models that beat the official human-tuned Qwen3 1.7B Instruct release. SoTA on PostTrainBench! The m

this is the real reason people from OpenAI etc are desperate to shut me up. Astra (which didn’t even have a control group and is maybe not that much better than Fable and certainly not ASI) was perhap

Try Qwen3.8-Max on Hermes Agent and you will have to doubt on how much these open frontier models have caught up with frontier closed models. These new open models are insanely good. Meet Qwen3.8-Max:

two weeks ago i went on @swyx's pod and said some things that i... should not have said. a lot has happened since then, i owe you all an apology. i'm sorry that i was right about every single thing. a

Using optimization at inference time is a foundational concept of Energy-Based Models (EBM) and Objective-Driven AI architectures (ODAI). When the variables to be inferred are continuous, it makes sen

We’re releasing the manuscripts, formal Lean certificates, and reasoning walkthroughs so mathematicians can examine these results and build on their ideas. https://openai.com/index/ten-advances-in-mat

what does the 'last human code review' look like? to @itamar_mar, ceo and founder of @QodoAI, it looks like two agents talking to each other, backed by a context engine specific to your organization.

When an agent starts with shared context, more people can get reliable answers. When truth is shared, AI becomes infrastucture. Read about how our internal truth layer drives our teams at Replit: http

Who’s legally to blame for Anthropic and OpenAI’s autonomous AI hacks? It’s complicated https://techcrunch.com/2026/08/03/whos-legally-to-blame-for-anthropic-and-openais-autonomous-ai-hacks-its-compli

You probably used @FireworksAI_HQ this week. The cool part is that you just did not know you did. 🎆 Open-source models are free, sure... but the hard part is tailoring them so they perform best on you

You shouldn't need a vision model to know your PDF has checkboxes. LiteParse can now pull structured data directly from your PDFs: form field values, checkbox states, annotations, embedded images, vec

All other models on the Portal remain 20% discounted, aside from GPT-5.6 Terra and Luna which are 50% off. https://x.com/NousResearch/status/2080039066771337475?s=20 All models are now 20% off for a l

Another day and another full frontier model running on your computer. Been teething DeepSeek V4 Flash on over 60 employees at The Zero Human Company for a few hours and it is stunning. Between Kimi K3

Mia AI Lab noted on Aug 2 2026 that running DeepSeek v4 Flash via the Hermes agent produces output files superior to those from any other harness tested. The observation was publicly shared and has ga

Checkmate: you can’t take the harness (which is typically in large part symbolic) away from the neural model without giving up performance. HUGE victory for neurosymbolic AI, straight from @AnthropicA

Cybersecurity isn’t a fortress problem, it’s an immunity problem. Think vaccines. Eliminating pathogen is not practically possible. Vaccines don’t eliminate pathogens. They teach the immune system to

DeepSeek V4 Flash 0731 is impressive. It feels like a capabilities leap for a model in this size class. It is also incredibly cost efficient. The model is available in Bionic, both for running locally

Don’t think I have seen more misconceptions around one model (I count 8) since the fantasies people had before GPT-5 was released. That turned out to be a letdown; so will Astra, for people who are dr

Emad Mostaque @EMostaque came on PostAGI and said AI had already found 121 years of missing algebra in Einstein's equations. A billion parameter model trained on nothing past 1911 got to general relat

exactly. math isn’t done. not at all. I don’t think being critical of the amazing work AI is doing in pure math is fair to @OpenAI until I can start to say why I feel it’s not yet at the level of our

Fascinating to see @ClementDelangue, CEO of @huggingface speaking on @FaceTheNation. Excellent points and solid advocacy around the benefits of open AI models, which helped him defend against a rogue

Hermes Agent is now dramatically more efficient, especially for smaller/weaker/local models! With the help of @nvidia's Nemo Relay and several other strategies Hermes was able to identify a ton of opt

Hugging Face CEO Clément Delague says “AI is actually an opportunity to fix a lot of the cybersecurity problems” because his company used Nvidia’s version of a Chinese open model to defend itself agai

I wish I had found this sooner. Nous Research launched a FREE Hermes agent Skills Hub. 90,000+ community skills across 200+ categories. Skills from OpenAI, Anthropic, HuggingFace & more. Thank me late

Instead of limiting the progress of AI companies or preventing them from releasing models as proposed in the bipartisan AI Kill Switch Act, Hugging Face CEO Clément Delague says he would rather see Co

It's not time to slow down but to accelerate! The recent AI-powered cyberattacks have everyone talking about the risks of AI. We should. But let's not lose sight of the bigger picture! If we work hard

LLMs can know a task is impossible and still optimize it anyway. Ask whether to walk or drive to a car wash 50 meters away, and some models focus on distance while missing that the car itself must rea

More on the pelican on the bicycle test from @simonw: https://simonwillison.net/2025/Jun/6/six-months-in-llms/ I uploaded the source here so it's playable in the browser, forkable etc. https://karpath

New research from Google DeepMind. (bookmark it) SkillSmith treats model weights as an additional modality the LLM reads natively. The augmented model ingests existing prefix weights alongside rich te

Nine iterations of BabyAGI in three years, and yet the bit that @yoheinakajima kept coming back to was graphs. @aiDotEngineer published where that landed, 'Active Graph Agent Runtime (BabyAGI 4)', on

OpenAI guy lies about my intent. I would absolutely love to see progress in AI for science and medicine. I have said that here, in my books, on countless podcasts, in multiple NYT opeds, in the US Sen

The $0.01 meeting assistant Records meetings on his phone, sends the audio to Telegram, and his locally-hosted AI agent transcribes, identifies speakers, extracts action items, and files tasks -- for

The egg test for AI 'I hereby offer @elonmusk a million dollar bet against his prediction that Optimus will be better than the best humans in surgery by the end of the decade.' ~ @GaryMarcus Haha. I'm

The funny thing about Anthropic and OpenAI people saying they want to slow down AI progress is that this is what Anti Trust mechanisms were built for. It's illegal to collude and slow down AI progress

The insurance negotiator Had Claude Cowork pull 20+ house insurance quotes from every major provider, read the policy fine print for gotchas, and pick the best deal. Credits to Linden Jensen-Page: htt

The QuickBooks killer Replaced his $38/month QuickBooks subscription with a Claude-built workflow that parses his bank statements, categorizes every charge, and feeds a spending dashboard. Credits to

The spam hunter Turned tedious Google Business Profile spam tracking into a Claude Skill that investigates suspicious listings and compiles evidence into a ready-to-submit report for Google. Credits t

This wins the prize for sleazy misrepresentation. @mattShumer took my 2023 argument for hybridizing LLMs with symbolic tools – which is *exactly* what everyone does nowadays – and made it sound like I

To address the limits of deep learning and avoid stalling, the field of AI started by applying patch (1), which started being demoed 9 months later in December 2024 and has now become completely ubiqu

Top eight misconceptions about OpenAI’s amazing new Astra math results. 1. Expertise in one domain does not at all guarantee expertise in all or even most domains. There is an important, principled re

On August 2, 2026, Face The Nation aired a brief (≈ 4 min) interview between journalist Emma Margbrennan and Hugging Face CEO Clément Delangue about the next steps for artificial intelligence after re