A few papers & benchmarks on LLM failure to follow instructions 🧵 I. Can LLMs Unlearn Stubborn Training Conventions? “LLMs often exhibit cognitive inertia, struggling when tasked with instructions that explicitly deviate from their training norms” arxiv.org/abs/2509.04292 1/

Former Manchester City manager Pep Guardiola has urged the world not to forget Palestinians in Gaza. In a video with the UN agency for Palestinian refugees, UNRWA, he condemns the dehumanisation of victims and the suffering of children.

The most built-upon model on the Hugging Face Hub isn’t Llama or Qwen. It’s FLUX.1-dev from Black Forest Labs, a German lab: 43k derivative models. Europe trained ~8% of the open-model compute on the Hub, but ~18% of what people build on top of it is European. Build or surrender? Build, and build open. See you at GTC Berlin this week

Google has another banger 🤯 It’s called PaperBanana. It turns your methodology text into academic illustrations, helping you explain your research visually. 100% free.

I personally think bears waiting for a lower bottom are on the wrong side of the trade. Capital flowing in/out of BTC bottomed August 9th 2026.

Back when calculators were a little more controversial, v/@JonErlichman.

BREAKING: At least 74,475 people have been killed and 175,517 wounded by Israel’s genocidal war on Gaza since October 2023, according to the enclave’s Health Ministry. 🔴 LIVE updates: aje.news/e9uqrl

DeepSeek v4.1 Flash 1017 is basically GPT 5.6 Sol at home Incredible

RT @nntaleb: Another way the current administration is trapped in the ZERO-SUM FALLACY: The H-1B visa situation. + You are NOT taking a job…

Today, @Meta is open sourcing the weights for Muse Spark Open, the most capable US open-weights model to date. It’s integrated with the PyTorch ecosystem: it ships with Helion kernels, and TorchTitan can be used to further tune the model. More in the blog: pytorch.org/blog/muse-spark-…

Israeli spyware disguised as a digital advertising firm has targeted journalists’ devices at Al Jazeera and other companies. Here is what we know.

A 3B active model beats Opus 5 at memory. We open-sourced the full stack — model, engine, and app. Hy Memory handles chat, work, and code memory in one layer, and it's first on all four benchmarks we ran, leading MemoryArena by 10.2 points while cutting memory volume 51.8% and recall tokens 39.3%. The model isn't tied to our engine. Drop Hy-Memory-30B-A3B into Mem0, MemOS, or EverMemOS and it still comes out on top, even on the one system it never trained on. It's already running in our own products, where one assistant's multi-hop recall went from 50% to 90%. It also unifies memory across Codex, Cursor, WorkBuddy, and more. Repo: github.com/Tencent-Hunyuan/H… Weights: huggingface.co/tencent/Hy-Me… Questions welcome in the replies.

how do you fine-tune a decision model on your own data? in these two notebooks, you'll learn: • how to prepare the training data • how to fine-tune our d1-3B decision model • using @UnslothAI and @huggingface TRL all within 30 minutes on a free-tier GPU. unsloth: github.com/Liquid4All/cookbo… TRL: github.com/Liquid4All/cookbo… ps: how adorable are these halloween kitties? 👻

Since we started offering employee profit sharing in 2020, we've paid out over $24,000,000 to employees. We've never offered stock options. Not a fan of future maybe money. We pay real cash to real people so they can spend it on real things right now. When you choose to be our customer, you're making this possible. You're supporting an independent, fully customer-funded business that shares its success with the people who make it happen. You don't need investors, stock options, or complicated cap tables to share your success. You can just share the profits. It really can be that simple. If you're curious, this is how the profit sharing plan works: basecamp.com/handbook/benefi…

Meta releases Muse Spark Open, a new SOTA open model. Run on a Mac Studio via Dynamic 2-bit Unsloth GGUFs. Muse-Spark-Open-816B-A42B supports text, images, video and audio input with 1M context. GGUF: huggingface.co/unsloth/Muse-… Guide: unsloth.ai/docs/models/muse-…

"The odds of an academic "researcher" producing anything eventually used by society is of the order of .00001%. That includes scientists. The odds for a baker: 100%" - Nassim Nicholas Taleb

Today @GoogleResearch is launching Deep Review — an agentic system designed to stress-test manuscripts, design docs, and complex proposals before submission or publication. After pilot partnerships analyzing over 15,000 research papers, we’re making Deep Review globally accessible to help researchers, engineers, and professionals bring structured critique to their daily workflows.

Can you identify which LLM wrote a passage from the text alone? 👀 Each AI lab leaves behind distinct stylistic markers in vocabulary, phrasing, and punctuation. In Name That LLM, you'll identify which of 18 leading models from 6 organizations (Anthropic, Google, Moonshot AI, OpenAI, Qwen, and ZAI) wrote each unlabeled passage. 💰 Prize Pool: $50,000 ⏰ Entry Deadline: January 6, 2027

A finetune of EmbeddingGemma2 on Vietnamese audio is now a leading model on Vietnamese ASR, impressive. Check it out: huggingface.co/qnguyen3/QUAN…

Would love to see a nice open source version of this with Sentence Transformers!