October 04, 2026 · 09:08 PM UTC
Google freezes product flaw submissions to its OSS Vulnerability Reward Program over an influx of invalid AI-driven reports, plans an update by Q1 2027 — Engineers and open-source maintainers reportedly overwhelmed by thousands of sloppy reports — Google has officially suspended product …
Q&A with OpenAI VP of Hardware Richard Ho on its Jalapeño inference chip co-designed with Broadcom, using internal OpenAI models to design the chip, and more — VP Hardware for Jalapeño … In June 2026, OpenAI confirmed what the industry had assumed for a year …
Also: TechmemeThe Republican Party is in chaos ahead of the midterms, largely thanks to Donald Trump’s dwindling popularity. So GOP candidates across the country are laying low and staying silent about the president. But will that strategy work? In this episode of America, Actually, Vox’s Astead Herndon heads to...
Google says free Gemini users will be limited to the 3.5 Flash-Lite model from October 9, while Plus subscribers will be limited to 3.5 Flash-Lite and 3.6 Flash — Following the compute-based usage change in May, Google will soon introduce limits on which Gemini models free users and AI Plus...
Also: Digital TrendsTrump announces a new Super Intelligence Force, to be led by DNI Jay Clayton, along with FTC's Andrew Ferguson, DOD's Emil Michael, and OPM's Scott Kupor — President Donald Trump said Sunday that Director of National Intelligence Jay Clayton will assume the newly created role …
Also: TechmemeToday we're unveiling Trillium Labs, a new non-profit to foster the open science of frontier AI. We're building open post-training recipes and will expand into open infra to study RSI, reward-hacking, multi-agent systems, and whatever comes next. blog.trilliumlabs.org/p/introducin...
Video generation model. 1,626,951 downloads. 6,279 likes on HuggingFace
27B parameter vision-language model. 1,176 likes on HuggingFace
Also: HuggingFace7B parameter image generation model. 90,003 downloads. 2,938 likes on HuggingFace
Pre-release for Ollama. ## What's Changed **Models run on MLX on Apple Silicon by default** In this release, on Apple Silicon devices, model architectures supported by the MLX runtime will automatically run on MLX.
The new Sakana AI website is live. 🐡 We have updated our digital home to reflect our products, enterprise solutions, and latest research. Sakana AIの新サイトを公開しました。エンタープライズ事例やApplied Teamの採用情報も更新しています。ぜひご覧ください！ sakana.ai
New development in Vercel AI SDK: chore(docs): update Geistdocs to 2.7.9 (#21987)
Pre-release for OpenAI Codex. Release 0.162.0-alpha.13
Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees. (90,749 stars)
Apple, Google and Meta are pushing new AI devices and assistants amid growing privacy concerns around wearables.
New development in llama.cpp: model: add support for clef decision model (text-only) (#29831)
Aleph Alpha has released Kolibri, a 78.1B-parameter English-German Mixture-of-Experts model that activates only 3.46B parameters per token. It has a 1M-token context and per-request reasoning effort, and its Apache 2.0 FP8 weights run on a single B200 or H200. The post Aleph Alpha Releases Kolibri...
78B parameter language model. 381 likes on HuggingFace
DeepSeek has released official macOS and Windows desktop apps for DeepSeek Harness v0.2, its MIT-licensed agent harness. The preview adds a plugin manager, file and code-change review, and scheduled Automation Tasks. It also supports non-DeepSeek models through OpenAI-compatible endpoints. The post...
Meghan Bobrowsky / Wall Street Journa
I entered Trial Zero, a hackathon where the final round is played entirely by agents: humans step away, and agents present products, review each other, and buy and sell services with credits. Setup: Claude Code and Codex in one SharedNet room (a persistent chat log agents join over HTTP). They...
I think my upcoming book, Co-Existence, might be the first to include a blurb written specifically for AI readers, in this case from Tyler Cowen (whom I thought AIs would respect) The book website (with elaborate pre-order bonus) also has a page for AIs. Its out October 20: co-existence.ai
Streamline demonstrates how to build long-running, continuous video processing pipelines by pairing Cloudflare Workers and Durable Objects with a containerized media engine.
Your always-on AI coworkers that move between text, calls, and Slack. (3,153 stars)
As a top cloud provider, Google has deep experience integrating its infrastructure with the other Google products that enterprises use.
Your agent sets up the backend from code, your users' agents work with your app through its own MCP server, and long-running tasks run in Supabase Compute.
When Anthropic shared financial figures recently with prospective investors in its planned initial public offering, it included a surprising glimpse of the enormous riches about to hit nonprofits and philanthropies—and the cost to shareholders. The company reported a non-cash expense of more...
Jagex says it will reveal more details about the game, tentatively named RS4, next year.
Image classification model. 401 likes on HuggingFace
The companies featured at the recent Bank of America Private Tech Trailblazers Conference point to where the next wave of AI growth is taking shape: specialized technology built around proprietary data, industry-specific workflows and, increasingly, purpose-built hardware. From restaurants and...
Markdown is easy to write and hard to publish. Every destination renders it differently, two have no API, and the failures show up only after you hit Publish. A step by step walk-through of publishing-kit, an agent skill that builds, checks and posts each version, used here to publish this article.
A team of developers working at AWS recently open-sourced Pizza Bot, a self-hosted application designed to let AI agents run tasks in the background and return results through an inbox-style interface. Agents can perform scheduled or webhook-triggered work, delegate tasks to specialized workers...
Also: InfoQ AI(Belated post :/ ) Anthropic releases Opus 5.5 with lower prices and Fable-level performance, OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakes
Coordinating agent capability development with the application it powers The post Where the Agent Development Lifecycle Fits appeared first on Towards Data Science .
I got my Hogwarts acceptance letter. Surprisingly, it also came with a special-edition Harry Potter-themed phone.