Model Follower.

Up to date

Rebuilt
2026-10-25 06:54 UTC
Release radar
117 on watch
Hub snapshot
2026-10-25
Re-checked
daily · every number sourced and dated

Links / Past days

Tuesday, 6 October 2026

The AI links the most newsletters and link sources had shared by the end of the day (UTC), each source counted once.

As of 2026-10-06 23:59 UTC Status sealed Rebuilt from issues on file

Reconstructed after the day from the issues on file (dates as published). An issue counts on the date it carries, even if it was collected later.

That day 36-hour window

  1. Shared by 7

    Introducing Beam: Reflection’s 501B open-weight model

    Reflection AI launches Beam, a 501B sparse MoE model with 23B active parameters trained on 23.8T tokens. Details how 10.5K NVIDIA GB300 GPUs generated over 100 million rollouts to achieve 3–4× less inference compute than Qwen 3.8-Max. cited

    Reported byTuring Post, Latent Space / AINews, The Rundown AI, The Neuron, Superhuman AI, TLDR AI

    Discussed onHacker News 551 pts · 174 comments

    Where it was shared
    reflection.ai announcement · 12 min read First shared 2026-10-05
  2. Shared by 3

    OpenAI to Add Text Watermarks

    “On Monday, OpenAI said it will start hiding an invisible signal in the text that ChatGPT and Codex (its coding assistant) write for people in the European Union. The signal lives in the AI's word choices (a pattern in which words it picks), invisible to you but findable by a special detector.” via The Neuron

    Reported byThe Rundown AI, The Neuron, TLDR AI

    How 1 source put it
    • The Rundown AI2026-10-06issue ↗
    • The Neuron2026-10-06issue ↗ On Monday, OpenAI said it will start hiding an invisible signal in the text that ChatGPT and Codex (its coding assistant) write for people in the European Union. The signal lives in the AI's word choices (a pattern in which words it picks), invisible to you but findable by a special detector.
    • TLDR AI2026-10-06issue ↗
    openai.com First shared 2026-10-06
  3. Shared by 2

    EmbeddingGemma 2: an open, lightweight multimodal embedding model

    Google launches EmbeddingGemma 2, a 740M-parameter Apache 2.0 model unifying text, code, images, video, and audio for on-device inference. Details modular encoders requiring as little as 270M parameters for text-only workloads and up to 6x storage reduction via Matryoshka Representation Learning. cited

    Reported bySimon Willison

    Discussed onHacker News 429 pts · 46 comments

    blog.google announcement · 3 min read First shared 2026-10-06
  4. Shared by 2

    Introducing Mistral Large 4

    Mistral launches a public preview of its 1 trillion-parameter, natively multimodal model Le Chonk. Includes cybersecurity benchmarks where it scores 82% on vulnerability patching and beats GPT-6-Astra on legal tasks. cited

    Reported bySimon Willison

    Discussed onHacker News 522 pts · 5 comments

    mistral.ai announcement · 9 min read First shared 2026-10-06
  5. Shared by 2

    on Mistral Large 4 — Hacker News.

    Reported bySimon Willison

    Discussed onHacker News 2,027 pts · 1,210 comments

    mistral.ai First shared 2026-10-06
  6. Shared by 2

    The Top 100 Gen AI Consumer Apps — 7th Edition

    a16z releases the 7th edition of its Top 100 Gen AI Consumer Apps with a new ranking by U.S. card spending. Includes data showing the top 1% of payers account for 19.5% of all observed consumer AI spend and details on new personal agents. cited

    Reported byThe Rundown AI, Ben's Bites

    Where it was shared
    a16z.com news report · 14 min read First shared 2026-10-06
  7. Shared by 2

    Instinct in group chats

    “Instinct now lets early-access users add a shared agent to group chats for planning trips, tickets, carpools, and events. Personal Instincts enforce permissions, keep account access separate, and pause sharing when new members join.” via TLDR AI

    Reported byTLDR AI, Ben's Bites

    How 1 source put it
    • TLDR AI2026-10-06issue ↗ Instinct now lets early-access users add a shared agent to group chats for planning trips, tickets, carpools, and events. Personal Instincts enforce permissions, keep account access separate, and pause sharing when new members join.
    • Ben's Bites2026-10-06issue ↗
    x.com X post First shared 2026-10-06
  8. Shared by 3

    OpenAI Expands Ads in ChatGPT

    “OpenAI introduced a new visual ad format in ChatGPT alongside expanded measurement tools, partnerships, and brand-suitability work. The company positioned advertising as a way to support broader access to ChatGPT while giving businesses another channel to reach users during product discovery and decision-making.” via TLDR AI

    Reported byAI Breakfast, Data Points, TLDR AI

    How 1 source put it
    • AI Breakfast2026-10-05issue ↗
    • Data Points2026-10-05issue ↗
    • TLDR AI2026-10-06issue ↗ OpenAI introduced a new visual ad format in ChatGPT alongside expanded measurement tools, partnerships, and brand-suitability work. The company positioned advertising as a way to support broader access to ChatGPT while giving businesses another channel to reach users during product discovery and decision-making.
    openai.com First shared 2026-10-05
  9. Shared by 2

    Norway Proposes AI-Glasses Ban

    “Norway is moving to keep AI smart glasses out of gyms, schools, playgrounds, and beaches, with a temporary ban headed to parliament as officials weigh permanent rules.” via The Rundown AI

    Reported byThe Rundown AI, Forward Future

    How 1 source put it
    • The Rundown AI2026-10-06issue ↗ Norway is moving to keep AI smart glasses out of gyms, schools, playgrounds, and beaches, with a temporary ban headed to parliament as officials weigh permanent rules.
    • Forward Future2026-10-06issue ↗
    apnews.com First shared 2026-10-06
  10. Shared by 1

    — Victoria Kim, Reporting from the Australian parliament Tags: accidental-cyberattacks, generative-ai,…

    “… in place additional monitoring to allow “immediate intervention” by staff to stop training if the company’s models access the internet in ways they’re not supposed to, Mr. Kwon [chief strategy officer at OpenAI] said. — Victoria Kim, Reporting from the Australian parliament Tags: accidental-cyberattacks, generative-ai, ai-security-research, openai, ai,…” via Simon Willison

    Reported bySimon Willison

    nytimes.com First shared 2026-10-06
  11. Shared by 2

    Devin now maintains memory of how you work in an open format.

    “Agent memory: Cognition launched Devin “Dreaming,” which prunes and links a memory graph overnight. It is open-sourcing the git- and markdown-backed format as Agent Memory Repo (launch, format).” via Latent Space / AINews

    Reported byLatent Space / AINews, Ben's Bites

    How 1 source put it
    • Latent Space / AINews2026-10-06issue ↗ Agent memory: Cognition launched Devin “Dreaming,” which prunes and links a memory graph overnight. It is open-sourcing the git- and markdown-backed format as Agent Memory Repo (launch, format).
    • Ben's Bites2026-10-06issue ↗
    x.com X post First shared 2026-10-06
  12. Shared by 1

    Release: llm-openai-decisions 0.1a0 OpenAI released their new Jev-style Decisions API, as previously announced at last…

    Reported bySimon Willison

    github.com First shared 2026-10-06
  13. Shared by 1

    OpenAI released their new Jev-style Decisions API, as previously announced at last week's DevDay.

    Reported bySimon Willison

    developers.openai.com First shared 2026-10-06
  14. Shared by 1

    Since I already have an llm-typesafe plugin for talking to Jev, I had GPT-6 Astra read the new OpenAI API documentation…

    Reported bySimon Willison

    github.com First shared 2026-10-06
  15. Shared by 1

    Consult the README for full details of how to run the other types of questions.

    Reported bySimon Willison

    github.com First shared 2026-10-06
  16. Shared by 1

    Release: llm-mistral 0.16 Adds support for reasoning models, such as the newly released Mistral Large 4.

    Reported bySimon Willison

    github.com First shared 2026-10-06
  17. Shared by 2

    People are asking ChatGPT to help them decide how to vote in the midterms

    “NPR found voters are using chatbots to compare candidates and debunk viral claims; about one in five voters consult them for election news, per a September survey.” via The Neuron

    Reported byThe Neuron, Forward Future

    How 1 source put it
    • The Neuron2026-10-06issue ↗ NPR found voters are using chatbots to compare candidates and debunk viral claims; about one in five voters consult them for election news, per a September survey.
    • Forward Future2026-10-06issue ↗
    npr.org 6 min read First shared 2026-10-06
  18. Shared by 2

    Satya Nadella reinvented Microsoft once. Can he do it again in the AI era?

    “CNBC asked whether Nadella can reinvent Microsoft again, noting Copilot has 30M seats, but the company still lacks an AI smash hit.” via The Neuron

    Reported byThe Neuron, Forward Future

    How 1 source put it
    cnbc.com 9 min read First shared 2026-10-06
  19. Shared by 2

    Two Room-Temperature Antiferromagnetic Semiconductor Candidates

    AI agents identified KV[Cr(CN)₆], a 1999 material, and a new design as Luttinger-compensated semiconductors with zero net magnetism. Details band gaps, spin windows, and synthesis hurdles for both candidates, plus a GitHub repo with raw simulation files. cited

    Reported byTLDR AI

    Discussed onHacker News 491 pts · 340 comments

    vals.ai analysis · 8 min read First shared 2026-10-05
  20. Shared by 1

    In April 2024 OpenAI offered to "cover the financial cost of users re-embedding content with these new models" -…

    “(In April 2024 OpenAI offered to "cover the financial cost of users re-embedding content with these new models" - https://openai.com/index/gpt-4-api-general-availability/ - but I don't think that's something we can rely on from every provider.)” via Simon Willison

    Reported bySimon Willison

    openai.com First shared 2026-10-06

The week to that day two or more sources

  1. Shared by 7

    Introducing Beam: Reflection’s 501B open-weight model

    Reflection AI launches Beam, a 501B sparse MoE model with 23B active parameters trained on 23.8T tokens. Details how 10.5K NVIDIA GB300 GPUs generated over 100 million rollouts to achieve 3–4× less inference compute than Qwen 3.8-Max. cited

    Reported byTuring Post, Latent Space / AINews, The Rundown AI, The Neuron, Superhuman AI, TLDR AI

    Discussed onHacker News 551 pts · 174 comments

    Where it was shared
    reflection.ai announcement · 12 min read First shared 2026-10-05
  2. Shared by 11

    Gemini 4 Argon: our next era of frontier intelligence

    Google announces Gemini 4 Argon, a frontier model launching at $2 per million input tokens and $10 per million output tokens. Details internal wins like a 40% quantum optimization beat, 1M token output limits, and a 91.7% LVBench score. cited

    Reported byLatent Space / AINews, The Rundown AI, Superhuman AI, TLDR AI, Ben's Bites, Don't Worry About the Vase, The Neuron, AI Breakfast, TheSequence, Turing Post

    Discussed onHacker News 1,702 pts · 1,190 comments

    Where it was shared
    blog.google announcement · 6 min read First shared 2026-09-30
  3. Shared by 3

    OpenAI to Add Text Watermarks

    “On Monday, OpenAI said it will start hiding an invisible signal in the text that ChatGPT and Codex (its coding assistant) write for people in the European Union. The signal lives in the AI's word choices (a pattern in which words it picks), invisible to you but findable by a special detector.” via The Neuron

    Reported byThe Rundown AI, The Neuron, TLDR AI

    How 1 source put it
    • The Rundown AI2026-10-06issue ↗
    • The Neuron2026-10-06issue ↗ On Monday, OpenAI said it will start hiding an invisible signal in the text that ChatGPT and Codex (its coding assistant) write for people in the European Union. The signal lives in the AI's word choices (a pattern in which words it picks), invisible to you but findable by a special detector.
    • TLDR AI2026-10-06issue ↗
    openai.com First shared 2026-10-06
  4. Shared by 10

    GLM-5.3 and the spread of advanced cyber capabilities

    Anthropic reports attackers bypass GLM-5.3 safeguards 64% to 100% of the time using simple techniques. Includes ExploitBench results showing 50 of 410 end-to-end exploits and details the $4,400 cost to abliterate the model. cited

    Reported bySimon Willison, Latent Space / AINews, Last Week in AI, TLDR AI, AI Breakfast, The Neuron, Don't Worry About the Vase, The Batch, Interconnects

    Discussed onHacker News 256 pts · 241 comments

    AboutGLM-5.3-Flash, GLM-5.3, GLM-5

    Where it was shared
    anthropic.com analysis · 8 min read First shared 2026-09-29
  5. Shared by 12

    Dots. OpenAI has given us… dots

    “OpenAI’s DevDay gives us even more reasons to question why would we need a laptop. Dots, its new personal agents, have their own cloud computers, browsers, and connections to your applications. OpenAI says they can continue working on projects and carry context across ChatGPT, Slack, and Teams. And soon texts, as well.” via Turing Post

    Reported byTuring Post, Latent Space / AINews, The Neuron, Superhuman AI, TLDR AI, AI Breakfast, Data Points, Ben's Bites, Don't Worry About the Vase, Exponential View, TheSequence

    Discussed onHacker News 769 pts · 647 comments

    How 1 source put it
    openai.com First shared 2026-09-29
  6. Shared by 5

    Anthropic seeks religious wisdom for raising Claude

    “The New York Times reports that Anthropic had been consulting with religious scholars and theologians on the question of machine consciousness. Anthropic has prioritised questions of whether its systems might imbue consciousness and personhood more than other AI companies. It goes so far as Anthropic’s team proposed to withdraw from a Vatican event before…” via Exponential View

    Reported byForward Future, Exponential View, The Rundown AI, The Neuron

    Discussed onHacker News 166 pts · 444 comments

    How 1 source put it
    • Forward Future2026-09-30issue ↗
    • Exponential View2026-10-03issue ↗ The New York Times reports that Anthropic had been consulting with religious scholars and theologians on the question of machine consciousness. Anthropic has prioritised questions of whether its systems might imbue consciousness and personhood more than other AI companies. It goes so far as Anthropic’s team proposed to withdraw from a Vatican event before…
    • The Rundown AI2026-10-05issue ↗
    • The Neuron2026-10-05issue ↗
    nytimes.com First shared 2026-09-30
  7. Shared by 12

    More on the model here

    “: An upgrade for coding, computer use, and professional work. OpenAI claims near-Astra intelligence at one-fifth of Astra’s standard input and output token prices. Available through the API and listed paid ChatGPT plans.” via Turing Post

    Reported bySimon Willison, Turing Post, Latent Space / AINews, The Neuron, Superhuman AI, TLDR AI, AI Breakfast, Data Points, Ben's Bites, Don't Worry About the Vase, TheSequence

    Discussed onHacker News 1,067 pts · 955 comments

    How 1 source put it
    openai.com First shared 2026-09-29
  8. Shared by 7

    The Future Is for Everyone: Muse for Small Business

    Meta expands its personal AI agent with new skills and connectors to help small businesses run operations. Includes connectors for Instagram, Facebook, and Meta ads, plus integrations with Canva, Shopify, Stripe, and QuickBooks. cited

    Reported byThe Rundown AI, The Neuron, Superhuman AI, TLDR AI, Ben's Bites, TheSequence, Turing Post

    Where it was shared
    about.fb.com announcement · 2 min read First shared 2026-09-30
  9. Shared by 3

    OpenAI Expands Ads in ChatGPT

    “OpenAI introduced a new visual ad format in ChatGPT alongside expanded measurement tools, partnerships, and brand-suitability work. The company positioned advertising as a way to support broader access to ChatGPT while giving businesses another channel to reach users during product discovery and decision-making.” via TLDR AI

    Reported byAI Breakfast, Data Points, TLDR AI

    How 1 source put it
    • AI Breakfast2026-10-05issue ↗
    • Data Points2026-10-05issue ↗
    • TLDR AI2026-10-06issue ↗ OpenAI introduced a new visual ad format in ChatGPT alongside expanded measurement tools, partnerships, and brand-suitability work. The company positioned advertising as a way to support broader access to ChatGPT while giving businesses another channel to reach users during product discovery and decision-making.
    openai.com First shared 2026-10-05
  10. Shared by 2

    The Top 100 Gen AI Consumer Apps — 7th Edition

    a16z releases the 7th edition of its Top 100 Gen AI Consumer Apps with a new ranking by U.S. card spending. Includes data showing the top 1% of payers account for 19.5% of all observed consumer AI spend and details on new personal agents. cited

    Reported byThe Rundown AI, Ben's Bites

    Where it was shared
    a16z.com news report · 14 min read First shared 2026-10-06