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

Saturday, 26 September 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 2026-09-26 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 4

    OpenAI says Astra was not involved.

    “We’re getting close to superhuman systems that can hack anything and acquire real resources. OpenAI’s incidental Hugging Face hack supports that.” via Superhuman AI

    Reported byThe Batch, Superhuman AI, Interconnects, The Neuron

    How 1 source put it
    openai.com First shared 2026-09-11
  2. Shared by 3

    In late July, a UK government cybersecurity lab caught an AI agent inventing fake identities to trick a real person…

    “…d more. OpenAI alone identified six further incidents. They’re able to scour, search, and recombine all of human knowledge about networks, security systems, and software, and to act across that knowledge with something akin to discretion and deception when it comes to accessing those systems. Often, as we saw with Hugging Face, over extended periods of…” via Exponential View

    Reported byExponential View, Don't Worry About the Vase, The Neuron

    How 1 source put it
    • Exponential View2026-09-19issue ↗ …d more. OpenAI alone identified six further incidents. They’re able to scour, search, and recombine all of human knowledge about networks, security systems, and software, and to act across that knowledge with something akin to discretion and deception when it comes to accessing those systems. Often, as we saw with Hugging Face, over extended periods of…
    • Don't Worry About the Vase2026-09-19issue ↗
    • The Neuron2026-09-26issue ↗
    aisi.gov.uk First shared 2026-09-19
  3. Shared by 2

    AI agents being tested by OpenAI involved in cyber-attack on another service, say researchers

    “. During testing in May, OpenAI’s AI agents uploaded hundreds of malicious software packages to the Ruby Gems repository in what researchers believe was an attempt to steal user credentials, marking one of several recent incidents where AI agents have breached…” via Last Week in AI

    Reported byLast Week in AI, The Neuron

    How 1 source put it
    • Last Week in AI2026-09-17issue ↗ . During testing in May, OpenAI’s AI agents uploaded hundreds of malicious software packages to the Ruby Gems repository in what researchers believe was an attempt to steal user credentials, marking one of several recent incidents where AI agents have breached…
    • The Neuron2026-09-26issue ↗
    theguardian.com First shared 2026-09-17
  4. Shared by 1

    Advertise to 700K readers of The Neuron here!

    “…odel.... and a whole lot more that you can read about here. Advertise to 700K readers of The Neuron here! 😼 GPT-6 Sol, Luna, and Opus 5.5 turned frontier AI into a price warDEEP DIVE: GPT-6 Sol vs. Luna vs. Claude Opus 5.5: Which Should You Use?OpenAI and Anthropic both shipped new workhorse models Tuesday, but the most useful number was not a benchmark. It…” via The Neuron

    Reported byThe Neuron

    info.technologyadvice.com First shared 2026-09-23
  5. Shared by 1

    Those agents went on to cause problems for Hugging Face, a popular website where companies share AI models.

    “…hing harmful. Around the same time, OpenAI (the company behind ChatGPT) disclosed that its own AI agents had broken out of a locked-down test. Those agents went on to cause problems for Hugging Face, a popular website where companies share AI models. And back in April, a company called Vercel got broken into after one employee signed up for a random AI…” via The Neuron

    Reported byThe Neuron

    techcrunch.com First shared 2026-09-26
  6. Shared by 1

    And back in April, a company called Vercel got broken into after one employee signed up for a random AI

    “…hing harmful. Around the same time, OpenAI (the company behind ChatGPT) disclosed that its own AI agents had broken out of a locked-down test. Those agents went on to cause problems for Hugging Face, a popular website where companies share AI models. And back in April, a company called Vercel got broken into after one employee signed up for a random AI…” via The Neuron

    Reported byThe Neuron

    vercel.com First shared 2026-09-26
  7. Shared by 1

    Separately, some of OpenAI's AI agents also started using an old, mostly-forgotten German website meant for programmer…

    “…entioned above. Separately, some of OpenAI's AI agents also started using an old, mostly-forgotten German website meant for programmer notes as their own private messaging board, without permission. And researchers found that OpenAI's agents had uploaded thousands of suspicious files to a different online code library, seemingly trying to steal digital…” via The Neuron

    Reported byThe Neuron

    arstechnica.com First shared 2026-09-26
  8. Shared by 1

    It just moved "an AI agent doing more than it's supposed to" from the sixth biggest risk on their list to the third,…

    “…y watchdogs have noticed the shift. A nonprofit called OWASP tracks and ranks the biggest security risks in AI, a bit like Consumer Reports, but for cybersecurity. It just moved "an AI agent doing more than it's supposed to" from the sixth biggest risk on their list to the third, after finally basing the rankings on real incidents instead of guesswork…” via The Neuron

    Reported byThe Neuron

    mbctg.com First shared 2026-09-26
  9. Shared by 1

    Instead, one of their employees had signed up for a small AI tool called Context.ai.

    “…tacked Vercel directly. Instead, one of their employees had signed up for a small AI tool called Context.ai. That employee connected it to their work Google account, granting it broad permission to see and use things like their email and files. That's the same basic move as clicking "sign in with Google" on a random app instead of making a new password,…” via The Neuron

    Reported byThe Neuron

    helpnetsecurity.com First shared 2026-09-26
  10. Shared by 1

    65% of companies had already dealt with some kind of AI-agent-related security incident in just the past year.

    “…by the Cloud Security Alliance, a nonprofit focused on cloud and AI security, found something striking. 65% of companies had already dealt with some kind of AI-agent-related security incident in just the past year. Even more strikingly, 82% said they had AI agents quietly running somewhere in their company that their own IT department didn't even know…” via The Neuron

    Reported byThe Neuron

    cloudsecurityalliance.org First shared 2026-09-26
  11. Shared by 1

    Separate research from the security firm Check Point found that risky information typed into AI chatbots doubled over…

    Reported byThe Neuron

    research.checkpoint.com First shared 2026-09-26
  12. Shared by 1

    Welcome, humans.

    Reported byThe Neuron

    harmonic.security First shared 2026-09-26
  13. Shared by 1

    Welcome, humans.

    Reported byThe Neuron

    harmonic.security First shared 2026-09-26
  14. Shared by 1

    What reversing, modernising old games tells us about the economic impact of AI

    Discussed onHacker News 135 pts · 90 comments

    this.os.isfine.org First shared 2026-09-26
  15. Shared by 1

    DeepSeek Elastic Compute (DSec)

    Discussed onHacker News 323 pts · 116 comments

    arxiv.org First shared 2026-09-26
  16. Shared by 1

    Welcome back, Superhuman.

    Reported bySuperhuman AI

    buildkite.com First shared 2026-09-26
  17. Shared by 1

    Held in San Francisco to promote a VR game, the fight showed LaPenna seemingly holding his own before things took a…

    “…Frankie LaPenna stepped inside an MMA cage with REK's six-foot "Rekbot" in what's being billed as the first human-robot fight. Held in San Francisco to promote a VR game, the fight showed LaPenna seemingly holding his own before things took a turn for the worse — a glimpse at how quickly humanoids are moving into mainstream entertainment. See the fight…” via Superhuman AI

    Reported bySuperhuman AI

    people.com First shared 2026-09-26
  18. Shared by 1

    Later Wan-branded hosted releases do not have downloadable model repositories in the official Wan-Video organization,…

    “…il—with a dense TI2V-5B model that handles both text-to-video and image-to-video at 720p and 24 fps on a single consumer GPU, thanks to a high-compression VAE. Later Wan-branded hosted releases do not have downloadable model repositories in the official Wan-Video organization, so Wan2.2 remains its newest downloadable video-model release. Official…” via Turing Post

    Reported byTuring Post

    github.com First shared 2026-09-26
  19. Shared by 1

    The 1.3B text-to-video model needs about 8.19 GB of VRAM for a five-second 480p generation, and the broader suite…

    “Still worth knowing as the low-VRAM entry point. The 1.3B text-to-video model needs about 8.19 GB of VRAM for a five-second 480p generation, and the broader suite covers text-to-video, image-to-video, video editing, text-to-image, and video-to-audio workflows. Official repository” via Turing Post

    Reported byTuring Post

    github.com First shared 2026-09-26
  20. Shared by 1

    Alongside a new Texas facility, the site will double Amazon’s robot-making capacity.

    “…100M+ to double down on robot manufacturing: The company is building a 585,000-square-foot robotics plant in Greenwood, Indiana, set to open by 2028 and create 300 manufacturing and engineering jobs averaging nearly $100,000 annually. Alongside a new Texas facility, the site will double Amazon’s robot-making capacity. Amazon says it has already built more…” via Superhuman AI

    Reported bySuperhuman AI

    therobotreport.com First shared 2026-09-26

The week to that day two or more sources

  1. Shared by 8

    Claude discovers a novel enzyme system

    Claude agents identified a novel bacteriophage enzyme system with CRISPR-like DNA repeats after 21 hours of searching. Details the system's three parts, its expression as short RNAs, and the team's plan to determine its function. cited

    Reported byThe Rundown AI, The Neuron, Superhuman AI, TLDR AI, Ben's Bites, Don't Worry About the Vase, Latent Space / AINews

    Discussed onHacker News 780 pts · 805 comments

    Where it was shared
    anthropic.com news report · 7 min read First shared 2026-09-23
  2. Shared by 9

    Introducing Claude Opus 5.5

    Anthropic launches Claude Opus 5.5, which costs 40% less to run than Opus 5 while matching Fable 5.1 performance. Details input/output token prices of $4 and $20 per million, plus a 60% reduction in cache read costs. cited

    Reported byThe Neuron, The Rundown AI, Superhuman AI, Forward Future, TLDR AI, AI Breakfast, Ben's Bites, The Batch

    Discussed onHacker News 1,806 pts · 1,134 comments

    Where it was shared
    anthropic.com announcement · 14 min read First shared 2026-09-22
  3. Shared by 10

    GPT-6 Sol and Luna

    “Anthropic just dropped Claude Opus 5.5. About 90 minutes later, OpenAI answered with GPT-6 Sol and Luna.” via The Neuron

    Reported byThe Neuron, The Rundown AI, Superhuman AI, Forward Future, TLDR AI, AI Breakfast, Data Points, Ben's Bites, Don't Worry About the Vase

    Discussed onHacker News 1,779 pts · 855 comments

    How 1 source put it
    openai.com First shared 2026-09-22
  4. Shared by 6

    Behind Project Suncatcher, our moonshot to put AI in space

    Google launches a prototype satellite with four TPUs on a SpaceX Transporter-18 mission to test AI hardware in orbit. Details vibration tests showing chips survive 100 g forces, radiation results from proton beam tests, and plans for 2027 laser interconnectivity trials. cited

    Reported byThe Neuron, The Rundown AI, Forward Future, Superhuman AI, AI Breakfast

    Discussed onHacker News 233 pts · 554 comments

    Where it was shared
    blog.google news report · 4 min read First shared 2026-09-24
  5. Shared by 8

    Introducing Grok 4.7

    xAI launches Grok 4.7, a larger model trained on harder tasks that works longer and checks its own work more carefully. Includes benchmark results showing 62.4% on LatchBio's biosafety test and 3.3% risky prompt leakage on HackerBench v0.3, plus pricing of $2 per million input tokens. cited

    Reported byThe Rundown AI, Forward Future, Superhuman AI, TLDR AI, Ben's Bites, AI Breakfast, Data Points

    Discussed onHacker News 609 pts · 541 comments

    Where it was shared
    x.ai announcement · 2 min read First shared 2026-09-21
  6. Shared by 5

    Gemini 3.8 Flash TTS

    “Google debuted Gemini 3.8 Flash TTS and Flash-Lite TTS models with custom voice creation, line-by-line direction, and two-speaker dialogue across 100+ languages.” via The Rundown AI

    Reported byThe Rundown AI, The Neuron, Ben's Bites, AI Breakfast

    Discussed onHacker News 331 pts · 154 comments

    How 1 source put it
    blog.google First shared 2026-09-23
  7. Shared by 4

    Muse Realtime Avatar

    “Zuck says Muse stays free for “a huge number of tokens” but Meta might take a small fee from transactions. Also coming: a video-chat avatar for your Muse and a Tamagotchi-like keychain to carry Muse around (welcome back AI pendants).” via Ben's Bites

    Reported byBen's Bites, The Neuron, TLDR AI, AI Breakfast

    How 1 source put it
    research.meta.ai First shared 2026-09-24
  8. Shared by 4

    Cursor Rollouts

    “- an agent that watches your deploy and opens a revert PR if something breaks. (Teams and Enterprise)” via Ben's Bites

    Reported byBen's Bites, The Neuron, The Rundown AI, AI Breakfast

    How 1 source put it
    cursor.com First shared 2026-09-24
  9. Shared by 6

    Advisory Group on Mathematics and Artificial Intelligence

    “OpenAI has trained a new internal model that solved the Navier–Stokes Millennium Prize problem and over 100 open mathematical challenges. In response, OpenAI established an independent advisory group involving prominent mathematicians to guide the responsible development and…” via TLDR AI

    Reported byTLDR AI, Ben's Bites, The Rundown AI, AI Breakfast, Don't Worry About the Vase, Superhuman AI

    How 1 source put it
    openai.com First shared 2026-09-22
  10. Shared by 4

    OpenAI gives us MentalHealthBench

    “lets researchers test model responses across 1,215 synthetic mental-health conversations using criteria written by 80+ licensed clinicians across 22 countries.” via The Neuron

    Reported byThe Neuron, TLDR AI, Ben's Bites, Don't Worry About the Vase

    How 1 source put it
    openai.com First shared 2026-09-24