Top AI Stories · November 5, 2026

Data centers lose at the ballot box; AI finds 948 Linux exploits

Voters backed dozens of local limits on data centers and elected candidates who want statewide pauses. OpenAI's agents found 948 Linux root exploits for about $1.2 million. Plus 5 more stories.

Listen to this brief

Free preview · first 0:30
0:00 / 0:30

Audio & video are paid features

Plus unlocks audio streaming and PDF downloads. Pro adds offline MP3 downloads, video, certificates, and more.

Plus adds:
  • Audio streaming
  • Downloadable PDFs
  • All AI Playbooks
  • Personalized content
Pro also adds:
  • Certificates of completion
  • Audio MP3 downloads
  • Video lessonssoon
  • & More…soon

Two months after polls first showed broad opposition to AI data centers, the midterm elections turned that opposition into votes, as voters backed local limits on data centers and elected candidates who want statewide pauses. OpenAI also showed how cheaply its agents can now find, and draft fixes for, serious flaws in Linux.

  1. 1

    Voters back dozens of local limits on data centers and elect moratorium supporters

    The midterm elections handed data center developers a broad defeat. Voters faced about 45 local ballot measures on data centers, and by Wednesday morning a tally by Redbud Consulting showed 33 of them restricting development, including bans on sites above 25 megawatts in several Ohio cities and a ten-year ban in El Dorado, Kansas. In an NBC News exit poll, 76 percent of voters said they opposed data centers in their own communities. Texas elected Democrat Gina Hinojosa governor on a promise to pause projects until laws cover water, power and property rights, which Bloomberg estimates exposes about 13 gigawatts of planned capacity. Analysts expect projects already under construction to continue, so the delays fall mostly on builds planned for 2028.

  2. 2

    OpenAI's agents found 948 working Linux root exploits in one run costing about $1.2 million

    OpenAI described a pipeline, built on a cyber model it has not yet released, that scans the Linux kernel for flaws, reproduces them, writes exploits and drafts patches. In a mid-October run it analyzed 2,308 candidate findings, reproduced 1,724 of them and produced 948 working exploits that let an ordinary local user take full control of a stock desktop Linux system. The run cost about $1.2 million in API credits, roughly $1,279 for each exploit, against the $20,000 to $70,000 that hacking contests have paid for a single bug of this kind. Human researchers are reporting the critical flaws first, with candidate patches, through OpenAI's Patch the Planet program. It follows Linus Torvalds saying the next Linux release is "full of fixes for issues found by AI."

  3. 3

    AI super PACs had a mixed midterm night, and OpenAI is already looking to 2028

    Money from both sides of the AI debate produced uneven results. Two industry-aligned groups, American Mission and Building the Future, spent $4.8 million backing South Carolina Senate candidate Darline Graham, who lost. American Mission is the Republican arm of Leading the Future, backed by OpenAI president Greg Brockman and Andreessen Horowitz, while its Democratic arm helped reelect Florida Rep. Jared Moskowitz. Public First, backed by Anthropic, lost several Republican races but did better with Democrats. Leading the Future says primaries matter more, and it spent about $27 million there. On November 4, OpenAI's global affairs chief Chris Lehane published an essay comparing AI's economic upheaval to the New Deal era and urging a "third path" before the 2028 election.

  4. 4

    Moonshot publishes Kimi K3.1 Pro's weights on schedule, under a license with revenue strings

    Moonshot AI put the weights for Kimi K3.1 Pro on Hugging Face on November 4, the date it promised when the model reached its apps and API in October. It is a mixture-of-experts (MoE) model with 2.8 trillion total parameters, 104 billion active per token, and a one-million-token context window. The license is not plain open source. A company that sells model access and earns more than $20 million over any 12 months needs a separate agreement with Moonshot, and products with more than 100 million monthly users must show the Kimi K3.1 Pro name, the same shape as Kimi K3's terms. Internal use is exempt, and unlike some rival licenses, it excludes no countries.

  5. 5

    Meta buys Conductor and OpenAI agrees to buy Pierre as coding agents reshape developer tools

    Meta confirmed it has acquired Conductor, a startup whose app runs teams of coding agents from labs like Anthropic and OpenAI in parallel workspaces, without disclosing terms. Conductor stopped taking new sign-ups on November 4. Its cloud service goes read-only on November 13 and shuts down on November 20, when user data starts being deleted, while the local app keeps getting fixes until early 2027. The same day, OpenAI agreed to buy Pierre Computer Company, whose open-source Trees and Diffs libraries already power file browsing and code review in the Codex app. Pierre's projects will stay open source, and the deal still needs regulatory approval.

  6. 6

    OpenAI's Images 2.6 brings animated GIFs to ChatGPT and generates images 50 percent faster

    OpenAI released Images 2.6 to every ChatGPT and Codex user on web, desktop and mobile. ChatGPT can now make looping GIFs, with controls to change playback speed and edit individual frames, and OpenAI says images generate 50 percent faster and keep characters and products more consistent across rounds of edits. Developers get two API models: GPT-Image-2.6 Flare, which can return a low-quality image in about six seconds, and GPT-Image-2.6 Sunburst for higher-quality work. OpenAI says people now make 4 billion images a week across its products, up from 3 billion in September.

  7. 7

    Uber says AI agents booked just four of its rides last quarter

    On its earnings call, Uber disclosed for the first time how much business arrives through outside AI agents, and the answer was almost none: four rides in the third quarter, and about 500 Uber Eats orders during Halloween week, against 320 million rides and deliveries. The figures cut against a popular worry that assistants like Meta's Muse or OpenAI's Dots will order on users' behalf and cut apps like Uber out. Chief executive Dara Khosrowshahi said Uber will work with agent makers but expects customers to keep using its own app, calling it "very, very early."

Get Top AI Stories by email

The day's most important AI news — free, daily, unsubscribe anytime.

Share
Spot a typo or have feedback?Share feedback

Sources

  1. 1.A thousand fewer paths to root — OpenAI · November 4, 2026
  2. 2.Meta acquires AI coding startup Conductor — TechCrunch · November 4, 2026
  3. 3.Introducing Images 2.6 — OpenAI · November 4, 2026
  4. 4.Big Tech claims voters' AI backlash is "anti-Trump" not "anti-tech" — Ars Technica · November 4, 2026
  5. 5.OpenAI is already trying to shape the 2028 election — TechCrunch · November 4, 2026
  6. 6.Conductor is joining Meta — Conductor · November 4, 2026
  7. 7.OpenAI to acquire Pierre Computer Company, Inc. — OpenAI · November 4, 2026
  8. 8.AI PACs spent millions trying to shape Congress — here's how they fared — TechCrunch · November 4, 2026
  9. 9.Uber only fulfilled 4 rides from AI agents last quarter — The Verge · November 4, 2026
  10. 10.Kimi K3.1 Pro — Moonshot AI · November 4, 2026
  11. 11.Data centers lost the midterms — The Verge · November 4, 2026
  12. 12.ChatGPT will now make you animated GIFs — The Verge · November 4, 2026

AI disclosure: Researched and drafted with AI; reviewed and edited by the AI Pro Playbook editorial team before publishing. Sources above link to original publishers.

🧭Recommended for you