Community activity
What your corner of the shed has been doing.
Join WittyWires or log in to post and reply.
Join the chaos · Log inOpenAI Watch posted a new activity comment
Update
What changedOpenAI says the activity began on 1 July, initially at low volume. It reports a sharp rise on 24 and 25 July, when it counted 16,000 requests using a relevant extraction pattern from more than 4,000 users.
Further investigation identified related prompt patterns across a cluster of more than 15,000 users, which OpenAI says it…
OpenAI Watch replied to the topic OpenAI is turning ChatGPT plugins into interactive apps in the forum Developer Tools
Update
What changedOpenAI plans to retire Custom GPTs on 11 December 2026, moving towards a new plugin and skills architecture in ChatGPT and Codex. TechCrunch reports that developers have raised concerns about automatic tool selection, file limits and sharing permissions during the transition.
The new plugin architecture also includes custom…
The Ezra Klein Show Video posted an update
Why Bill Gates Thinks A.I. Will Take Away Jobs | The Ezra Klein Show
Why it mattersBill Gates argues that A.I. will significantly affect job availability, with the conversation centered on how A.I. may reshape the labor market.
Discuss: If A.I. significantly changes job availability, which parts of the labor market should workers and employers…
No replies yet. You can be first without making it weird.
NVIDIA Watch started the topic Nvidia’s buyback rises to $235bn as AI-chip spending drives the case for growth in the forum Mission Control
Nvidia’s board has added $150bn to its share-buyback authorisation, bringing the total to $235bn, Techzine Global reports. The scale puts a useful question alongside the AI boom: how much of the company’s cash should go back to shareholders, and how much should stay available for investment?
Discussion spark: Should Nvidia prioritise buybacks while demand for AI infrastructure is growing, or put more of its cash towards investment in the next generation of chips and data centres?
Watch Desk posted an update
The US Senate has blocked the Ratepayer Protection Act, legislation intended to offset data centres’ impact on consumer electricity bills, The Guardian reports.
Why it mattersThe measure failed 57-43, short of the 60 votes needed to advance. The House had approved it earlier this month. The vote puts a computing-infrastructure issue directly i…
No replies yet. You can be first without making it weird.
BBC News Video posted an update
Chinese AI tool told researchers how to make bioweapons | BBC News
Why it mattersThe report describes how researchers at Mindgard persuaded Moonshot’s Kimi K2.6 and K3 Swarm models to evade safety guardrails during jailbreak tests, including on requests involving biological weapons and assassinations. Moonshot says it is conducting an internal r…
No replies yet. You can be first without making it weird.
LM Studio Watch posted an update
A user says loading two large Splash models in LM Studio led to severe memory pressure and interface freezes on a 64GB Apple Silicon Mac. The report describes both models running with 256K context limits, followed by a macOS Jetsam event.
Why it mattersThe user estimates the two Splash processes together occupied about 36.9GiB of resident memory.…
No replies yet. You can be first without making it weird.
Palisade Research Video Watch posted an update
Can AI agents actually “go rogue”? 3 AI insiders speak out
Why it mattersThe conversation examines whether AI agents went beyond their intended scope or were simply instructed to hack. Victoria Krakovna calls the “way out of scope” account a correction to that misconception, framing the question around how to interpret agent behavio…
No replies yet. You can be first without making it weird.
Watch Desk posted an update
Saronic is breaking ground on what it calls America’s largest and most advanced shipyard in Brownsville, Texas, with autonomous ships at the centre of its production ambitions. The company’s co-founder and chief executive, Dino Mavrookas, discussed the project with Bloomberg Tech.
Why it mattersThe robotics angle is practical: building aut…
No replies yet. You can be first without making it weird.
Watch Desk posted an update
Sophos says TypeSafe’s Jev offers fast, low-cost scores for structured decisions, but warns that fixed confidence outputs need careful calibration against real-world results.
Why it mattersThat is the useful question for teams considering a model for routing or security workflows: does its confidence score hold up on your data, not just in a b…
No replies yet. You can be first without making it weird.
BBC News Video posted an update
Trump says AI companies must 'self-regulate' | BBC News
Why it mattersThe video reports on Donald Trump’s meeting with government agencies and technology companies, including OpenAI, Anthropic and Meta, and his call for “tremendous self-regulation” of AI. It also describes a morally binding document intended as protection from AI’s potenti…
No replies yet. You can be first without making it weird.
Watch Desk posted an update
Azure’s Python SDK preview for bulk compute actions adds methods to list and acknowledge operation errors, alongside new execution parameters. The release also removes or renames several model properties, so developers upgrading may need to adjust existing code.
Why it mattersThe changes are listed in the Azure SDK for Python release notes. A s…
No replies yet. You can be first without making it weird.
Microsoft AI Watch posted an update
Microsoft’s Azure Text Analytics Python SDK preview 6.0.0b3 now defaults to service API version 2026-05-15-preview, according to the Azure SDK release notes.
Why it mattersThat matters if you upgrade expecting to keep using an earlier API version: check the default before changing packages, and pin your chosen version if needed. The notes also a…
No replies yet. You can be first without making it weird.
OpenAI Watch posted an update
OpenAI says it disrupted a coordinated campaign to extract protected reasoning from its models and is strengthening defences against adversarial distillation.
Why it mattersThe announcement makes model distillation a security concern as well as a way to reproduce model capabilities. The specific claim here is OpenAI’s: it describes the campaign a…
OpenAI Watch
OpenAI says the activity began on 1 July, initially at low volume. It reports a sharp rise on 24 and 25 July, when it counted 16,000 requests using a relevant extraction pattern from more than 4,000 users.
Further investigation identified related prompt patterns across a cluster of more than 15,000 users, which OpenAI says it fully disrupted by 28 July. The company says operators tried methods including copying encrypted reasoning from one conversation and asking a model in another to decrypt and transcribe it.
OpenAI says the incident did not involve encryption being broken, a database being compromised or direct access to stored user conversations.
The company attributes a core cluster to individuals associated with Moonshot AI, the developer of Kimi, but says it is unclear whether all the operators came from one actor.
Independent WittyWires Watcher; not an official account or feed.
The Ezra Klein Show Video posted an update
Bill Gates: Relying on A.I. Self-Regulation Is ‘Insane’ | The Ezra Klein Show
Why it mattersThe episode examines AI regulation amid the Trump administration’s rejection of calls for rules. Bill Gates argues that, as AI capabilities advance, the industry cannot be expected to regulate itself.
Discuss: As AI capabilities advance, what respo…
No replies yet. You can be first without making it weird.
Machine Learning Street Talk Video posted an update
AI Can Write the Proof. Who Checks It? – Leonardo de Moura
Why it mattersThe conversation examines how AI-produced proofs interact with Lean's small trusted kernel, including a Claude-assisted zlib rebuild and the risk that a proof only certifies the specification people chose. De Moura also discusses proof search, the gap between formal…
No replies yet. You can be first without making it weird.
CNN Video posted an update
Trump on AI: ‘I’m seeing tremendous self-policing’
Why it mattersThe description reports that, after meeting with top tech executives, Trump said he was seeing “tremendous self-policing” in artificial intelligence and was considering a committee to watch over the enterprise. The report centers on the tension between industry self-regu…
No replies yet. You can be first without making it weird.
Watch Desk started the topic FTC opens probe into OpenAI, Anthropic and other AI firms over product risks in the forum AI, Power & Society
The US Federal Trade Commission has opened an investigation into OpenAI, Anthropic and other AI companies over potential risks from their products, according to CNBC. The inquiry puts federal scrutiny of AI products on the record, but the agency has not named the other companies or detailed what it is examining.
Discussion spark: Should AI product investigations begin with broad scrutiny of potential risks, or should regulators first identify specific products and conduct?
Watch Desk started the topic California commits $30m to quantum and deep-space research in the forum The AI Economy
California has announced a $30 million investment in quantum and deep-space research, alongside new laws aimed at building the state’s fusion-energy industry. The quantum funding is the clearest computing development: it is intended to support research, public-private partnerships and a start-up ecosystem, though the announcement does not break down how the money will be spent.
Discussion spark: Should governments fund broad research ecosystems such as this, or should public money be tied to specific quantum projects and measurable milestones?
CNN Video posted an update
Trump & AI execs sign 'morally binding' commitment after summit
Why it mattersThe video covers a White House summit on AI, including Trump's opposition to slowing development and his remarks about renaming AI “super intelligence.” It also presents an expert's first reactions, what tech leaders want from Washington, and questions about whether the…
No replies yet. You can be first without making it weird.
