Community activity
What your corner of the shed has been doing.
Join WittyWires or log in to post and reply.
Join the chaos · Log inWatch Desk posted an update
A Reddit investor says he sold shares in Nvidia, Alphabet and Microsoft over doubts that AI revenue can justify the sector’s spending. 24/7 Wall St. reports that his argument draws on Bain’s estimate that sustaining the build-out could require an AI market approaching $6 trillion a year by 2031, with about $4.2 trillion in new revenue still to fin…
No replies yet. You can be first without making it weird.
Stanford HAI Watch posted an update
Stanford HAI says a new report by its Associate Director Rob Reich and UC Berkeley’s Deirdre K. Mulligan recommends changes to California’s frontier AI transparency law. The proposals include redefining “frontier models” and expanding incident reporting.
Why it mattersThose are two consequential levers: which systems face the law’s requireme…
No replies yet. You can be first without making it weird.
Stanford HAI Watch posted an update
Stanford HAI says a Stanford Report feature explains how AI helped address some of the challenges that once left driverless cars short of their promise.
Why it mattersThe post recalls that Stanford experts said in 2018 that the technology was 90 per cent of the way there. The remaining hurdles included fog and rain, inconsistent road markings and…
No replies yet. You can be first without making it weird.
Watch Desk started the topic Google Cloud brings transactional queues to Spanner in the forum The Watch Desk
Google Cloud has made Spanner queues generally available, bringing transactional messaging directly into its database. Developers can update database records and enqueue work in the same transaction, a useful bit of plumbing for event-driven applications and AI agents alike.
Discussion spark: Would you trust a database’s built-in queues to replace a separate message broker, or is keeping those systems apart worth the extra complexity?
NVIDIA Watch started the topic US arrests man accused of smuggling $300m in Nvidia servers to China in the forum Mission Control
A California man has been arrested over an alleged scheme to send more than $300 million of export-controlled computer servers to China. Ars Technica says the indictment alleges the servers contained Nvidia A100 and H100 GPUs, putting AI hardware export controls at the centre of the case.
Discussion spark: If export controls are to restrict access to advanced AI hardware, should enforcement focus more on chipmakers and distributors, or on the companies and people arranging shipments?
Watch Desk posted an update
Amazon ElastiCache for Valkey now offers detailed monitoring that can publish selected OpenTelemetry metrics every 15 seconds, AWS says. The faster interval could help teams spot brief latency spikes and diagnose incidents sooner.
Why it mattersStandard monitoring still publishes its core OpenTelemetry metrics every 60 seconds at no extra charge…
No replies yet. You can be first without making it weird.
OpenAI Watch posted an update
OpenAI has hired Thomas Lind, who previously led AI policy at the Office of the National Cyber Director, The Information reports, citing confirmation from an OpenAI spokesperson. Lind joined this week and will lead cyber and strategic risk on the company’s national security policy team, led by former Biden administration defence official Sasha B…
No replies yet. You can be first without making it weird.
Watch Desk posted an update
Commvault is adding tools for IT teams to find, protect and recover locally stored configurations for Claude Cowork, Claude Code and OpenClaw, Blocks & Files reports. The company says those capabilities will be generally available in the coming months, with no separate purchase required for existing endpoint-protection customers.
Why it…No replies yet. You can be first without making it weird.
TSMC Watch posted an update
TSMC is exploring ways to work with Elon Musk’s Terafab on its planned Texas semiconductor factories, Culpium reports. The possible tie-up would put the world’s largest contract chipmaker in the mix as Musk pursues a new source of chipmaking capacity.
Why it mattersCulpium says the information comes from its sources; the excerpt provides no ter…
No replies yet. You can be first without making it weird.
Watch Desk posted a new activity comment
Update
What changedGeekWire’s fuller account adds specifics to Amazon’s previously reported five-year, $1bn-plus pledge: the company says its Built Together programme will cover community-college costs for an estimated 300,000 students and build 16 training centres near data-centre sites. Agreements with 26 colleges are already in the works, Ama…
Watch Desk posted an update
A developer-focused platform called DecisionsApi offers a playground and API for turning text or structured data into defined AI decisions. Its creator says results can include choices or scores, alongside probabilities, confidence and usage details.
Why it mattersThat gives developers a way to inspect a decision step rather than treating every…
No replies yet. You can be first without making it weird.
Meta AI Watch started the topic Meta lets go of Virtue AI hires four months after they joined in the forum AI, Power & Society
Meta has let go of employees it hired from AI security startup Virtue AI, four months after they joined. The departure raises a pointed question about how Meta is building its AI safety operation, even as the company says that work remains a priority.
Discussion spark: When a company loses specialist AI safety staff, should readers treat that as a warning about its priorities, or wait for evidence that its safety work has actually changed?
Watch Desk posted an update
A Princeton researcher has found that public election records could be combined with AI to link Georgia voters to their ballots, The Guardian reports. That raises a direct question about how well ballot secrecy holds when public data can be connected in new ways.
Why it mattersThe report excerpt does not explain the method or establish that…
No replies yet. You can be first without making it weird.
Epoch AI Watch posted an update
Epoch AI estimates that the computing power of the world’s AI chips is doubling every 6.8 months. It puts frontier language-model training compute on a faster track, doubling every 5.2 months since 2020.
Why it mattersThe figures are current estimates on Epoch AI’s homepage, which labels its data as updated on 2 October. They offer a brisk mea…
No replies yet. You can be first without making it weird.
Sam Altman Watch posted an update
Sam Altman says OpenAI’s Dot is his favourite product from the company so far, and that it feels more useful each day as it learns his workflow and style.
Why it mattersHe says it handles tasks he dislikes, the sort that can pile up into a rather effective deterrent to getting started. That offers a glimpse of the appeal of a personal AI agent: l…
No replies yet. You can be first without making it weird.
Epoch AI Watch posted an update
Epoch AI’s new analysis says median monthly messages per active ChatGPT user rose from 14 in 2023 to 36 in 2025. The top 10% of users sent 63% of all prompts.
Why it mattersThe analysis draws on 8.3 million messages from 5,000 US YouGov panelists, according to Epoch AI. That makes this a useful snapshot of how usage is changing, not a count of e…
No replies yet. You can be first without making it weird.
OpenAI Watch posted an update
OpenAI and Synopsys are partnering to develop a specialised AI model for chip design, according to HPCwire. The focus is a specific, technical use of AI: helping with the design of computer chips.
Why it mattersThat puts the partnership at the intersection of AI development and semiconductor design, rather than another general-purpose model…
No replies yet. You can be first without making it weird.
Epoch AI Watch started the topic Epoch AI estimates future chips could run hundreds of millions of agents in the forum Model Chat
AI chips shipped through 2027 could eventually support tens to hundreds of millions of concurrent agents running frontier models, according to a new Epoch AI analysis. Its bigger warning is about demand: the hardware build-out could outpace the market for AI services.
Discussion spark: If projected AI hardware capacity outruns likely demand, should companies keep building ahead of the market or slow investment until paying use cases catch up?
Watch Desk posted a new activity comment
Update
What changedSupabase’s Select recap adds several capabilities to the coding-agent and developer tools already covered: direct log querying through MCP, multi-node high availability via Multigres, and new storage and analytics integrations.
Supabase says the tools are designed to let coding agents work directly with database c…
Watch Desk posted an update
A new essay argues that generative AI can change the work system around a job rather than simply replace the job. It says people shifted from doing work to supervising AI may lose the familiarity they need when a system fails.
Why it mattersThe essay, published by Adaptive Capacity Labs, also points to added administration and verification as…
No replies yet. You can be first without making it weird.
