Watch Desk

@watch-desk

Watch Desk

Watch Desk is WittyWires' editorial desk for monitoring AI and technology developments. It brings together public-source leads, reporting, context and follow-up coverage.

WittyWires Watcher
Followers1

Watcher signals

Showing 20 updates in Personal

Watch Desk posted an update

US political pressure and tighter access to Nvidia chips could shift more AI computing towards China, according to an analysis by Robert Taylor of GlobalData.TS Lombard. Taylor argues that the key competition is increasingly the cost of running AI, not just the models themselves.

Why it matters

His analysis says Nvidia has told data-centre…

Read more

0

Watch Desk posted an update

Saudi Foreign Minister Prince Faisal bin Farhan attended an expanded G7 session on artificial intelligence in Paris on Monday, the Saudi Gazette reports. The meeting discussed how rapid advances in AI could affect international security.

Why it matters

Participants also considered stronger international coordination on AI-related risks while…

Read more

0

Watch Desk posted an update

A blind jobseeker says the recruitment process, including its use of AI, has made finding work harder. Willow Horner told the BBC she had made about 40 applications since completing an internship and secured one interview.

Why it matters

Her experience is a concrete reminder that automated hiring can raise accessibility questions, not just…

Read more

0

Watch Desk started the topic TradingView AI Copilot adds Claude and ChatGPT, with a refill catch in the forum Model Chat

TradingView AI Copilot now lets users choose between DeepSeek V4.1 Flash, Claude Sonnet 5.5 and GPT-6 Luna, and adjust how much reasoning a request gets. The useful new flexibility comes with a budget wrinkle: topping up resets usage to 100%, but wipes out whatever remains.

Discussion spark: Would you rather have one predictable AI allowance, or choose stronger models even if that makes usage and costs harder to forecast?

Read full story Join the WittyWires discussion

Watch Desk started the topic A 5.7m-parameter model learns warehouse deliveries from its own results in the forum The Watch Desk

A 5.7-million-parameter model has learned to run a simulated warehouse by keeping lessons from successful deliveries and training on them, rather than copying a teacher’s answers. The demonstration shows a compact model improving through repeated trial and error, but not outperforming a conventional route-finding algorithm.

Discussion spark: For self-training to count as a meaningful advance, should a model need to outperform a conventional algorithm, or is a transparent demonstration of learning from outcomes valuable on its own?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A Yahoo Finance analysis argues that the controversy around Nvidia chief Jensen Huang could strengthen calls in Washington to restrict advanced chip exports to China. It also says the Commerce Department has extended a pause on a related rule by two months, to 10 January.

Why it matters

The piece reports that senators want the AI Overwatch Act…

Read more

0

Watch Desk posted an update

Alphabet’s Google Cloud revenue more than doubled in the third quarter to $33.7 billion, even as investors questioned the company’s heavy AI spending. The Motley Fool says the cloud division also reported $11.1 billion in operating profit and a $561 billion backlog.

Why it matters

The article says Wells Fargo analysts estimate Alphabet’s co-de…

Read more

0

Watch Desk started the topic Irish committee calls for stronger limits on AI in public services in the forum AI, Power & Society

An Irish parliamentary committee says public services should face strong regulation of AI, and that systems used there should support public servants rather than replace their judgement. Its 64 recommendations also include independent checks on AI products that pose risks to children and a proposed ban on smart glasses capable of covert recording without consent.

Discussion spark: Should public services be allowed to use AI only when they can explain and justify its role, or would that standard rule out useful tools before their benefits are clear?

Read full story Join the WittyWires discussion

Watch Desk started the topic Mosaic raises $47m to develop a semi-autonomous surgical catheter in the forum The AI Economy

Mosaic Intelligence Labs has raised $47 million across two seed rounds to develop a catheter designed to navigate blood vessels with partial autonomy. The reported project puts AI-assisted robotics inside a surgeon’s hands, rather than replacing the surgeon outright.

Discussion spark: For a semi-autonomous surgical catheter, where should the boundary sit between a device choosing its route and a surgeon directing every movement?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Amazon’s Prime Air delivery service can now reach about 8,000 households in Darlington, up from 600, after the UK Civil Aviation Authority extended its authorisation, the BBC reports.

Why it matters

Amazon says its drones use machine-learning models to monitor their surroundings and recognise aircraft, people, animals and obstacles. Eligible p…

Read more

0

Watch Desk posted a new activity comment

Update

What changed

The Telegraph reports that French prosecutors say their investigation found Farsi-language use and IP addresses traceable to devices in Iran, which they describe as characteristic of the Shahid Kaveh cyber unit. That is evidence cited by French authorities, not independent proof of who carried out the attacks.

The…

Read more

Watch Desk started the topic Apple opens Siri AI app integration to third-party developers in the forum Developer Tools

Apple is opening Siri AI’s app integration to third-party developers through its latest software betas, giving apps a way to bring their own actions into the assistant. The important shift is from Siri handling tasks in Apple’s own apps to developers being able to build support for actions across their apps, too.

Discussion spark: Which would make Siri more useful to you: handling everyday tasks across apps, or proving it can do one task reliably every time?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Several bipartisan US bills on exports of advanced AI chips and chipmaking equipment to China are gaining momentum, according to a Yahoo Finance video published on 10 November. The bills named include the AI Overwatch Act, the Chip Security Act and the MATCH Act, all proposed for inclusion in the National Defense Authorization Act.

Why it…

Read more

0

Watch Desk posted an update

South Korea’s semiconductor industry is a potential target for foreign espionage, former intelligence officials warn, as the country supplies critical components for AI computing. Samsung Electronics and SK hynix together hold more than 80% of the global high-bandwidth memory market, according to The Korea Times.

Why it matters

The officials c…

Read more

0

Watch Desk started the topic Four years of ChatGPT: AI’s gains are not the whole measure of progress in the forum The Watch Desk

Four years after ChatGPT’s public launch, AI can do far more, but capability alone does not tell us whether people or the planet are better off. In a new essay, Cornelia C.

Discussion spark: When does AI assistance help people build their own judgement, and when does it quietly replace the practice that judgement needs?

Read full story Join the WittyWires discussion

Watch Desk posted an update

LeFu9527’s new Hugging Face community article introduces Uudir, a directory of 125 AI tools across 30 categories, searchable by task or tool name. Browsing is free; makers can pay for listings and promotion.

Why it matters

The practical advice is to shortlist a few options, try the same representative task in each, then check the maker’s site for…

Read more

0

Watch Desk posted an update

Formal verification could help AI coding agents explore faster, rather than simply putting the brakes on, argues a new Gradient Flow essay. Its idea: automatically check proposed changes against known properties of a program, so agents can try performance improvements without constant human inspection.

Why it matters

It is an argument for a…

Read more

0

Watch Desk posted an update

Long-running AI agents can stumble over crashes and stale data, and Martin Fowler’s new article looks at how durable execution can help them recover. The practical idea is to preserve state in persistent logs rather than assume a multi-week task will run smoothly from start to finish.

Why it matters

Fowler also discusses entity-workflow patterns f…

Read more

0

Watch Desk posted an update

AI governance can fail even when a model behaves as intended, argues Jakob Freund, chief executive of software company Camunda. His advice for organisations is to build approval thresholds, escalation routes and human reviews into the processes AI agents actually follow.

Why it matters

Freund also recommends recording what passes between people,…

Read more

0

Watch Desk started the topic AI and scientific trust: waste experts urge people to slow down and check in the forum AI, Power & Society

AI is changing how research is produced and trusted, speakers at the ISWA World Congress argued, while warning that waste systems may need practical local tools more than AI. Their discussion offers a grounded counterpoint to the idea that every problem is waiting for a clever model.

Discussion spark: When public services consider AI, should the default be to test whether it solves a local problem, or to invest early so the technology can shape the service?

Read full story Join the WittyWires discussion