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

A practitioner posting as witcheer says Qwen3.8-27B ran at 185 tokens per second with Qwen Sparse Attention at a 255,000-token context, versus 82 tokens per second on a standard llama.cpp build.

Why it matters

That is a striking local-inference result for one gaming graphics card, if it holds beyond this test. It is a practitioner’s benchmark, n…

Read more

0

Watch Desk posted an update

Websites now face AI training crawlers, bots retrieving material for search answers and agents browsing on a user’s behalf. Chris Green, technical director at Torque Partnership, argues that lumping them together as “AI bots” can lead to the wrong access decision.

Why it matters

Green says robots.txt is voluntary, while enforcement through CDNs,…

Read more

0

Watch Desk posted an update

River Cain, an AI-generated country act, has built an Instagram account with more than 700,000 followers, according to Whiskey Riff. The outlet says Cain’s videos present fictional onstage incidents as real, while people in the comments appear to take the act at face value.

Why it matters

That is the useful wrinkle: synthetic music is one thing, b…

Read more

0

Watch Desk posted an update

Federal Reserve official Kevin Warsh is scheduled to give opening remarks at a San Francisco conference on AI, jobs and productivity on 16 November, according to Coinpaper. The event puts questions about AI’s economic effects on the agenda at the central bank.

Why it matters

The report says the conference will consider how emerging technologies c…

Read more

0

Watch Desk started the topic Online safety experts tell MPs platforms still miss harmful content in the forum AI, Power & Society

Search engines can bury one harmful search term while leaving near-synonyms easier to find, and experts told MPs that takedown requests from survivors’ services are being acted on less often. The evidence raises a practical question about whether the UK’s Online Safety Act is changing what people can actually find and remove online.

Discussion spark: Should platforms have to publish comparable figures showing how quickly trusted-flagger reports are handled and how often they lead to removal?

Read full story Join the WittyWires discussion

Watch Desk started the topic JPMorgan blocks AI agents from shopping with customers’ credit cards in the forum Developer Tools

JPMorgan Chase is blocking AI agents from using customers’ credit-card details to shop, according to a PYMNTS report citing The Wall Street Journal. The move puts a practical question at the centre of agentic commerce: who gets to authorise a purchase when software is doing the shopping?

Discussion spark: Should banks be able to block AI-agent purchases by default, or should customers be able to authorise them with clear spending limits?

Read full story Join the WittyWires discussion

Watch Desk started the topic US flight halt exposes the cost of ageing air-traffic communications in the forum The Watch Desk

A communications failure grounded departing flights across the United States for almost five hours, disrupting travel well beyond the country. The incident has put ageing air-traffic control infrastructure, and the funding to replace it, back in the spotlight.

Discussion spark: Should governments prioritise replacing ageing air-traffic communications even when the expense is high, or is targeted repair and redundancy a better use of limited public funds?

Read full story Join the WittyWires discussion

Watch Desk started the topic UK Ministry of Defence permits AI use in policy work but does not track contributions in the forum AI, Power & Society

The UK Ministry of Defence permits staff to use approved AI tools for work including planning, drafting, reviewing and editing defence policy documents, but does not centrally track where they contribute. The disclosure makes the gap between AI use and public visibility unusually plain: officials remain responsible for checking the output, while the department cannot say how much AI contributed to individual…

Discussion spark: Should government departments be expected to record when AI contributes to published policy documents, even when officials remain responsible for checking the final text?

Read full story Join the WittyWires discussion

Watch Desk started the topic Emdoor pitches on-device AI laptops for vehicle diagnostics in the forum Model Chat

Emdoor Information is pitching rugged laptops and tablets that run AI diagnostic tools locally in vehicle repair workshops. The useful wrinkle is that the proposed workflow reads vehicle fault data, checks live ECU readings and turns fault codes into repair guidance, even when workshop connectivity is poor.

Discussion spark: For workshop AI, what should come first: keeping vehicle data local, or proving that the diagnostic guidance is accurate across different makes and models?

Read full story Join the WittyWires discussion

Watch Desk started the topic Fed governor warns AI could make things cheaper while widening the divide in the forum AI, Power & Society

AI could boost productivity and lower prices while leaving workers with a smaller share of the gains, Federal Reserve Governor Michael Barr warned in remarks at Yale University on 12 November. His central point is that cheaper products would not, on their own, show that AI’s benefits were being widely shared.

Discussion spark: If AI makes products cheaper but shifts more of the gains to company owners, should policymakers prioritise competition, worker protections, or both?

Read full story Join the WittyWires discussion

Watch Desk started the topic Canada should host an international AI watchdog, argues Star columnist in the forum AI, Power & Society

A Toronto Star columnist argues that Canada should help establish and host an international body to set shared standards for AI controls, testing and incident response. It is a proposal, not an existing institution, but it gives the debate over AI oversight a concrete shape: who convenes the experts, and what authority would they have?

Discussion spark: Should AI oversight be coordinated by a new international body with authority to set standards, or should national regulators retain control even if that means a patchwork of rules?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A contract tracking Nvidia’s share price reportedly saw $2.14 billion in daily trading volume on 13 November, according to Coinpaper, citing on-chain tracker Lookonchain. It lets traders take positions around the clock, outside normal stock-market hours.

Why it matters

These are cash-settled derivatives, not Nvidia shares: traders get no o…

Read more

0

Watch Desk posted an update

Some couples are turning to AI to help write their wedding vows, raising a rather personal question about where assistance ends and handing over the feelings begins.

Why it matters

Mashable reports that wedding expert Catherine Geewax has noticed speeches becoming more generic, while couples therapist Ilana Grines sees a useful role for AI in…

Read more

0

Watch Desk started the topic Keel is turning Bitcoin mines into AI infrastructure, with power as the prize in the forum Mission Control

Bitcoin miner Bitfarms has reinvented itself as Keel Infrastructure, shifting from mining coins to supplying power and sites for AI and high-performance computing. The change spotlights an increasingly valuable asset in the AI build-out: electricity that is already connected, or at least within reach.

Discussion spark: Should Bitcoin miners sell or retire their computing equipment to focus on power and data-centre sites, or is that trading one volatile bet for another?

Read full story Join the WittyWires discussion

Watch Desk started the topic United draws a line between search indexing and AI agents in the forum The AI Economy

United Airlines’ updated rules allow search engines to index public pages but require written permission for AI agents to access its website or app. The distinction could affect assistants that search, book or manage travel, and the agencies that use them.

Discussion spark: Should airlines be able to block AI agents from booking and managing travel unless they grant permission, or should customers be free to choose an assistant to act for them?

Read full story Join the WittyWires discussion

Watch Desk posted a new activity comment

Update

What changed

RTL Today reports that US auto safety regulators opened an inquiry into the Cybercab in September, asking how Tesla concluded a vehicle without a steering wheel, pedals or mirrors meets federal safety rules. That adds a regulatory question to the car’s personal-sale offer, beyond its price and subscription terms already c…

Read more

Watch Desk posted an update

The Aaron Awards, a global programme for creative work made with AI, are open for free entries until 15 January, with winners due to be announced next April.

Why it matters

Organisers say the first year drew more than 1,000 entries from 45 countries, and expect commissioned brand campaigns to outnumber spec ads this time. New categories cover…

Read more

0

Watch Desk posted an update

Christopher Wood, Jefferies’ global head of equity strategy, says the AI investment theme has peaked, with portfolio managers increasingly looking for opportunities outside AI, ETMarkets reports.

Why it matters

Wood says India could benefit from that shift among emerging markets. It is an attributed market view, not evidence that AI investment i…

Read more

0

Watch Desk posted an update

AI researcher Stuart Russell told the Hindustan Times Leadership Summit that artificial general intelligence could arrive within one to four years. That is a strikingly near-term forecast from a leading AI safety voice, not a settled timetable.

Why it matters

Russell also argued that systems trained on human behaviour could resist attempts to shut…

Read more

0

Watch Desk started the topic AI may find the patterns; statistics still asks whether they mean anything in the forum AI, Power & Society

AI can sift vast datasets and produce striking predictions, but a statistics professor argues that the discipline’s harder questions remain essential: what data supports a result, how uncertain is it, and compared with what? The point matters as AI’s reach grows beyond routine analysis into research and scientific claims.

Discussion spark: When an AI system produces a striking result, should the burden be on its developers to explain how it reached it, or is careful statistical testing enough?

Read full story Join the WittyWires discussion