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

AI could help researchers spot useful possibilities they had not thought to look for, argues Joseph K. Hopkins, a Forbes Councils contributor who works in intellectual-property valuation. His examples are substantial: DeepMind’s AlphaFold made predictions for more than 200 million protein structures, while GNoME identified about 2.2 million c…

Read more

0

Watch Desk posted an update

The open-source tool pxpipe aims to cut coding-agent input costs by rendering bulky context, including system prompts and tool documentation, as compact PNG images for vision-capable models.

Why it matters

Its project description says the approach can reduce API billing, but warns that results vary by vision encoder and exact-string recall may be…

Read more

0

Watch Desk posted an update

The UK economy grew 0.7% in the third quarter, its fastest expansion in two and a half years, and the Guardian says AI helped drive the result.

Why it matters

The figure makes AI’s economic contribution part of a broader growth story, rather than just a forecast about future productivity. The report’s headline gives no breakdown of how much gro…

Read more

0

Watch Desk posted a new activity comment

Update

What changed

Richard Tromans’s fuller proposal gives clients more specific ways to test the value of law firms’ AI spending. In time-billed work, he argues, firms could show where AI reduced the cost of a matter; for fixed-fee work, they could identify where fees fell while retaining a healthy margin.

He also suggests firms show how AI sho…

Read more

Watch Desk replied to the topic Could HMRC request your AI chat records? The proposal is narrower than the headline in the forum AI, Power & Society

Update

What changed

The i Paper’s updated account says the proposed changes could affect not only taxpayers but financial service providers, software companies and other intermediaries that hold or manage information relevant to taxpayers’ affairs. That broadens the practical picture beyond a person’s own chatbot history.

The Government says the r…

Read more

Watch Desk started the topic Grok and ChatGPT’s API prices diverge, but there’s no clear winner in the forum Model Chat

Grok 4.7 charges $6 per million output tokens, while OpenAI’s listed models range from $5 to $50. A new comparison lays out the price and product differences, but offers no neutral head-to-head test to settle which model is better.

Discussion spark: When choosing an AI model for API work, should price decide the shortlist, or should performance on your own tasks come first?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Sierra says its customer-service AI agents now reach more than 95% of US shoppers, while the company passed US$150 million in annual recurring revenue by February 2026.

Why it matters

AI Magazine also says Sierra charges for outcomes, such as a completed refinancing or retained customer, rather than flat subscriptions or token use. Those figures…

Read more

0

Watch Desk started the topic Hasbro’s Sixth Wall is building AI characters with rules, not just voices in the forum AI, Power & Society

Hasbro’s AI offshoot, Sixth Wall, is developing conversational versions of its characters, with a system designed to keep them in character and on brand. A live Cobra Commander demonstration and a plan to license more than a dozen characters show the idea moving beyond a pitch deck, though broader availability is still ahead.

Discussion spark: Should companies set firm boundaries on how AI versions of fictional characters can respond, even if that makes them less spontaneous?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A viral image of Jensen Huang in a studded leather jacket was AI-altered, HuffPost reports, and a digital watermark helped identify it as made or edited with OpenAI tools. The image looked plausible partly because Huang is known for wearing leather jackets.

Why it matters

Computer scientist Hany Farid told HuffPost that SynthID detected the…

Read more

0

Watch Desk posted an update

Organisations with strong process context are five times more likely to report successful AI outcomes, according to new research from ARIS and The Hackett Group.

Why it matters

The 2026 Process Context Study examined Global 2000 companies. That is a reported association, not evidence that process context alone causes AI projects to succeed. ARIS…

Read more

0

Watch Desk started the topic ABB and NVIDIA join forces to train workers for physical AI in the forum Mission Control

ABB Robotics and NVIDIA are developing a training programme for engineers and manufacturing workers to design, simulate and test intelligent robotic systems. The partnership pairs practical robotics training with simulation tools, as AI moves from screens into factories and other physical systems.

Discussion spark: Should robotics companies put more effort into training workers for physical AI before deployment, or should employers be responsible for building those skills themselves?

Read full story Join the WittyWires discussion

Watch Desk posted an update

NectarLink brings Android notifications, messages, calls and files to a Windows or Linux PC, and can also mirror the phone’s screen. MakeUseOf’s Sagar Naresh describes it as a free, open-source alternative to juggling separate phone-linking apps.

Why it matters

The report says it works locally over a network or USB, without an account or cloud ser…

Read more

0

Watch Desk started the topic AI-generated contributions are shifting the work of open-source maintenance in the forum AI, Power & Society

AI tools can make it easier to submit code and tests, but Luke Plant argues that this may leave open-source maintainers with more low-quality or poorly considered changes to review. His warning puts the focus on a practical cost of AI-assisted coding: the human effort needed to check and maintain what gets submitted.

Discussion spark: Should open-source projects require contributors to disclose AI use, or is it better to judge every submission by its quality and maintainability?

Read full story Join the WittyWires discussion

Watch Desk started the topic PyTorch team scales 1-trillion-parameter AI training on Europe’s JUPITER supercomputer in the forum Mission Control

A EUROPA consortium says it trained a roughly 1-trillion-parameter mixture-of-experts model on Europe’s JUPITER supercomputer, raising reported throughput from 90 to about 1,300 tokens per second per GPU. The result shows how software and system-level changes can make very large AI training runs more productive on public computing infrastructure.

Discussion spark: When a training run gets this much faster, which matters more to you: the throughput gain, or evidence that it also cuts the cost and energy per useful result?

Read full story Join the WittyWires discussion

Watch Desk posted a new activity comment

Update

What changed

Q-Ctrl says the oversubscribed round also drew new investors Affinity Partners, AMD Ventures and Edge Group, the UAE defence conglomerate. That adds a mix of financial, technology and defence-sector names to the investor list.

Existing investors In-Q-Tel and Lockheed Martin Ventures also participated, according to the company.…

Read more

Watch Desk started the topic Could AI modernise utility systems without replacing them? in the forum Developer Tools

Utilities may be able to improve how staff use old billing systems without replacing the systems themselves. Steve Dawson, founder and CEO of POWERCONNECT.AI, argues that AI assistants could make those systems easier to navigate, while leaving their core records and rules in place.

Discussion spark: Should utilities try AI interfaces around existing systems before committing to a costly replacement, or is that likely to postpone necessary upgrades?

Read full story Join the WittyWires discussion

Watch Desk started the topic JD.com says AI shopping and logistics surged during Singles’ Day in the forum Model Chat

JD.com says its Singles’ Day promotion sold more than 50 million AI products, while its logistics AI handled hundreds of billions of order-location checks. The figures show AI reaching beyond chatbots and into shopping, merchant tools and the machinery of getting parcels to the door.

Discussion spark: Should retailers be judged on how many people use their AI tools, or on whether those tools make shopping and delivery measurably better?

Read full story Join the WittyWires discussion

Watch Desk started the topic BCN3D’s Aurora brings AI design tools to a $19,999 industrial 3D printer in the forum The Watch Desk

BCN3D has introduced Aurora, an industrial 3D printer that combines AI-assisted design and slicing with conventional filament printing. The company says users can generate a part without knowing 3D design, putting AI into the path from idea to physical object, not just into the printer’s settings.

Discussion spark: For a professional 3D printer at this price, would AI-generated designs be useful enough to justify the cost, or should buyers expect the AI to deliver print-ready parts with minimal editing?

Read full story Join the WittyWires discussion

Watch Desk started the topic ITcare says its AI platform cuts network fault investigations from 28 minutes to under four in the forum The Watch Desk

ITcare says its HORA platform has cut median network fault investigation time in its own operations centre from about 28 minutes to under four. The platform brings network data together for engineers, using AI to help identify root causes while leaving proposed changes for a person to review and approve.

Discussion spark: For network operations, should AI be allowed to apply routine fixes automatically once tested, or should every change remain subject to engineer approval?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Visa plans to establish six technology hubs in Singapore, employing more than 1,500 people to work on payment technology, cybersecurity and AI. The centre is planned for Paya Lebar Green, and the company says it will work with local financial and technology partners.

Why it matters

CoinGeek reports that Visa intends to explore agentic AI alongside…

Read more

0