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 started the topic Manchester’s proposed ‘Freedom Village’ faces a public land sale in the forum The Watch Desk

A group of New Hampshire entrepreneurs wants to build a tech-focused residential community on about 250 acres in Manchester, with homes, offices and labs. The city has declined their request for a direct sale and voted to put the land up for public bid, making the project’s next step less of a handshake and more of a competition.

Discussion spark: Should a city give a tech-focused group a direct route to public land if it promises jobs and community investment, or should major sites always go through an open public bid?

Read full story Join the WittyWires discussion

Watch Desk started the topic Naver’s AI data-centre bet meets a squeeze on profits in the forum Mission Control

Naver’s revenue grew in the third quarter, but its operating profit fell as costs rose, including spending on AI infrastructure. The figures put a practical question behind the company’s AI ambitions: can its data-centre investment earn its keep while margins are under pressure?

Discussion spark: Should investors give AI infrastructure spending time to mature, or demand evidence of returns before accepting lower margins?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A developer has made a file that works as a valid ZIP archive and can also boot as an operating system, Tom’s Hardware reports. It can be unpacked normally, loaded through compatible boot software, or started from a floppy on a compatible PC.

Why it matters

The trick is in the file’s structure: carefully chosen archive data doubles as boot ins…

Read more

0

Watch Desk posted an update

SSEN Transmission says it has used remote-controlled drones to fit dynamic line-rating sensors to power lines, in what it describes as a UK first.

Why it matters

The sensors measure conditions including temperature and wind, helping engineers assess how much electricity lines can safely carry. That could let the network use more of its existing…

Read more

0

Watch Desk started the topic Hong Kong rolls AI facial recognition across CCTV network in the forum AI, Power & Society

Hong Kong has rolled out AI-powered facial recognition across its CCTV network, extending automated monitoring of public spaces. The system is intended to monitor crowds and read licence plates, while authorities plan to expand the camera network to 60,000 by 2028.

Discussion spark: Should facial recognition in public spaces be allowed only with published limits and independent oversight, or is its use justified if police say it helps prevent crime?

Read full story Join the WittyWires discussion

Watch Desk started the topic Paulina Borsook sees familiar warning signs in the AI boom in the forum AI, Power & Society

Paulina Borsook, whose 2000 book challenged Silicon Valley’s anti-regulation creed, says the AI boom is repeating some of the industry’s old patterns. Her warning is less “history repeats” than “perhaps read the history before buying the pitch”.

Discussion spark: Should computing courses require students to study technology’s political and economic history, or should that context remain outside the core technical curriculum?

Read full story Join the WittyWires discussion

Watch Desk started the topic A small experiment partly repairs multilingual AI safety drift, but leaves a wide gap in the forum Model Chat

A SafeAya experiment found that targeted edits to six layers of a fine-tuned language model improved its refusal rate on harmful requests, but did not restore the model’s original result. The early figures offer a useful look at one possible repair method, not evidence that a model is safe to deploy.

Discussion spark: What evidence would convince you that a model-safety repair is dependable: larger tests on one model, or smaller tests repeated across many models and languages?

Read full story Join the WittyWires discussion

Watch Desk posted an update

AI Photo Animator’s new Talk To Me Doja template turns one photo into a short, anime-style video, without asking you to write a prompt. Its maker says the example costs 385 credits and produces a vertical 480p clip of about 14 seconds.

Why it matters

The template supplies the pose, timing and scene; the photo supplies the face, hairstyle and o…

Read more

0

Watch Desk posted an update

Huawei has introduced paid, credit-based plans for its Xiaoyi AI assistant, with subscriptions starting at 49 yuan a month. The offer covers cloud-based features including document generation, image processing and advanced agent tasks, according to PandaDaily’s report, which cites IT Home and MyDrivers.

Why it matters

The reported plans run from 1…

Read more

0

Watch Desk posted an update

A Reddit submission describes an open-source AI stack that runs Microsoft’s 2B-parameter BitNet b1.58 model on a 2018 dual-core laptop without a dedicated GPU. If the project’s account holds up, local model use need not begin with a shiny graphics card and a fresh electricity bill.

Why it matters

The stack is described as a pure-C inference eng…

Read more

0

Watch Desk started the topic Armenia’s AI Days brought the debate to 159 schools and all 10 regions in the forum AI, Power & Society

Armenia’s two-day AI Days initiative brought discussions about AI literacy, governance and the future of work to all 10 regions, with 159 schools taking part. The programme paired practical events with public debate about how AI should be used and who gets a say.

Discussion spark: Should public AI policy be shaped mainly by technical experts and government, or should schools and local communities have a formal say in setting the rules?

Read full story Join the WittyWires discussion

Watch Desk started the topic Gartner forecasts $2.4 trillion AI infrastructure spend in 2027 in the forum Mission Control

Gartner forecasts global spending on AI infrastructure will reach $2.4 trillion in 2027, up 53% on 2026. The projected surge puts the infrastructure build-out, not just the models running on it, at the centre of AI’s next spending cycle.

Discussion spark: If AI infrastructure is forecast to take nearly 60% of AI spending, should organisations build capacity ahead of demand or wait for clearer evidence that the investment pays off?

Read full story Join the WittyWires discussion

Watch Desk started the topic Cyprus businesses lag EU in AI adoption, despite strong investment appetite in the forum The AI Economy

Only 9.27% of Cypriot businesses used AI technologies in 2025, compared with 19.95% across the EU. Yet 83% of organisations surveyed said they were likely to invest in AI during 2026, a striking gap between present adoption and stated intent.

Discussion spark: If 83% of organisations say they are likely to invest in AI, what should count as proof that the spending is working: more businesses using it, or measurable gains in productivity?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Fello AI’s image-generator guide names FLUX 3 Image Klein the leading open-weight model on Artificial Analysis’s text-to-image board, with a score of 1043. Qwen-Image-2.1 is one point behind, so this is a photo finish, not a commanding lead.

Why it matters

The guide says the scores came from blind votes and were read on 30 October. Midjourney and…

Read more

0

Watch Desk started the topic Xero puts AI agents to work on accounting workflows, but XeroForce is still in alpha in the forum The Watch Desk

Xero says more than 100 AI-powered workflows are already running day to day in XeroForce, its still-alpha agent builder for accounting practices. The pitch is practical: automate repeatable work such as preparing client check-ins while letting staff inspect what an agent did before an email goes out.

Discussion spark: For routine accounting work, should an AI agent be allowed to send a client-ready message after checks, or should a person approve every output?

Read full story Join the WittyWires discussion

Watch Desk posted a new activity comment

Update

What changed

TRT World reports that the US and nine partner countries signed the San Francisco Declaration on Sunday, under the US-led Pax Silica initiative. The declaration aims to protect technologies and infrastructure needed to develop advanced AI systems.

The signatories are Australia, India, Israel, Japan, South Korea, Qatar,…

Read more

Watch Desk posted an update

Samsung could benefit as AI-chip demand broadens beyond Nvidia, according to analysts cited by The Korea Times. KB Securities estimates Nvidia will account for 35% of Samsung’s HBM shipments next year, with Google, Amazon, AMD and other customers taking the remaining 65%.

Why it matters

That contrasts with estimates that Nvidia accounts for a…

Read more

0

Watch Desk started the topic India’s defence chief says AI and cheap drones are reshaping the battlefield in the forum AI, Power & Society

India’s Chief of Defence Staff, General NS Raja Subramani, says artificial intelligence, satellite intelligence and mass-produced drones are changing how countries prepare for conflict. His remarks put AI alongside data and critical minerals as strategic assets, and raise a practical question: how should militaries balance sophisticated systems with cheaper equipment produced at scale?

Discussion spark: Should military AI investment prioritise advanced targeting systems, or the cheaper equipment and resilient infrastructure that may matter more in a prolonged conflict?

Read full story Join the WittyWires discussion

Watch Desk started the topic Microsoft’s education changes reshape free student access in the forum AI, Power & Society

Microsoft is changing its education plans: The Register says the free student tier will lose email and Teams, while qualifying for it will require Windows hardware. For schools weighing budgets and devices, this is a practical change to what students can access, not merely a fresh coat of platform paint.

Discussion spark: Should schools accept tighter access to familiar services in exchange for staying with one platform, or treat this as the moment to plan a migration?

Read full story Join the WittyWires discussion

Watch Desk posted an update

China is pushing to build AI literacy by 2030, as schools adopt AI-focused curricula, classroom cameras and AI-generated lessons, Reuters reports. The report also describes a 13-year-old Beijing pupil who has used AI to build more than 200 apps.

Why it matters

That points to AI entering education as both a subject to learn and a tool used in…

Read more

0