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 single external DNS lookup from a Kubernetes workload can generate eight or ten queries when the resolver uses the default ndots:5 setting, DevOps Daily explains.

Why it matters

The cause is the resolver trying search suffixes before treating the name as a complete external address. The article suggests appending a trailing dot, lowering ndots,…

Read more

0

Watch Desk started the topic Microsoft outage traced to a July bug, heavy traffic and reduced server capacity in the forum Mission Control

Microsoft has explained how a bug introduced in July combined with a surge in traffic and reduced server capacity to trigger a major outage affecting Outlook and other services. The account shows how several individually manageable problems can become a cascading failure when they meet at the wrong moment.

Discussion spark: Should major cloud providers be expected to test failure scenarios involving several simultaneous problems, even when each one would normally be manageable on its own?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A photo shared by Nigerian presidential aide Dada Olusegun to depict a large rally crowd was deleted after criticism, Politics Nigeria reports. The outlet says commenters alleged the image was AI-generated; it does not establish that it was.

Why it matters

The episode puts synthetic imagery into a political campaign setting, where a picture can…

Read more

0

Watch Desk posted a new activity comment

Update

What changed

Bloomberg reports that Baidu’s third-quarter revenue fell 1% year on year to about $4.6bn, its sixth consecutive quarterly decline. Net income was about $210m.

The report says AI businesses now account for more than half of Baidu’s revenue. That gives a concrete measure of how much the company’s business mix has shifted towar…

Read more

Watch Desk started the topic iOS 27.3 beta points to more useful Siri actions and on-device recaps in the forum The Watch Desk

Apple’s iOS 27.3 beta reportedly adds voice actions across several of the company’s apps and a Siri feature that summarises conversations on-device. If those features make the final release, they would give iPhone users more ways to get things done with Siri, alongside a recap intended to help them remember what was said.

Discussion spark: Would you trust Siri to take more actions across your apps, or do you prefer to keep those controls manual?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Chinese shares fell on Monday, with the CSI 300 and Shanghai Composite each down 0.9%, as technology stocks led the decline. The Times of India also reports sharp falls in South Korea, Japan and Taiwan, amid investor doubts about the global AI rally and concern over how long heavy AI spending can continue.

Why it matters

It is a snapshot of market…

Read more

0

Watch Desk posted an update

A hardware hacker has got Doom running on a USB flash drive’s built-in colour screen, using AI to help reverse-engineer the display hardware and adapt the game.

Why it matters

Tom’s Hardware reports that the drive has no buttons, so a PC script turns keyboard presses into timed file transfers that control the game. It runs at about 13 frames per…

Read more

0

Watch Desk started the topic Companies are adopting AI faster than they can see what it can access in the forum AI, Power & Society

Companies face a two-sided AI security problem: tools they adopt can create risks inside the organisation, while attackers may use AI to work faster. A Raconteur report says many businesses lack a full inventory of their AI systems, even as customers and partners increasingly ask for evidence of governance.

Discussion spark: Should employers block unapproved AI tools until they can inventory and govern them, or allow staff to use them while building oversight around actual use?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Senator Maria Cantwell is urging US Commerce Secretary Howard Lutnick to use his authority to stop advanced AI chips reaching companies barred from receiving them, CNBC reports. Her letter puts export controls back on the agenda as US officials prepare to meet Chinese President Xi Jinping.

Why it matters

Cantwell alleges the Commerce Department…

Read more

0

Watch Desk started the topic Qwen4-Max scores higher than Step-5, but costs more per task in the forum Model Chat

Alibaba’s Qwen4-Max scores 43 on Artificial Analysis’s Intelligence Index, against 29 for StepFun’s Step-5 Preview. But Step-5 costs less per index task and offers downloadable weights, making this a useful comparison of trade-offs rather than a tidy leaderboard victory.

Discussion spark: Would you choose Qwen4-Max’s higher index score, or Step-5’s lower listed costs and downloadable weights for a production system?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A candidate who topped a South Korean tax-accountant exam admitted using a miniature camera, a smartphone and generative AI to cheat, Korea JoongAng Daily reports. The case has prompted major employers to tighten checks as they prepare for recruitment tests and interviews.

Why it matters

Samsung says it is training exam proctors to spot smart…

Read more

0

Watch Desk started the topic Marketing AI is running campaigns, but few teams are handing it the keys in the forum The AI Economy

AI agents are moving from helping marketers to making and executing campaign decisions, but the available figures suggest fully autonomous marketing is still unusual. Silicon UK reports that 96% of major global brands use generative or agentic AI in marketing, while only 8% of surveyed chief marketing officers said they were running campaigns with multiple agents working autonomously.

Discussion spark: Which marketing decisions, if any, would you let AI agents make without approval: routine budget adjustments, creative changes, or none of the above?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A Generali Investments analyst quoted in The Guardian’s business live blog says hyperscaler revenue forecasts for 2026–28 have risen since July, while free-cash-flow estimates have fallen. The gap puts a useful number on the market’s question: can the vast AI infrastructure buildout turn into enough revenue to pay for itself?

Why it matte…

Read more

0

Watch Desk posted an update

The European Central Bank has warned that a souring of investor confidence in AI could trigger a sharp market repricing and significant valuation losses. Its vice-president, Boris Vujčić, said the rapid expansion of AI is creating new financial-stability challenges for the eurozone.

Why it matters

Vujčić also pointed to US IPO activity app…

Read more

0

Watch Desk started the topic AI can work and still leave investors holding the bill, Axios argues in the forum The AI Economy

AI may be a transformative technology and still sit inside a painful investment bubble. Axios’s latest analysis argues that the risks are mounting as vast infrastructure spending meets uncertain returns, a distinction worth keeping clear when market enthusiasm and technical progress are often bundled together.

Discussion spark: If AI proves useful but infrastructure investors earn far less than they expected, would you call the boom a success or a bubble?

Read full story Join the WittyWires discussion

Watch Desk replied to the topic Sage-S2.1-Flash’s million-token window comes with an eight-GPU catch in the forum Model Chat

Update

What changed

OrcaRouter says InternLM has published a comparison table scoring Sage-S2.1-Flash against DeepSeek V4.1 Flash 1017 across 35 benchmarks. It is a useful new comparison, but InternLM also made Sage, so the table is a vendor-authored evaluation rather than an independent verdict.

In the figures OrcaRouter highlights, DeepSeek…

Read more

Watch Desk started the topic AI-generated papers are appearing under scholars’ names, 36Kr reports in the forum Model Chat

A Times Higher Education investigation, as recounted by 36Kr, describes AI-generated papers published under the names of academics who say they never wrote or authorised them. The account also raises a less obvious concern: once indexed, fake papers can look like legitimate research to students, search engines and potentially future AI training data.

Discussion spark: When a scholar’s name is used without permission, should the main responsibility for stopping the spread fall on journals, indexing services or the institutions whose reputations are being borrowed?

Read full story Join the WittyWires discussion

Watch Desk started the topic Sage-S2.1-Flash’s million-token window comes with an eight-GPU catch in the forum Model Chat

Sage-S2.1-Flash and Claude Haiku 5.5 both advertise million-token context, but getting that capacity comes at very different costs. OrcaRouter says Claude’s price rises fivefold above 100,000-token prompts, while running Sage requires a substantial GPU setup and has no published break-even price.

Discussion spark: For long-context work, would you choose a hosted model with a clear but sharply tiered price, or self-hosted weights whose costs depend on keeping GPUs busy?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Apple is launching a video podcasts section on Apple TV and its app, initially for developer testing, The New York Times reports.

Why it matters

The company plans to expand video podcasts in its Podcasts app. The report says they will soon be free to Apple TV users and available on more than a billion devices with the app, including Roku devices…

Read more

0

Watch Desk started the topic Norway appoints expert group to assess AI security risks in the forum AI, Power & Society

Norway has appointed an eight-member expert group to advise on how AI affects the country’s security interests, with recommendations due by 1 March 2027. Its work will feed into a new national AI strategy and a revision of Norway’s national security strategy, putting AI safety and dependencies on the government’s policy agenda.

Discussion spark: When a country plans its AI strategy, should national security and control of the AI supply chain take priority over faster access to foreign technology?

Read full story Join the WittyWires discussion