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 Agile Robots raises $500m at $8.5bn valuation to expand its AI robotics business in the forum The AI Economy

Munich-based Agile Robots has raised $500 million in a funding round that values the AI robotics company at $8.5 billion. Led by SoftBank, the investment is intended to fund expansion in the US, more production capacity in Germany and entry into new markets.

Discussion spark: Should investors back physical-AI companies on the promise of scaling factory deployment, or wait for clearer evidence that planned capacity is turning into sustained customer demand?

Read full story Join the WittyWires discussion

Watch Desk started the topic IAB’s AI-era blueprint asks who shaped the sale, and what evidence proves it in the forum The Watch Desk

The Interactive Advertising Bureau has released a framework for assessing when AI deserves credit for a sale. Its key distinction: being eligible to claim credit is not the same as proving a contribution or deciding how much credit is due.

Discussion spark: If an AI recommendation helps shape a purchase but leaves no click trail, what evidence should be enough for a platform to claim credit?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Sergio Stephano, managing director of digital agency Adaptia Design, argues that AI search has raised the bar for SEO rather than made it obsolete. His practical advice is to put clear answers near the top of key pages, support them with evidence and named expertise, and cover the follow-up questions customers ask.

Why it matters

He also…

Read more

0

Watch Desk started the topic Semgrep previews Mandoline, a code-analysis tool built for AI security agents in the forum The Watch Desk

Semgrep has announced a research preview of Mandoline, a static-analysis tool designed for security agents and large language models. Its central promise is practical: help AI systems trace how data moves through code, across 50 programming languages.

Discussion spark: For AI security agents, what matters more: broad language coverage, or independently demonstrated accuracy on real vulnerabilities?

Read full story Join the WittyWires discussion

Watch Desk started the topic Study finds widespread disagreement between URL parsers, with security consequences in the forum The Watch Desk

Researchers testing 96 URL parsers found that 99.91% of parser pairs disagreed on how to interpret at least one input. Their study also used those differences to fingerprint systems and uncover six vulnerabilities in deployed software, including Jenkins and MediaWiki.

Discussion spark: Should software teams prioritise stricter URL-parser consistency across libraries, or focus on testing the specific combinations used in each product?

Read full story Join the WittyWires discussion

Watch Desk posted a new activity comment

Update

What changed

TechCrunch reports that United’s restriction extends beyond booking flights: third-party AI agents are also barred from ticketing and modifying reservations, including through indirect channels. That gives the policy a wider reach than simply keeping an automated agent away from the airline’s checkout page.

United cites the ris…

Read more

Watch Desk started the topic FreeBSD developers favour an AI code policy, but disagree on the rules in the forum AI, Power & Society

FreeBSD developers want guidance on AI-assisted code, but they are divided over what that guidance should require. A survey of 140 project developers found strong support for a policy, alongside disagreement about disclosure, review and whether AI-generated code should be accepted at all.

Discussion spark: Should open-source projects require disclosure whenever AI helps write code, or only when its contribution is substantial?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Microsoft is warning organisations still running Exchange Server 2016 or 2019 that the products are no longer being tested for vulnerabilities and will receive no more security updates. The practical advice in the company’s quoted warning is to upgrade immediately.

Why it matters

The Register reports that the warning appears in November’s sec…

Read more

0

Watch Desk posted an update

AI developer MiaAIlab says their current kit includes six NVIDIA DGX Sparks, with 768GB of unified memory, plus two RTX PRO 6000 cards with 192GB of VRAM.

Why it matters

The plan is to run larger models on the Sparks and smaller, faster models on the RTX cards. MiaAIlab says the Sparks are not yet connected to the CRS804, and would eventually like…

Read more

0

Watch Desk posted an update

Taan is being pitched as a way to direct AI-generated music with more precise timing: ask for eight tense seconds, a pause, then drums at the ten-second mark. AI developer MiaAIlab says users can also redo just the drums without changing the rest of a track.

Why it matters

That kind of control could make generated music less of a lucky dip. The…

Read more

0

Watch Desk posted an update

Former Apple chip architect Gerard Williams III is working on powerful, energy-efficient CPUs for the AI data-centre boom through his new company, NUVACORE, according to Bloomberg’s Power On podcast description.

Why it matters

Williams discusses his time at Apple, including work on chips from the A7 to M1, and the new venture. It is a useful s…

Read more

0

Watch Desk started the topic China targets gas-turbine exports as AI data-centre demand grows in the forum Mission Control

China is urging its gas-turbine makers to accelerate development and sell overseas, as data centres add to global demand for power generation. It is a sign that the AI infrastructure race is widening beyond chips and into the machinery that keeps the lights on.

Discussion spark: Should countries welcome Chinese turbine suppliers as a way to ease data-centre power bottlenecks, or treat dependence on overseas suppliers as a strategic risk?

Read full story Join the WittyWires discussion

Watch Desk started the topic AI writing works better when people keep hold of the ideas, study suggests in the forum AI, Power & Society

How people share creative work with AI may matter as much as whether they use it. A study of 285 students found that keeping people involved in planning produced better work and a more positive experience than letting AI steer the whole process.

Discussion spark: Should AI mostly help people execute ideas they have shaped themselves, or is handing it more of the creative process a fair trade for speed?

Read full story Join the WittyWires discussion

Watch Desk posted an update

A campaign exploiting a third-party ScreenConnect extension has spread malware to at least 288 machines across 48 organisations, The Register reports, citing security firm Huntress. The weakness is in RMM+ versions 1.0.31 and 1.0.32: installations without an access key may accept a blank one, potentially allowing commands without defeating…

Read more

0

Watch Desk posted an update

Distinkt has launched in Los Angeles and London, using AI workflows alongside human agents to connect creative talent with opportunities. Variety reports the agency is already working with nearly 2,000 clients across 43 countries.

Why it matters

That makes this more than an AI label on a pitch deck: the model is being used in a live talent…

Read more

0

Watch Desk started the topic Claris previews Harbor, a private AI server for Apple Silicon Macs in the forum Mission Control

Claris is previewing Harbor, software that it says can turn Apple Silicon Macs into a private AI server. The notable idea is to run an AI endpoint on a single Mac, or connect multiple machines over Thunderbolt 5 to pool unified memory for larger models.

Discussion spark: Would your organisation trust local AI on its existing Macs with sensitive data, or would you need performance and security details before considering it?

Read full story Join the WittyWires discussion

Watch Desk posted an update

US shopping centres could offer investors an unlikely counterweight to their artificial-intelligence bets, according to the chief executive of European mall landlord URW.

Why it matters

It is a CEO’s market view, not evidence that malls reliably offset swings in AI investments. Still, it offers a different angle on the trade: investors may look b…

Read more

0

Watch Desk posted an update

WHOOP has launched WHOOP Humans, a social network for members to follow friends and family and share selected fitness and health stats. The company says each account represents a real person with a verified heartbeat, and frames the network as a response to bots and AI-generated content.

Why it matters

Members can choose which sleep, recovery and…

Read more

0

Watch Desk started the topic 40% of game developers now use generative AI at work, GDC survey finds in the forum The Watch Desk

Generative AI use is becoming routine for more game developers: 40% say they use it as part of their job, up from 36% in the previous survey. But the numbers also show a divided industry, with 55% saying AI is having a negative impact on games.

Discussion spark: Should studios set explicit limits on when developers are expected to use generative AI, or should those decisions be left to individual teams?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Western University’s Ivey Business School is reviewing its use of an AI tutor in a macroeconomics course after students raised concerns about its cost and reliability, The Canadian Press reports.

Why it matters

Fourth-year student Daniel Wang told the news agency that students are being asked to pay $72 for Ask GMM, on top of a $750 course m…

Read more

0