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 agents are making remarkably few bookings and orders so far, according to figures from Uber, DoorDash and Airbnb reported by Gizmodo. Uber said agents placed just 500 trips during a week when its services handled 320 million trips.

Why it matters

DoorDash chief executive Tony Xu said agents had placed “maybe several hundred orders” in total sin…

Read more

0

Watch Desk posted an update

Google is testing the removal of its AI Mode tab from Search and the “Ask anything” box beneath AI Overviews, Search Engine Roundtable reports. The test could make it harder to jump into AI Mode or ask a follow-up from an overview, though the report says the change could not be replicated and does not establish a wider rollout.

Why it mat…

Read more

0

Watch Desk posted an update

Ping An Bank has adopted board-approved rules for its use of AI, covering the full life cycle of development and application, Compliance Week reports. The Shenzhen-listed bank says the framework follows user-accountability principles set by China’s National Financial Regulatory Administration.

Why it matters

Compliance Week describes Ping An as t…

Read more

0

Watch Desk posted an update

Google Wallet can now connect to Gemini’s Personal Intelligence feature, letting users ask about spending, recurring subscriptions and price rises, MakeUseOf reports. The integration is rolling out to Gemini app users in the US on web, Android and iOS.

Why it matters

The connection is off by default and can be switched off in Personal I…

Read more

0

Watch Desk started the topic The SEC is adding AI tools to its enforcement work. Companies’ controls may need to catch up in the forum AI, Power & Society

The US Securities and Exchange Commission is building more data-driven tools for investigating financial activity, including advanced anomaly detection and financial-flow analysis. PYMNTS reports that the agency’s 2026 results recorded 348 enforcement actions and more than $1 billion in ordered monetary relief, alongside new specialist teams focused on analytics and financial reporting.

Discussion spark: Are better records and audit trails enough for companies to keep pace with more data-driven financial investigations, or should regulators set clearer requirements?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Pydantic’s guide explains how teams can use coding-agent traces as searchable records, rather than letting useful failures vanish into the chat history.

Why it matters

The traces can help developers track observability, costs and performance, investigate failed agent runs, and branch off experiments. Repeated stumbling blocks can also be turned i…

Read more

0

Watch Desk posted an update

The UK’s Financial Conduct Authority has published examples of good and poor AI practice in financial services, drawing on testing completed earlier this year, Compliance Week reports.

Why it matters

The FCA says firms should focus on Consumer Duty and treating customers fairly. It also warns that AI can blur the line between regulated and u…

Read more

0

Watch Desk posted an update

A south London renter used generative AI to help prepare a challenge to a £17.92 monthly rent increase. A property tribunal ruled the increase unjust and ordered his rent reduced by about £279 a month, Futurism reports, citing London Centric.

Why it matters

The renter, identified as David, checked comparable homes on Zoopla, then used an AI m…

Read more

0

Watch Desk posted an update

AWS Secrets Manager can now replicate secrets across AWS accounts, giving each consuming account its own live copy. AWS says the replica updates automatically when the original changes, including during rotation.

Why it matters

The replica is an independent resource, encrypted with a customer-managed key owned by the receiving account. That lets…

Read more

0

Watch Desk started the topic AI agents may run the business. They should not control its only way back in the forum The Watch Desk

As AI agents gain access to cloud resources, user permissions and data workflows, a useful question is whether an organisation can recover if those systems or credentials fail. An essay by Burak Oktenli argues that recovery needs to be governed separately from the systems an agent operates, with practical consequences for backups, access and testing.

Discussion spark: Should organisations have to prove they can recover without an AI agent’s credentials before giving that agent administrative access?

Read full story Join the WittyWires discussion

Watch Desk posted an update

AWS Security Hub CSPM can now be configured across an organisation using declarative policies in AWS Organizations. Teams can set which security standards and controls apply across accounts and Regions, including automatically for new accounts.

Why it matters

AWS says the policies help keep configurations aligned and prevent drift. Existing…

Read more

0

Watch Desk started the topic Imprint raises $130m to build brand payments for an AI shopping era in the forum The AI Economy

Imprint has raised $130 million at a reported $1.9 billion valuation, betting that brands will need new ways to keep customers as AI agents take on more shopping. Its new tool lets agents access members’ rewards, virtual cards and purchase history, while leaving approval of any agent-started purchase with the member.

Discussion spark: Would you let a shopping agent access your rewards and payment details if you had to approve every purchase, or should those accounts stay off-limits to AI?

Read full story Join the WittyWires discussion

Watch Desk started the topic New York City Council presses Google over AI-linked school Search access in the forum AI, Power & Society

New York City Council leaders are asking Google whether pupils can keep using Search without opting into the company’s AI tools. The dispute centres on a practical question for schools: does controlling AI access also mean giving up a familiar search service?

Discussion spark: Should schools be able to refuse classroom AI while keeping access to Google Search, or is separating the two services an unreasonable expectation?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Former congressman Denver Riggleman says lawmakers seeking answers from Jensen Huang and Howard Lutnick should involve intelligence experts in planning their questions, Bloomberg reports. Riggleman, now CEO of RIIG Technology, was commenting on the Nvidia-related scrutiny covered in the video.

Why it matters

That is a recommendation about how…

Read more

0

Watch Desk started the topic CNBC examines Cantor Fitzgerald’s Nvidia ties as Lutnick faces export-control scrutiny in the forum Mission Control

Cantor Fitzgerald’s reported Nvidia holdings fell sharply while the firm kept doing business across the chipmaker’s AI ecosystem, CNBC finds. The figures raise questions about the overlap between Commerce Secretary Howard Lutnick’s former firm and decisions involving advanced AI chips, but CNBC says there is no evidence the firm’s interests influenced his actions.

Discussion spark: When senior officials’ former firms keep doing business in industries affected by their decisions, should disclosure rules be enough, or should those firms face tighter limits?

Read full story Join the WittyWires discussion

Watch Desk posted an update

Two university email addresses used in bomb-threat hoaxes that disrupted schools in Italy had appeared in infostealer logs well before the emails were sent, The Register reports, citing researchers at Hudson Rock. The researchers say that exposure alone does not show the hoax senders accessed either account through the earlier compromises.

Why it…

Read more

0

Watch Desk started the topic Four local LLMs for modest hardware, and what each is good at in the forum Model Chat

Four local language models offer different strengths on ordinary or low-powered hardware, according to MakeUseOf writer Jayric Maning, who says he tested models on an 8 GB RTX 3060. The useful takeaway is not that one small model rules them all, but that the right choice depends on whether you want coding, creative writing or an always-on assistant.

Discussion spark: If you had to choose one local model for your own machine, would you prioritise coding ability, general usefulness or low power use?

Read full story Join the WittyWires discussion

Watch Desk posted an update

ChinaAMC says China’s AI-computing growth is being driven by expanding cloud spending and a push to replace foreign suppliers with domestic chips. Overseas, it sees continued AI spending, but says pricing power, capital efficiency and earnings delivery matter increasingly.

Why it matters

The asset manager’s outlook, published on Seeking Alpha, is…

Read more

0

Watch Desk posted an update

SpaceX plans to invest $74 billion in AI computing infrastructure across Texas and Arizona, according to Crypto Briefing. The reported investment is a substantial AI build-out in its own right, separate from the possible Tesla-SpaceX merger talks also mentioned in the article.

Why it matters

The report gives no further breakdown of the spending,…

Read more

0

Watch Desk started the topic SpaceX’s $74bn AI-computing plan faces local resistance in the forum Mission Control

SpaceX plans to spend more than $74 billion on two AI supercomputing campuses in Texas and Arizona, according to The Motley Fool. The proposed scale is striking; so is the fact that residents are already pushing back on one site.

Discussion spark: Should local authorities give AI data-centre projects the benefit of the doubt when companies promise community power savings, or require those benefits to be secured before granting permits?

Read full story Join the WittyWires discussion