Why Trump just put a spy chief in charge of AI policy

TOGETHER WITH*

Which HR trends will shape your team in 2026?

AI, remote work, and global hiring are reshaping HR. What does that mean for how you plan, hire, and support your team?

Oyster's 2026 HR trends report breaks down the biggest trends shaping teams this year, so you can plan with them in mind.

Read the 2026 trends

*Sponsor

TECH BRIEF

Monday's briefing from Front Research: four stories on who gets to set the terms for AI, from the White House to the app on your phone.

• Trump creates a "Super Intelligence Force" chaired by spy chief Jay Clayton, with 120 days to report on AI risks

• Schneider Electric nears a roughly $20 billion all-cash deal for US industrial software maker PTC

• Extracted system prompts show Meta's Muse agent keeps a page on every person in a user's life

• Google cuts free Gemini users to Flash-Lite and removes Pro from the $4.99 AI Plus plan on October 9

Let's get into it.

Trump Puts Intelligence Chief Jay Clayton in Charge of a New "Super Intelligence Force"

• President Trump announced the Super Intelligence Force on Truth Social on Sunday, October 4, with Director of National Intelligence Jay Clayton as chair

• Vice chairs are FTC Chairman Andrew Ferguson, Pentagon technology chief Emil Michael and OPM Director Scott Kupor; members include JD Vance, Pete Hegseth and Scott Bessent

• The group has 120 days to report on AI risks and opportunities and to recommend what the federal government should be responsible for

• Elon Musk said SpaceXAI will be renamed SpaceXSI after Trump directed federal agencies to say "Super Intelligence" instead of "AI"

• Why it matters: federal AI policy now runs through the national security apparatus with a mandate built around being first, which lowers the odds of binding rules for the frontier labs in the near term

President Trump on Sunday created a task force he calls the Super Intelligence Force and named Director of National Intelligence Jay Clayton to chair it, Gizmodo reported. FTC Chairman Andrew Ferguson, Pentagon technology chief Emil Michael and Office of Personnel Management Director Scott Kupor serve as vice chairs, with Vice President JD Vance, Defense Secretary Pete Hegseth and Treasury Secretary Scott Bessent as members. In his Truth Social post, Trump said the group "will coordinate the Federal Government's engagement with Consumers, Public Interest Groups, Religious Organizations, Critical Infrastructure Providers, and Super Intelligence Companies."

The group has 120 days to report on AI risks and opportunities and to recommend the federal government's responsibilities, including response plans for AI-enabled threats that avoid regulation seen as slowing innovation, according to Implicator.ai, which also lists David Sacks and former Secretary of State Condoleezza Rice as outside advisers. Clayton, a former SEC chairman with no technology industry background, keeps his job overseeing 18 intelligence agencies. He framed the mission as a race: "The risk of not being first is high. Not being first increases identified, and unidentified, risks, particularly from our adversaries." The task force follows the September 29 White House Accord on Super Intelligence, under which AI companies signed voluntary safety standards, and comes as Anthropic and OpenAI press for federal guardrails.

The branding is part of the policy. Trump signed an order titled "Inaugurating the Era of Super Intelligence" that directs federal agencies to use "Super Intelligence" rather than "AI" in public communications, websites and reports, Fox Business reported. Elon Musk said SpaceXAI, formed in July 2026 when SpaceX absorbed xAI, will become SpaceXSI: "No more AI. SI, it's better."

Why it matters: the composition tells investors more than the name. A task force chaired by the intelligence director, with the Pentagon, the FTC and the federal personnel office as vice chairs, treats AI first as a national security and competitiveness question and only second as a consumer protection one. That points toward voluntary standards, federal procurement and threat response rather than licensing or liability regimes, which is favorable for the capital spending plans of Microsoft, Alphabet, Amazon, Meta, Nvidia and the private labs. The 120-day clock puts the recommendations in early February 2027, and the FTC chairman's seat is the one to watch: it is where any consumer or competition enforcement would originate. The companies asking for binding federal rules, Anthropic and OpenAI among them, did not get them on Sunday.

Schneider Electric Nears a $20 Billion All-Cash Deal for PTC

• The Financial Times reported on Sunday that Schneider Electric is close to buying Boston-based PTC for about $20 billion; Bloomberg put the price at more than $20 billion

• The all-cash deal could be announced as early as Monday, October 5, though talks could still fall apart

• PTC's market value was about $15.6 billion before the report, implying a premium of roughly 28%

• It would be Schneider's largest acquisition, four months after it bought industrial AI firm Cognite for $3.1 billion

• Why it matters: a French electrical equipment group is spending data center profits to buy one of the last large independent industrial software vendors

Schneider Electric is nearing an agreement to buy US engineering software company PTC for about $20 billion, the Financial Times reported on Sunday, citing people with knowledge of the matter. Reuters, in a story carried by Investing.com, said it could not immediately verify the report. Bloomberg separately reported that Schneider is nearing a deal at more than $20 billion.

The transaction would be all cash and could be announced as early as Monday, according to Global Banking and Finance Review, which published the Reuters account. Talks are ongoing and there is no certainty of a deal, and both companies declined to comment. PTC, which sells software for designing, manufacturing and servicing products, had a market capitalization of about $15.63 billion, so the reported price implies a premium of roughly 28%.

The deal extends a deliberate shift. Schneider began as a maker of fuses and circuit breakers and is now a major supplier of data center cooling, racks and power distribution, with US data center demand offsetting weakness in traditional electrical markets. It bought Cognite, a private industrial data and AI software provider, for $3.1 billion in June 2026, and raised its annual revenue guidance in July after record first-half results.

Why it matters: if signed, this is the largest deal in Schneider's history and one of the biggest industrial software transactions on record, and it puts PTC's design and product lifecycle tools inside a hardware vendor that already sells into factories and data centers. The strategic logic is that industrial AI needs engineering data, and PTC holds a great deal of it. The financial question is price and funding: $20 billion in cash is a large commitment for Schneider shareholders and will test how much of the data center windfall they want recycled into software. For the sector, the bid resets the valuation floor for the remaining listed industrial software names and raises the pressure on Siemens, Dassault Systemes, Autodesk and Rockwell Automation to respond. Because the price comes from press reports, the confirmed terms and any competing interest are the things to watch on Monday.

Meta's Muse Agent Keeps a Profile on Everyone in a User's Life, Extracted Prompts Show

• Wired reported on Sunday that extracted system prompts show Muse compiles "a page for every person in the user's life"

• The pages hold facts, relationship history and tips for improving each relationship, and are updated hourly, according to coverage of the report

• Muse has been downloaded by millions and has topped the US App Store's free chart since its September launch

• The people profiled include friends, relatives and colleagues who never agreed to be part of the product

• Why it matters: Meta's most successful AI product so far is built on data about non-users, the category of data that has drawn the heaviest privacy penalties in Europe

Meta's AI agent Muse builds and maintains a detailed profile of each person in a user's social circle, Wired reported on Sunday, based on system prompts extracted from the app. The instructions describe "a page for every person in the user's life," holding facts about them, the history of the relationship and suggestions for improving it.

According to AI Weekly, the pages are updated hourly and can include details such as birthdays and private arguments. The instructions were obtained by independent AI safety researcher Karan Joshi, who asked Muse through its chat interface to share its own files. The same account says Meta's position is that the operating files were designed to be accessible to users for transparency, and that Muse has been the number one free app on the US App Store since its launch in September.

The commercial momentum is not in doubt. TechBuzz noted that downloads have kept climbing despite the privacy concerns, and that the personalization depends on profiling relatives and colleagues who never opted in, at a time when European regulators are already scrutinizing AI data practices.

Why it matters: Muse is the first proof that Meta's AI spending can produce a chart-topping consumer product, and the feature that makes it useful is the one that creates the legal exposure. Profiles of people who are not users, compiled without their consent and refreshed every hour, sit squarely in the territory covered by GDPR and by US state privacy laws, and Meta has already paid billions of dollars in privacy penalties. The practical risks for investors are a forced redesign or a delayed rollout in the EU, and a feature set that becomes less sticky if the contact pages have to be limited. Competitors building agents with persistent memory, including OpenAI, Google and Apple, face the same design question, so whatever regulators decide about Muse will set the template for the category.

Google Cuts Free Gemini Users to Its Smallest Model and Pulls Pro From AI Plus

• From October 9, free Gemini users can select only Flash-Lite, losing access to Flash and Pro

• AI Plus subscribers ($4.99 a month) keep Flash and Flash-Lite but lose Pro

• AI Pro ($19.99 a month) becomes the cheapest plan with the Pro model and gains Deep Think, previously limited to AI Ultra

• Google is also adding low, medium and high effort settings that use up more of a user's allowance at higher levels

• Why it matters: Google is rationing compute by price tier, a sign that inference capacity, not demand, is the binding constraint on consumer AI

Google will restrict which Gemini models its lower tiers can use starting October 9, 9to5Google reported, citing a Google support document. Free users will be limited to Flash-Lite. AI Plus subscribers, who pay $4.99 a month, will have Flash-Lite and Flash but not Pro. AI Pro, at $19.99 a month, keeps every model and gains Deep Think, which until now was exclusive to AI Ultra at $99.99 to $199.99 a month.

The change governs which models a user can pick, not the size of the allowance, according to tbreak, which noted that AI Plus subscribers lose Pro on account-specific dates that Google will communicate by email. Usage continues to refresh every five hours until a weekly limit is reached. After the change, AI Pro is the lowest-priced plan that includes the Pro model.

Google is also preparing effort settings (low, medium and high) across models. The support document says higher levels "increase the model's ability to complete tasks and provide more thorough answers" while consuming more of a user's limits. The move follows the compute-based usage limits Google introduced in May 2026.

Why it matters: this is a price increase delivered through product tiers. Anyone who wants Google's best general model now has to pay $19.99 rather than $4.99 or nothing, which should lift average revenue per subscriber and reduce the cost of serving the free base. It also shows that even Alphabet, with its own chips and one of the largest capital budgets in the industry, is choosing to ration inference rather than give it away. The risk is share: free users who find Flash-Lite too weak can try ChatGPT, Claude or Meta's assistants, so engagement data over the next quarter matters. For the wider sector, the direction supports the bull case on AI monetization and on continued demand for compute from Nvidia and Broadcom.