- TLDR AI
- Posts
- Trump unleashes “un-woke” AI agenda
Trump unleashes “un-woke” AI agenda
PLUS: Chatbots send 1B clicks, Altman warns your therapy chats aren’t private, and Meta’s wristband turns gestures into keystrokes
Helena Liu
July 26, 2025
Good morning, AI enthusiasts. Washington just heard Donald Trump’s promise to “unleash American AI” by scrapping guardrails, web publishers saw a 357% jump in visits coming from chatbots, Sam Altman cautioned that ChatGPT confessions have zero legal privilege, and Meta demoed a wristband that turns muscle twitches into computer commands.
In today’s TLDR AI:
Trump’s AI action plan cuts safety rules, targets “woke” models, and fast-tracks power-hungry data centers
AI referrals hit 1.13 B in June, up 357% year-on-year, even as Google still dwarfs them
Altman says ChatGPT therapy isn’t confidential, so your deepest secrets could be subpoenaed
Meta’s sEMG wristband deciphers hand gestures without cameras or implants
LATEST DEVELOPMENTS
TLDR: Three new executive orders roll back Biden-era guardrails and demand federally funded models stay free of DEI “dogmas.”
One order brands the US an “AI export powerhouse,” easing export permits for frontier models and chips.
Another speeds up environmental review for mega data centers, a win for Nvidia and cloud builders.
Tech CEOs (Huang, Palantir, OpenAI) cheered from the front row; civil-rights groups filed a “People’s AI action plan” in protest.
Lobbying spend by the top eight tech firms hit $36M so far in 2025, about $320K for every day Congress is in session.
Why it matters: If the plan survives, Silicon Valley gets lighter regs and looser export rules, cementing Big Tech as the biggest winners of Trump’s AI pivot.
TLDR: Similarweb data show ChatGPT, Gemini, Grok and friends sent over a billion referrals to the top 1,000 sites in June.
ChatGPT accounts for 80% of all AI referrals; news sites saw clicks jump 770% year-on-year.
Google still delivered 191B referrals in June, two orders of magnitude more but publishers fear “Google Zero” as AI summaries expand.
Yahoo led news traffic from AI (2.3 M), followed by Reuters and The Guardian; NYT lags after blocking ChatGPT during its lawsuit.
E-commerce winners: Amazon (4.5 M AI clicks), Etsy (2.0 M) and eBay (1.8 M).
Why it matters: The referral trickle is turning into a river; anyone who relies on search traffic now needs an AI traffic strategy, fast.
TLDR: OpenAI’s CEO told a podcast there’s zero doctor-patient or attorney-client confidentiality when you pour your heart out to ChatGPT.
Courts can subpoena user chats; OpenAI is already fighting a NYT discovery request that could expose millions of logs.
Altman said, “People talk about the most personal sh**… we haven’t figured out privacy for AI confessions yet.”
He wants parity with medical or legal privilege but admits policy is miles behind usage.
Until then, OpenAI advises users to stick to Enterprise tiers if they need stronger data controls.
Why it matters: Chatbots feel private, but legally they’re not, raising red flags for mental-health startups and anyone treating AI like a diary.
TLDR: Reality Labs built a Bluetooth wristband that reads surface-electromyography signals and decodes hand gestures on the fly.
Recognizes handwriting-in-air at 21 WPM and pinch gestures without cameras or calibration.
Generalizes across users thanks to 100 hrs of sEMG data from 300 volunteers; personalization bumps accuracy another 16%.
Meta open-sourced the dataset to spur research and hinted the tech could ship with future Quest devices.
Accessibility researchers say it could help users with limited mobility control PCs without implants.
Why it matters: If sEMG UI works at scale, keyboards and touchscreens might one day feel as clunky as dial-up modems.
COMMUNITY
Want to master AI & automations? Ready to turn your AI skills into a new revenue stream? Get started with our FREE workshop.
Learn how you can monetize AI and get certified as an AI specialist during our free web class. Click here to register.
Reply