LT Weekly: Oct 2, 2026
The major market indices were fairly flat with the S&P 500 closing slightly down and Nasdaq closing slightly up, but under the hood, the equal weight S&P fell -0.7% and was down for the 7th straight week. A striking stat (per FactSet) that I saw was that fewer than 25% of index constituents are above their 50-day moving average. Yields remain a big focus with the 10 and 30-year yields hitting their highest levels since 2002, though each ended the week off those record levels as Fed rate cut expectations eased (and helped the Friday stock market rally, along with lower oil prices).
Fundamentally, we focused on the below themes and developments this week (it was hard to get away from AI!).
- The AI Industry Agrees To “Morally Binding” Self-Policing…But Will It Work?
- The Personal Agent Race Is Heating Up As OpenAI Connects The Dots
- Under Anthropic’s Hood: Accelerating Growth, Big Losses, Huge Commitments & Unusual Risk Disclosures
- The Telecom-Satellite Landscape Continues To Evolve
- Quick Hits From The Media Majors At Bloomberg Screentime
- More Retail Price Cuts Emerge While Personalized Pricing Remains A Hot Button
- How Quickly Is AI Search Increasing & Converting? VERY
- How Much Revenue Is Needed To Support The Enormous AI Build? $6 TRILLION
- Grab Bag: Robinhood Agents And Extended Traded / Amazon Driver Smart Glasses / Mattel Takeover Talks
Enoy the read and weekend!
Best,
Leslie
The AI Industry Agrees To “Morally Binding” Self-Policing…But Will It Work?
Following months of AI model breakouts (most notably OpenAI agents hacking Hugging Face, an Australian gov’t health portal and US agency websites, alongside incidents with Anthropic, Meta and Google models), the White House’s answer was a voluntary, “morally binding” accord signed by Trump and top AI CEOs on Tues vs pursuing new regulatory rules. However, pressure for more enforceable guardrails came this week from some House Democrats who demanded that the labs disclose every known breach by 10/2, and from Florida’s AG, who is seeking a court order restricting OpenAI from developing new models w/o independent 3rd-party guardrails and from offering ChatGPT to minors in the state. The FTC also jumped into the arena and is now probing Anthropic, OpenAI and other AI labs to determine dangers that the technology poses to consumers.
All in all, AI safety, or shall we now say SI safety, is far from solved, and the path ahead remains unclear and largely in the hands of the tech industry, at least for now. See more details below.
The “White House Accord on Super Intelligence” signed on Tues by Trump & top AI CEOs is “morally binding” (link/link/link/link/link/link/link)