Sunday, October 18
Platform Power, AI Shifts, Musk Turbulence
Meta is quietly walking back a land grab. The company removed language from its terms of service that critics said could have authorized it to scrape news sites (opens in a new tab) and use third-party content for AI training — a retreat that follows sustained pressure from publishers and regulators. Meanwhile, the 'No Kings' protest movement (opens in a new tab) says Meta cut off access to its Instagram account on the very day it needed it most, adding a political dimension to concerns about the company's grip on public discourse. Together, these episodes reveal a platform that shapes what information moves and what gets blocked, with little transparency into how those decisions get made.
TikTok has permanently banned the advertising account promoting Alex Gibney's documentary (opens in a new tab) on Elon Musk, a move that looks less like routine enforcement and more like editorial suppression given the subject matter. Musk himself settled lawsuits this week over his $1 million voter giveaway (opens in a new tab) during the 2024 election cycle, closing a legal chapter without admitting wrongdoing. And Tesla has now acknowledged in writing that it cannot guarantee Roadster reservation holders (opens in a new tab) will ever receive a vehicle — a stunning admission from a company that collected deposits years ago and has repeatedly delayed the car. The through-line is accountability: Musk-linked entities are simultaneously dodging legal consequences and enforcing silence around scrutiny.
The AI infrastructure layer is quietly reshuffling. Replit (opens in a new tab) has switched its default model provider from Anthropic to OpenAI, a signal that the competition for developer-facing AI integration is intensifying well beyond chatbots. Anthropic, for its part, is showing signs of corporate maturation — its workforce is becoming less junior and employees are spending more time in meetings, a pattern that typically marks the transition from scrappy startup to bureaucratic institution. Separately, websites are increasingly blocking AI agents (opens in a new tab) from crawling their content, with robots.txt-style restrictions now targeting autonomous systems rather than just search bots — a nascent but growing resistance to the assumption that the open web is AI's free data commons.