AI Headlines for Wednesday, September 4

Elon Musk: Now, witness the power of my fully armed and operational AI-shitposting supercomputer! - Fortune

In a shocking and yet utterly predictable meltdown, Trump, Musk post weird AI images of Harris. - Futurism

Remarkable that he thinks it’s a good idea to post about Kamala Harris being a Communist, Democrats being the party of criminals, and comparing Harris to Monica Lewinsky. For someone so smart he sure does a lot of dumb stuff, like saying people who say Hitler was misunderstood are ‘worth watching.’ Dare I utter the word deplorable, or is that still considered outrageously condescending?

---

Intensive mapping of Warner Brothers studio area is underway ahead of Tesla’s October 10 robotaxi unveiling. - The Verge

---

Can Waymo driverless robotaxis be a big profitable business? - NY Times

---

How robotaxis get remote help to extract themselves from jams. - NY Times

---

OpenAI is pitching big but still rather fuzzy plans for data centers, infrastructure, and semiconductor manufacturing in the US. - Bloomberg

---

OpenAI plans to use its own custom AI chips built by TSMC starting in late 2026. - The Register

---

DoJ subpoenas Nvidia as probe intensifies. - The Register

---

Nvidia antitrust investigation grows, Feds look into whether companies who don't exclusively use Nvidia get penalized, among other issues. - Bloomberg

---

You.com raises $50m for what sounds like enterprise super-Perplexity RAG agents. - TechCrunch

---

Some mixed reviews for OpenAI's SearchGPT. - Washington Post

---

Zillow's AI search will show you more amazing homes you can't afford. - The Verge

---

Yuval Harari: How AI threatens democracy - NY Times

---

Bots are astroturfing regulators' request for comment websites. People will probably have to prove they are human and American to comment. - Bloomberg

---

Using AI apps to create better headshots. - CNET

---

Carbon nanotube TPUs reduce power requirements by 3 orders of magnitude. - livescience.com

---

Man recorded of kids at Disney World to create AI-generated CSAM; Facial recognition startup Clearview got a huge fine from Dutch authorities. - Forbes

---

More ChatGPT voices coming soon. - Android Authority

---

AI to clarify hard to understand dialogue. - Digital Trends

---

Sometimes sweet-talking ChatGPT and asking nicely is all you need to hop the guardrails. - Reddit

---

How to use AI in Excel. - Geeky Gadgets

---

Zed is an AI-first code editor, built from the ground in Rust, in contrast to Cursor which is a fork of Electron-based VS Code. - AIM

---

Jason Brownlee has an Interior Design with Stable Diffusion mini-course. - Machine Learning Mastery

---

Google Gems help you jump-start custom prompts. - ZDNET

---

Users talk about how using ChatGPT as a therapist helped them. - Reddit

---

Intel dives into make-or-break Lunar Lake. The 120 claimed TOPS doubles the competition, but that's just a small part of the equation with single-core and multi-core benchmarks, battery life, and hopefully no disastrous defects. - The Register

---

Follow the latest AI headlines via SkynetAndChill.com on Bluesky