AI Reading for Tuesday October 1

Microsoft relaunches Copilot, which can now read your screen and talk to you, and 'think deeply', possibly leveraging OpenAIs o1 model - The Verge

---

Copilot “Click to Do” will let you circle stuff on your screen, get a context menu of stuff to do with it, like image search, send prompts to explain or answer questions about it. - The Verge

---

Google NotebookLM chief teases new features for AI generated podcasts - VentureBeat

---

Google's NotebookLM podcasts are a sensation. But it's not crazy hard to write a prompt to turn text into a Socratic dialogue and send it to TTS. That NPR emotional timbre is wild tho. - VentureBeat

---

Bill Gates worries about AI misuse for cybercrime and bioterrorism, the rapid rate of change leading to job losses, and potential loss of control over advanced AI systems. - Business Insider

---

OpenAI needs to keep a pipeline of new magic coming to keep raising more money at higher valuations and not implode - Benn Stancil

---

One could imagine a situation in the future where OpenAI can't raise more money and gets a bear hug from Microsoft. - Spyglass

---

As Apple steps back, SoftBank steps up with $500m OpenAI stake - FT

---

The narrative is that Apple's new AI iPhones aren't generating the splash that was expected. - The Conversation

---

Writing health AI rules: Why Micky Tripathi faces one of the toughest jobs in Washington - STAT

---

VCs are dispatching truckloads of cash to Mira Murati's house - Business Insider

---

An AI to let you talk with your future self. (to reduce anxiety, really?) - MIT News | Massachusetts Institute of Technology

---

Y Combinator is being criticized after it backed an AI startup that admitted it just cloned another AI startup's repo and slapped its own license on it - TechCrunch

---

AI can help you solve your 'someone who is good at the economy help me' budget questions - CNET

(help me budget meme)

---

Sam Altman Interview on AI, Capitalism, and the Power of Ambition - Geeky Gadgets

---

Maybe it's time for smart glasses without the dorky bulk and AR shades, without the AR and with just the camera, mike and AI. Still weird though. - MIT Technology Review

---

David Autor makes the case that AI will help the middle class with democratization of skills and lots of middle-skill jobs (doing what used to require high skills) - WSJ

---

EU appoints an expert panel to draft AI code. - euronews

---

If you are not sure how to use AI effectively on a problem, the quickest thing that might work is, ask the AI. - Hacker Noon

---

Smaller models are highly capable. - MIT Technology Review

Looking forward to seeing the new models like Llama 3.2 and Molmo in LMarena. Qwen 2.5 improved significantly. Crazy that a 9b-parameter model beats early GPT-4s! - lmarena.ai

---

Bengio: Humanity faces a 'catastrophic' future if we don’t regulate AI - livescience.com

---

Nvidia wants to be the OS for robots. - Business Insider

---

Cerebras files for IPO. I think there is a market for these giant inference chips but I'm not sure they found product-market fit yet. - CNBC

---

The Simpsons did an AI 'series finale'? wut? - New York Post

---

Follow the latest AI headlines via SkynetAndChill.com on Bluesky