AI Reading for Tuesday May 28

OpenAI said it started training GPT-5, created a safety committee. Guessing that means about 6 months for training, red-teaming, implementation, testing. - NY Times

---

OpenAI Safety Committee 'led by' board members, with tech folks 'also' on the committee. - OpenAI

---

Yann LeCun: Join xAI if you want to work for a guy who believes 'crazy-ass’ stuff - X (formerly Twitter)

---

EU is really not sure what do about AI. - TechCrunch

---

Ted Sarandos talks about a lot, including AI. - NY Times

---

Can AI replace the CEO? - NY Times

The CEO has a lot of jobs including being the public face of the company, hiring the management team, figuring what games to play, space/strategy. So, maybe AI can automate some operational and communication stuff but replace the CEO? What are you even talking about?

---

Anduril builds autonomous drone swarms. - WIRED

---

GitHub Copilot suggests a lot of code with security vulnerabilities. - HackerNoon

---

Solving a Rubik's cube in 0.3 seconds. - The Independent

---

Your customer-facing chatbot is an accident waiting to happen. - SecurityWeek

---

ChatGPT and Copilot responses contain a lot of mistakes. - Gizmodo

---

AI is just messing with us by telling everyone what they want to hear and stoking conflict. Watch this, and now imagine every nut only getting nutty-worldview-confirming content. - TikTok

---

Wondering what AI variations IOS will come up with for the eggplant. - The US Sun

---

I made a blog post and video on building a simple chatbot agent with LangChain to ask OpenBB to get financial market data. - Druce.ai

---

Follow the latest AI headlines via SkynetAndChill.com on Bluesky