AI Reading for Saturday March 15

Tech folks in Silicon Valley feel that AGI is coming real soon now - NY Times

---

On the other hand most AI scientists don't think scaling up current approaches leads to AGI. - New Scientist

if it doesn't have proper knowledge representation how can it be AGI? and also if it can't learn new skills and adapt on the fly.

---

Beating humans at a lot of skills, especially at speed and scale, is a different matter - NextBigFuture.com

---

Siri chief says the delay is ugly and embarrassing, but the team has done a lot, and will ship the world's best digital assistant. (Gift link) - Bloomberg

---

Intentionally training models with hidden objectives, and then having audit teams try to uncover the biases - Ars Technica

---

Anthropic’s plan to win the AI race with a somewhat less mass-market approach, according to CPO and former Instagram cofounder Krieger - The Verge

---

NHTSA documented Zoox robocars' apparent noncompliances with eight federal standards, including the lack of brake pedals, rearview mirrors, windshield wipers and defoggers. - Washington Post

---

DIY AI-powered doorbell screens visitors so you won’t even have to talk to those Mormons - Yanko Design - Modern Industrial Design News

---

Biodegradable nanobots on the rise for targeted drug delivery - GlobeNewswire News Room

---

New NIST AI framework eliminates “AI safety,” “responsible AI,” and “AI fairness”, adds “reducing ideological bias.” - WIRED

‘Orwellian’ has now been redefined to mean ‘unbelievable and tremendous’.

---

Enterprises use AI to counter cybersecurity threats - VentureBeat

---

Creators sell unposted videos for AI training at e.g. $3 a minute. - New York Post

(not sure why they should be required to permit training on posted videos for free)

---

'Open' model licenses are often restricted for commercial use, certain use cases - TechCrunch

---

Using AI to rewrite your own story - Fast Company

---

Follow the latest AI headlines via SkynetAndChill.com on Bluesky