AI Reading for Tuesday May 20
Microsoft's big event was all about the 'explosion' of AI agents - Business Insider
An interview with Microsoft CTO Kevin Scott on the birth of the agentic web - The Verge
NLWeb lets you add a chatbot to your web page, and agents would know how to talk to it. - TechCrunch
Microsoft did some stuff to support registering your agents, monitoring them, and red-teaming them - ZDNET
GitHub allows you to assign issues to a coding agent that will go away and think and submit a PR - The Verge
Today I was comparing Substack v. Ghost v. Beehiiv, tomorrow I might compare NYC mayoral candidates on my hot-button issues.
With NLWeb, instead of chatting with o3 like I did today, I could have an agent in my browser, that queries the chatbots on the product/campaign pages and synthesizes and gives me the result.
How Google decided to give websites no opt-out/in for AI search - Bloomberg
Google releases its NotebookLM mobile app - The Verge
Stratechery's Ben Thompson Interviews Nvidia CEO Jensen Huang About Chip Controls, AI Factories, and Enterprise Pragmatism - Stratechery by Ben Thompson
Trump signs the Take It Down Act criminalizing AI deepfakes. - Engadget
Hopefully it won't have unintended consequences like putting people in jail for Trump caricatures, sites not allowing image/video posting.
Klarna CEO says AI push may have gone too far, wants more humans now. - Futurism
Pushed revenue per worker way up - TechCrunch
But credit quality is down. Who could have predicted someone who buys a DoorDash burrito on a payment plan could be a bad credit risk?
Baiont’s Feng Ji: Quant managers who don’t adopt AI will be eliminated by the market - Financial Times
Time saved by AI offset by new work created, study suggests - Ars Technica
SAG-AFTRA filed an unfair labor practice complaint over Darth Vader's voice in Fortnite - Ars Technica
New Orleans called out for sketchiest use of facial recognition yet in the US - Ars Technica
AI Can Beat You in a Debate When It Knows Who You Are, Study Finds - Gizmodo
If your candidate's chatbot knows your personal info, maybe it can persuade you better when you or your agent visits its NLWeb chatbot.
Comparing ChatGPT's grammatical judgment with linguists and lay people finds it's pretty good. - Nature
Not that clear which GPT model. ChatGPT is not great at assigning calibrated scores, comparing 2 sentence is better. (I believe they found that in the study)
'Godfather of AI' Geoffrey Hinton says he trusts his chatbot more than he should - Business Insider
If You Can Trust An Airplane’s Autopilot, You Can Trust AI - Forbes
Ok, fair analogy, but you have an autopilot and also 2 highly trained pilots monitoring it and turning it on and off at appropriate times. Also, the amount of regulatory oversight over aircraft manufacturing and operation is intense, and every rule was written in blood.
A review of brain-computer interfaces, with number of studies and patients set to rise - The Wall Street Journal
Best free AI courses - Fast Company
or anything from deeplearning.ai
and Google’s intensive course
(I assume those last 2 are still available on video but not 100% sure)
Grok’s ‘white genocide’ meltdown nods to the real dangers of the AI arms race - CNN
Driverless vehicles that respond more like humans will cause less harm during road crashes, according to research - Financial Times
"the moment I realized AI could code better than me" - Reddit
Follow the latest AI headlines via SkynetAndChill.com on Bluesky