Good post from Balaji on the "verification gap". You could see it as there being two modes in creation. Borrowing GAN terminology: 1) generation and 2) discrimination. e.g. painting - you make a brush stroke (1) and then you look for a while to see if you improved the

More journalists killed in Gaza than World War 1, World War 2, Vietnam, Yugoslavia & Afghanistan combined

So I'm building a cookie compliance checker using #ClaudeSonnet4 (via the excellent Replit ⠕) and just had the following exchange. This is why #vibecoding is hard, in the way that parenting teenagers is hard.

Oh my fucking god. Classified data. Not the slightest clue how AI works, or what might happen to the data. The Director of National Intelligence did this. 🤦‍♂️

So far in the last few days • Apple reinforced what I have been saying for 30 years • Salesforce’s new study on the unreliability of agents further reinforced what I have been saying for years about the limits of reasoning in LLM-based approaches. • Disney and Universal

almost nobody here even cares. that tells me all i need to know.

Nice - my AI startup school talk is now up! Chapters: 0:00 Imo fair to say that software is changing quite fundamentally again. LLMs are a new kind of computer, and you program them *in English*. Hence I think they are well deserving of a major version upgrade in terms of

Since there is talk about bringing back ASBESTOS (this is somehow true), I thought it would be useful to describe just some what happens to you when you breath this stuff into your lungs. In short, it’s terrible. A 🧵

Glass had 5 – 50× higher plastic than plastic or cans The French tested water, soda, beer, iced tea, lemonade, and wine sold nationwide and found that glass-bottled beverages carried far more plastic shrapnel than plastic bottles or cans. 🧵

"A water company has confirmed one of its treatment works was the source of pollution that killed thousands of fish in a Cornish river." I see South West Water are up to their old tricks again. bbc.co.uk/news/articles/…

Good piece. share.google/0AOpurrckG4hKz…

Grok 4 decides what it thinks about Israel/Palestine by searching for Elon's thoughts. Not a confidence booster in "maximally truth seeking" behavior. h/t Cate Hall. Screenshots are mine.

🚨 BREAKING: Switzerland is launching a PUBLICLY developed LLM, focusing on compliance with Swiss data protection & copyright laws, and with the EU AI Act. Is this the European way? Should EU countries follow suit? The LLM will be launched by ETH Zurich and the Swiss Federal

How Britain privatised itself into failure Britain outsourced its state – now private firms profit while public services crumble. In Failed State, Sam Freedman explains what went wrong yorkshirebylines.co.uk/politics/the-s…

Happening now in Tel Aviv: we march with the pictures of the children in Gaza who were starved to death by our government and our army. We cannot accept it. We call people to refuse the starvation, to refuse the killing, to refuse the annihilation.

Facebook once bought a VPN app for $120M and turned it into a surveillance tool that spied on 33M+ users' entire phones for years. This app helped Zuck buy WhatsApp for a whopping $19B and break Snapchat's encryption. Thread

Tesla (TSLA) is not paying its bills and it is destroying small American businesses | Electrek share.google/lHvkcnBgK4t7Dh…

My work here is truly done. Nobody with intellectual integrity can still believe that pure scaling will get us to AGI. GPT-5 may be a moderate quantitative improvement (and it may be cheaper) but it still fails in all the same qualitative ways as its predecessors, on chess, on

this is one of the most remarkable technical blog posts I’ve ever read

LLM adoption among US workers is closing in on 50%. Meanwhile labor productivity growth is lower than in 2020. Many counter-arguments can be made here, e.g. "they don't know yet how to be productive with it, they've only been using for 1-2 years", "50% is still too low to see