I have a hobby of photographing garbage trucks, especially on my travels around the world. It seems to me that garbage collection is in so many ways the essence of civilization. It is underappreciated.

Dwarkesh and Dylan Patel got Satya to actually answer the hard questions about Microsoft's AI strategy. A few things that stood out: 1. On losing coding assistant market share from ~100% to 25% in one year: "I love this chart for so many reasons. One is we're still on the top.

Good post from @mattyglesias: "Democrats don’t need to learn from people who appeal to the voters who hate Donald Trump. They need to learn from people who appeal to the voters who don’t hate Donald Trump. That is literally the assignment. But absolutely everything about the

Claude Code + Bash + Skills is all you need! Jokes aside, after playing around with a couple of computer-using agents, I gave up on them. They are heavily optimized for clicking on common buttons and interfaces (e.g., order delivery), but suck at actual creative work like

If this figure of global mean temperature since 1658 doesn't shout that AI isn't the only disruption headed our way, you aren't paying attention. We are heading into a period where "normal" is changing in unprecedented ways. From https://t.co/xBr74LG0Cq via @adamtooze

I'm going to be hosting Kent Beck for an hour on my live with Tim show on https://t.co/l6ApZY0t5V this coming Wednesday November 12 at 11 am EST to talk about coding with AI. Kent is one of the most influential figures in modern software development. He created Extreme

So much of the thinking about how to deal with the risks of AI seems to start with laws and regulations. What if, rather than thinking about what should be legal, what if we were thinking about what should be normal? How might following norms enable a market, with laws and

From @pkedrosky: "I have little patience for most private equity musings, as readers will know, but I will concede that KKR calling speculative multi-gigawatt data centers "bragawatts" is something I wish I had come up with. It is handy because it captures the idea that much of

"It has never been easier to build an MVP and in turn, it has never been harder to keep focus. When new features always feel like they're just a prompt away, feature creep feels like a never ending battle. Being disciplined is more important than ever." I really feel this one!

Ben Thompson and Michael Mortenson don't use the terminology of @doctorow's enshittification, but they explain why Google and Amazon are screwed in the age of AI answers unless they change their ways. https://t.co/p5TrsnwRVv

OpenAI has been testing out different lines of defense all day. They have decided to go with “but we aren’t asking anything specific to ourselves” when in fact perhaps 80% of any subsidies/tax credits/loan credits might go to them or their partners. If any of this goes through

A good post by Bill Janeway, viewing the AI bubble through the history of past bubbles, asking whether the industry as currently constituted will ever be able to reach sufficient profitability to pay for the massive up front (and ongoing) capital investment. Perhaps his most

This post is gold.

I recently posted some thoughts here about the idea of AI as a worker vs. AI as a tool. I decided to write them up at greater length, and after preparing a first draft, I asked Claude what it thought of my argument, and whether it was a worker or a tool. Its response was

In a recent newsletter, Ben Thompson called a portion of Jensen Huang’s keynote at NVidia’s GPU Technology Conference (GTC) in DC “an excellent articulation of the thesis that the AI market is orders of magnitude bigger than the software market.” While I’m loath to contradict as

A while back, Anthropic shocked Cursor by suddenly raising its prices. Researcher Anjali Shrivavastava made the case that this wasn't just price gouging, but a clue to a fundamental economic issue at the heart of token pricing: the number of tokens that might be used by a given

I had a great interaction with a customer service chatbot last week, and yet the customer service itself was no better. That left me thinking about why that was, and why making the AI better isn't the answer. No matter how intelligent the AI powering the chatbot might be, it has

Practical #MachineLearning for #ComputerVision — End-to-End ML for Images: https://t.co/plRvmVeLF0 ———— #ML #DataScience #AI #DeepLearning #NeuralNetworks #DataScientist

Language is a map; getting the right map helps us see things correctly and get where we want to go. Here's a good map: