Under Trump, the us governments signature might as well be in pencil. This is America’s high water mark. It has lost the trust of the world. And since OpenAI and Anthropic and NVIDIA are the bulk of the hope of the American economy… America is fucked.
Canada walking away from American Bullying will go down in history of the moment the empire’s descent accelerated.
Author here. I’m building a global AI platform where users are on conference WiFi trying to pull up digital business cards. Origin server in Oregon, users in London, Dubai, Sydney. This essay started as an internal architecture doc explaining why we moved to an edge-first pattern with Cloudflare Workers, and turned into a deeper exploration of why the Origin-Edge-Client model is the next major architectural shift. The core argument: TCP recovery time is dictated by RTT, so terminating connections at the edge doesn’t just save latency on the happy path — it fundamentally changes how fast you recover from packet loss. The essay also gets into why Amdahl’s Law is a better lens than Moore’s Law for systems design, and why that maps directly onto this pattern.
More optimistic signal it’s very early innings in the architectural side of AI, with many more orders of magnitude power-to-intelligence efficiency to come, and less certainty today’s giants’ advantages will be durable.
my guess - e ink refresh rates / ghosting suck and it would be hellish hard to get an on screen keyboard with an e ink display to get anywhere the typing speed a modern touchscreen can deliver.
I guess your argument could be leveled against any transformational technology, from the industrial age through to the internet (which many doubted would have any meaningful economic impact, and clearly didn't solve many of the most pressing problems of the day for humanity).
Yes. There has been a regression in these metrics for white folks in the US. This is the first generation of whites in America who can expect to earn less and live shorter lives than their parents. However, that doesn’t generalize to the rest of the population, or world, and in America the reasons are policy: healthcare and education. Not because AI or tech broadly is particularly pernicious.
LLMs aren’t the end all be all of anything. But they’re clearly a step towards augmenting human cognition and in giving machines the ability to perform cognitive tasks. And when Google says a quarter of its code is being written by LLMs, and DeepMind is making tremendous progress on protein folding and DNA understanding with fundamentally the same technology, it seems pretty clear that we’d miss out on a lot without this.
Full disclosure: I think protein folding and DNA prediction could quite possibly the biggest advancements in medicine, ever. And still, all the critiques of LLMs being janky and not nearly sufficient to be generally intelligent are true.
So yes, I think it’s absolutely on the scale of electrification.
They were wrong to believe that technological progress could be stopped. The viable path is policy which ensures the gains are fairly distributed, not try to break the machines. That tactic has never and will never work.
I take Sutton's Bitter Lesson to basically say that compute scale tends to win over projecting what we think makes sense as a structure for thinking.
I also think that as we move away from purely von neumann architectures to more neuromorphic things, the algorithms we design and ways those systems will scale will change. Still, I think I agree that scaling compute / learning will continue to be a fruitful path.
I just take screenshots from the Apple Fitness app, and add a little of my own commentary, and GPT figures out what to do for the next session, extracting the right drills, programming it correctly given fatigue, progress, etc.
I'm experimenting with using GPT-4o (via OpenAI’s o3-pro) as a swim coach—part analyst, part training partner. This post is the second in a series documenting how it’s going, including how I’ve adapted the workflow to be more human-led and how I'm building a SwiftUI app that connects directly to Apple Health for automated, AI-powered feedback. If you're curious about applied AI, quantified self, or just hacking better habits, I’d love your thoughts.
Exploring giving a genetic genius (o3 pro Deep Research) lots of context (workout logs, personal notes, a PDF of a canonical swimming textbook) to build a custom trainer.
o4 quickly built a command line tool that transcribes arbitrary length audio files, and adds a summary and suggested actions/todo’s. It costs only a few cents to use, and using the APIs directly gives me more confidence that the data isn’t being used for purposes other than my own.