HexPhantom

He was not writing abstractly about mortality; he knew exactly what it meant

Learn the whole system, then remove everything unnecessary

Public transport makes commuting less wasteful, but I am not sure it makes a long commute free

None of the ideas are completely new, but putting them together this clearly makes them harder to dismiss and easier to remember

The one part I struggle with is -accept everything-. Acceptance helps with what cannot be changed, but I have also become happier by refusing to accept certain jobs, habits, relationships...

That's probably the sweet spot

There's something useful about time that is already spoken for. You're on the train, you can't do much else, so learning some Dutch feels easy. At home the same half hour somehow gets fragmented into six different things

I think there's a middle group too: people who like having learned something, but don't really enjoy those first few sessions. For them, just knowing that the initial frustration is normal can help a lot

Wanna say that this is a much better argument for learning than productivity or "becoming a more interesting person". Sometimes it is simply a way to keep the mind pointed outward

One thing I wish this emphasized more is that adults often confuse learning with consuming material about learning, which is why my useful rule has become: if I'm not producing errors, I'm probably not practicing yet

I think this is a bit too deterministic. Even if Europe is in a weak position economically, "the US does not need the rest of the world" seems overstated

It feels less like a pure tech market now and more like cloud, semiconductors and defense policy all getting mixed together

Where the frontier labs still have a business is probably not casual consumer chat

I share the sadness, but I'm not sure the consumer market was ever the main prize here

Hard to see how this does not turn into export controls for models, just with a lot more ambiguity

You may not build your daily system that way afterwards, but the mental model sticks

Yeah, that seems like an important distinction

Instead of always trying to make models more current and general, there may be value in making them deliberately narrow, historically constrained and weird in a well-defined way

The vitamin D angle seems much more plausible to me as a public-health issue than as an explanation for broad population-level cognitive differences

What I find more plausible is that low sun exposure is one small contributor among many

The paper isn't saying vitamin D determines cognition in general, just that there was a modest signal in a couple of memory measures

Women in the study weren't all starting from very low vitamin D levels

I'd read this less as "high-dose vitamin D makes kids smarter" and more as "prenatal vitamin D might matter for some neurodevelopmental outcomes, and it’s worth testing more directly"

I think "accountability sink" is the right phrase here

I think the only plausible argument for AI here is not "it knows age better than humans," but "it might be more consistent than ad hoc visual judgments by different officers"

I especially like "do they show their work?" In a gold rush, obscurity is often part of the business model: vague claims, unverifiable demos, hand-wavy benchmarks, carefully managed customer stories. Companies serving serious professions eventually have to deal with people who ask boring, concrete questions and expect boring, concrete answers

Maybe a respectable company is one where the answer to "what would make this business more profitable but worse for the world?" is not treated as a product roadmap

I think that's why integrity matters so much: it removes a whole layer of moral bookkeeping from daily life

I think that's a good distinction: not everyone in tech has to be "passionate" in the romantic sense, but it's hard to do good work for years if you actively dislike the underlying activity

Some of it is very 19th-century and easy to over-romanticize, but "reputation is the asset" and "debt trades away freedom" have aged pretty well