mlyle

23 ms·

361.
by mlyle 1y ago
They've gotta hope they get to cheap AGI, though. Any stall in progress either on chips or smartness/FLOP means there's a lot of surplus previous generation gear that can hang and commoditize it all out to open models. Just l
362.
by mlyle 1y ago
> I don't think WAL can ever really have a durable transaction; it is essentially based on leaving a transaction open until it gets "check-pointed" or actually committed to the database file. In general: WAL means you writ
363.
by mlyle 1y ago
I buy stuff -- small circuit boards, electronic components, unusual milling cutters, bearings, etc. The existence of these things as commodities lets me do a lot of deep EE and ME work with high school students. The US will not build a sup
364.
by mlyle 1y ago
Yup. P=IV is really surprising when you get to high power parts at low core voltages. Needless to say, you need lots of transistors and phases on voltage conversion, and you need lots and lots of plane area. (And,... 200A is the average
365.
by mlyle 1y ago
> You’re really wound up. Jsut consider that. It's hard to read emotions via text. I'm reading that -you- seem to be, with periods between each word, short phrases, &c. > “Shut up and don’t ask questions” is a very toxi
366.
by mlyle 1y ago
> and more consolidation into Wal-Mart scale corporations. I think this has more to do with the effective abandonment of antitrust mechanisms. > Raising the prices of those goods makes artisan products more cost competitive. Look, I d
367.
by mlyle 1y ago
I don't think making things more expensive for consumers is going to help that happen. I think it's going to steepen the divide between capital and labor. And most of the products we're talking about that are coming in via d
368.
by mlyle 1y ago
> We will have to pay more, and we should, for those things. No. US labor costs are high and working conditions are better, in large part, because US labor is worth it and US productivity is high. That labor is spent in high value indu
369.
by mlyle 1y ago
I think "this is your missing analysis" is a strong assertion to make to another human-- that sounds like an invitation to argue the merits. Through text, we don't have the benefit of tone. If your intention was to be curiou
370.
by mlyle 1y ago
You came here in your first comment blindly disagreeing with "This is your missing analysis." Then you seemed determined to misunderstand, e.g. > > But it's hard to escape your consumers paying most or all of the costs
371.
by mlyle 1y ago
That's the opposite of how tax incidence works. Elasticity means you can change your amount produced in response to changes in price. Producers can’t easily change output, so they bear more of the tax burden themselves. But in the long
372.
by mlyle 1y ago
I'm not talking about trade imbalance. I'm simply saying, tax incidence is predicted by elasticity, and in the long run suppliers have very high elasticity. You can possibly improve trade imbalance with tariffs (though retaliatio
373.
by mlyle 1y ago
Google uses plenty of water in cooling towers and chillers. Sure, water cooling loops to the server may reduce the amount a little bit compared to fans, but this isn't "recycling" in any normal sense. It's mostly a play
374.
by mlyle 1y ago
"We need demand more than supply" is a macro diagnosis. But tariff incidence is a micro question. Elasticity analysis doesn’t care whether the world has a demand shortfall or a supply glut. It asks: when a tax raises transaction c
375.
by mlyle 1y ago
You’re mixing up two different questions. "What the market will bear" is a monopoly pricing story. Food is messy because it's a commodity with a whole lot of substitution-- consumers have a high elasticity as a result. We are
376.
by mlyle 1y ago
In the long run, tariffs basically all fall on the consumer because producer and distributor behavior is near infinitely elastic. Econ 101 predicts that the party who is less able to adjust behavior in reaction to the tax pays most of the
377.
by mlyle 1y ago
> If you build on Intel, you are making a decision that exempts you from competing in the top of market for your products. Your designs will run hotter, consume more energy, have less yield, and lower clock. Most of semiconductor product
378.
by mlyle 1y ago
Yah. I think an AI playtester that could go "hey... this itch that lots of players seem to have doesn't get scratched often in your main gameplay loop" or "there's a valley 1/3rd of the way into the game where
379.
by mlyle 1y ago
I think there's surely a little bit of a middle ground? Subsidizing and protecting local production has a cost: there are gains from trade and operating less efficiently makes everyone's standard of living goes down. But you don&#
380.
by mlyle 1y ago
The only things that matters for this consideration are: A. How bad is the consequence of not having them? B. How long does it take to tool up domestic production if there's a supply disruption? C. How long can you tolerate the shortag
381.
by mlyle 1y ago
Just like you don’t want to wait many months to tool up to make food if there is conflict… You don’t want to wait many years to tool up to make chips that go in weapons, either. Else, you risk being structurally defeated very early, and thi
382.
by mlyle 1y ago
I think this is a silly take. Not all species had evolutionary pressure for cross-organism understanding. A different species that has selective pressure for social understanding and greater mental understanding is probably going to under
383.
by mlyle 1y ago
Nah, it's not that simple. Excessive passiveness causes ambiguity which causes its own risks. You want the cars to follow norms, modifying them down slightly for safety in cases where it's a clear benefit.
384.
by mlyle 1y ago
Yah. Latest thing I wrote was * Code using sympy to generate math problems testing different skills for students, with difficulty values affecting what kinds of things are selected, and various transforms to problems possible (e.g. having
385.
by mlyle 1y ago
Words have a definition and a connotation, and meaning is inferred from context cues. What a word means is based on its prevailing word in this context. No one reasonably uses "facial recognition" or "I recognized a face&quo
386.
by mlyle 1y ago
> I'm just interested in the undercurrent of why people seem to like this story Nah, it's "holy shit radioactive cows causing single-event-upset!@" It's the legend of the impossible to troubleshoot magical proble
387.
by mlyle 1y ago
Detecting a face is not the same as recognizing a face in either engineering parlance or typical usage. If I don't determine this is a face that I've seen before, I've not recognized the face (maybe I have recognized that
388.
by mlyle 1y ago
What if systems develop that are able to understand typical human intention and wants with an accuracy that is superior to most humans? You seem to rule this out, but despite having similar biology and wants, humans misunderstand others&#
389.
by mlyle 1y ago
I don't trip over CloudFlare except when in a weird VPN, and then it always gets out of my way after the challenge. Anubis screws with me a lot, and often doesn't work.
390.
by mlyle 1y ago
(I could believe that a very sick cow could possibly be a significant alpha emitter after being exposed to Chernobyl... but an alpha-emitting cow isn't going to screw up a computer inside a building).

More ›