bo1024

68 ms·

931.
by bo1024 7y ago
I'm not saying anything there is wrong. I'm saying they don't answer the question - they're incomplete. They focus on listing differences between cardiac muscle and skeletal muscle, but those differences don't obvio
932.
by bo1024 7y ago
I was thinking of that answer too. It talks about the ways heart muscle is different from skeletal muscle, which is cool, but I don't see it carry those differences through into explaining about fatigue...
933.
by bo1024 7y ago
I don't think any of the top answers are good at all. I suspect part of the reason is how poorly we understand muscle fatigue. Many answers talk about lactate and how it builds up in muscles but not the heart. However they don't s
934.
by bo1024 7y ago
That's a tradeoff worth discussing. I think there should be an entirely different cross-platform layer for interactive and intense SPAs like Google Docs and 3D sites. But making it possible for 0.00001% of sites to have 3D comes at a h
935.
by bo1024 7y ago
The first thing that comes to my mind is accessibility. If browsers were simpler to create and the web were simpler, people could much more easily develop accessibility tools. I think the privacy and tracker-blocking crowd would have a lot
936.
by bo1024 7y ago
Agree very strongly. The amount of innovation we’ve lost because of the high barrier to entry is also staggering to try to conceive.
937.
by bo1024 7y ago
I don’t know much about finance and I find this a bit terrifying. Are they hoping banks will buy lots of stocks or what is the rationale?
938.
by bo1024 7y ago
It says in the two weeks closest to peak rate. So if the peak is mid-May, those 75% would all fall in May.
939.
by bo1024 7y ago
Of course our currently attention model is poorly developed, because the companies developing it are literally competing for your attention. Capturing attention translates directly into money or power for them. To realize your vision, we ne
940.
by bo1024 7y ago
Thanks! Very informative.
941.
by bo1024 7y ago
I think this discussion got off-track. The question is about whether the company that sold you the car is allowed to impose EULA-type limits on how you modify, use, or resell it, above and beyond safety laws that apply equally to ever perso
942.
by bo1024 7y ago
You would think, by the same arguments, that the stock market could always just trade at 15-minute intervals. Why not? But people go crazy when researchers (e.g. Eric Budish at U. Chicago) suggest lowering the frequency to milliseconds, let
943.
by bo1024 7y ago
I have a question. They show that any given depth-ell network, computing F, is w.h.p. approximated by some subnetwork of a random depth-2ell network. But there is a theorem that even depth-2 networks can approximate any continuous functio
944.
by bo1024 7y ago
This is the COST (common-ownership self-assessed tax) from the recent book Radical Markets by Weyl and Posner.
945.
by bo1024 7y ago
Because the apartment next door would be $1000 cheaper and tenants would leave and go there.
946.
by bo1024 7y ago
I don't follow that. HQ2 was about creation of new jobs and immigration of more people into the area. For a fixed supply, higher rents are caused by a larger number of people demanding housing at the current prices. Higher wages only a
947.
by bo1024 7y ago
I should have cited it; great read on this stuff and was ahead of the curve!
948.
by bo1024 7y ago
If you asked a researcher on fairness and discrimination in machine learning, "what's an application where ML could go horribly wrong", they would describe exactly this kind of example. I don't know a thing about this co
949.
by bo1024 7y ago
That makes sense for one company's stock price versus another's. But it doesn't help me understand how the average behaves...
950.
by bo1024 7y ago
Can you explain how that impacts my mental model / questions? Are you pointing out that some companies may be barely affected by coronavirus and some a lot?
951.
by bo1024 7y ago
I wonder though if coronavirus is additionally the excuse (or maybe "catalyst") for a correction that was going to happen sooner or later anyway. My mental model of the stock market recently is a bunch of people driving too fast o
952.
by bo1024 7y ago
I think you're supporting the parent poster's point: Given that there are a number of petty people out there, GOG may be hoping those kinds of folks will just torrent instead of buy+refund.
953.
by bo1024 7y ago
I think usually the problem is "almost getting it" and trying to move forward, which means small uncertainties add up and all the sudden one is totally lost without being sure exactly why. So it's important to go back and mak
954.
by bo1024 7y ago
You're right, the paper's messaging is very confusing. What they want to do (eventually) is propose loss functions for machine learning . NOT algorithms for sorting per se. The consider ML models of the form f: x --> r that ta
955.
by bo1024 7y ago
This is one reason behind universal basic income as a welfare program. There are no hoops, no cutoffs, no distortion of incentives.
956.
by bo1024 7y ago
Hmm, what about letting another car in front of you? I think trying to solve this with pure data will end up like the chat bots that seem real for two sentences and then fail disastrously when you try to actually converse.
957.
by bo1024 7y ago
Absolutely right. * Object detection is not solved. * Even if it were, prediction requires a theory of mind. What does the other person see, know, and want to do? * Even if we solved prediction, driving requires social interaction, coordi
958.
by bo1024 7y ago
I'm not sure I agree with this -- how do you distinguish between "the system works for 80% of the people" and "79% of the people don't have medical emergencies (but the system would fail them if they did)"?
959.
by bo1024 7y ago
I've always been pretty skeptical of secure boot and trusted hardware modules since it seems to generally mean "trusted by the manufacturer, not the user". The model feels like the manufacturer owns and controls the device, a
960.
by bo1024 7y ago
Strongly disagree. ML is traditionally deployed in low-risk scenarios like image labeling or recommendations where 1% or 5% error rate is totally "routine" and acceptable. Doesn't mean this kind of ML pipeline can translate t

More ›