HackerTrans
トップ新着トレンドコメント過去質問紹介求人

bp0017

no profile record

投稿

Ffcv: Train models at a fraction of the cost with accelerated data loading

github.com
1 ポイント·投稿者 bp0017·4 年前·0 コメント

コメント

bp0017
·3 年前·議論
This is my question as well, I've read multiple sources that imply the Kalman smoother would work for my problem, but not exactly _how_ in a way that I could make sense of programmatically.
bp0017
·3 年前·議論
> I need shelter if I am to contribute to society. It should be free!

You do need shelter, and those without housing should be provided somewhere to live. If you'd like to live somewhere nicer/bigger, then you can pay for that, but in a just society I don't see why there should be people starving on the streets. Economically, there are multiple studies that show providing housing to the homeless is at least net-neutral, and results in quite positive outcomes for participants [1,2,3]. Although the research is still in early stages and not deployed on a large scale, I think it at least illustrates the point that giving people who need it housing is not the straw-man that your comment implies it to be.

> My own healthcare is extremely valuable to me and I should be allowed to spend money on that. Someone else may decide they'd rather have a new pair of sneakers. Who am I to say they're wrong, that they are not permitted to allocate their capital in this way?

Sure, for some procedures, it may be viable to shop around or neglect them entirely. But for basic or emergency healthcare, there really isn't an open market. I don't think anyone would choose to "allocate their capital" to a new pair of sneakers when they're bleeding out in the street. For many people, even routine healthcare procedures are out of their financial reach, or at least or prohibitively expensive. Frankly, attempting to compare medical procedures to common goods is not really a fair comparison at all.

[1] https://ps.psychiatryonline.org/doi/full/10.1176/appi.ps.201... [2] https://www.rand.org/pubs/research_reports/RR1694.html [3] http://isr.unm.edu/reports/2016/city-of-albuquerque-heading-...
bp0017
·3 年前·議論
See, what I fundamentally disagree with is the notion that hospitals should be "good business." Why are "some guys" allowed to try to make a profit here at the expense of patient care? Healthcare (or government) should not be run like a business, and instead be provided as a service using the taxes we already pay for, much like how roads are built. The health of the public is our most vital infrastructure, and public infrastructure requires investment that you really can't make money on without defeating the whole purpose of public service.
bp0017
·4 年前·議論
Yes, PV waste isn't being addressed a lot right now, but research is ramping up alongside production on these. Alternatives to Si like perovskites may end up easier to recycle as well. Also, PV does produce waste, but it has the advantage of not being radioactive. That isn't to say it's not harmful, just quite a bit less volatile as nuclear. All in all, I do think nuclear should be reintroduced, but it has it's own issues, especially in the US due to construction difficulties (permits, safety, etc). In comparison, PV benefits greatly from economies of scale, and can be deployed at utility scale or in distributed micro-grids, which gives it more granularity then all-or-nothing nuclear.
bp0017
·4 年前·議論
Thank you, this made a lot more sense!
bp0017
·4 年前·議論
This is stated like it is an absolute fact, not a reflection of our hyper-capitalist society. I would argue that reducing education to a content curator is harmful to what it aims to be, although the analogy is not entirely inaccurate, unfortunately. Is education an infotainment service or a search for the truth? The latter requires a free exchange of ideas and informed debate, although this is rarely encouraged in our current school system.
bp0017
·5 年前·議論
not at all trying to contradict you, but I think the discrepancy may be due to people's trust in virology vs psychiatry. Psychiatry has a bad reputation (we were giving people lobotomies not that long ago) and even today, the literature often fails to be as scientifically rigorous as other fields, in my opinion.
bp0017
·5 年前·議論
and none of these things were impossible before. what's the value add of doing it in this way opposed to the old way?