HN user
timothy.green+hellohn@gmail.com
The original post of 11% is referring to leaked numbers about _MFU_, which has erroneously been re-reported as fraction of GPUs being used at all. The parent post is trying to correct this misconception.
BTW the reporter looks like Cotten Hill if he was real, and actually fought in all those wars. I'm quite surprised they had him hosting the video. I'm curious what decisions led to this.
I hope positively surprised :) That's Paul Carter - he's a regular presenter on the BBC, particularly their tech show "BBC Click". Here's a nice interview (https://disabilityhorizons.com/2019/09/paul-carter-journalis...) he gave to Disability Horizons a few years ago about his experiences.
And one of the interesting (and hard) computer science problems is how to turn an arbitrary einsum expression into the most efficient series of operations (transpose, matmul, etc.) available in hardware kernels. This is mentioned in the post under "How Contractions Are Actually Computed".
Really sad - he was often the sensible voice in the UK media (e.g. BBC articles) when it came to security issues.
There was actually a lot of pushback against the austere aesthetic by government ministers - they wanted fancy looking photo banners and pretty things. This was pushed back on in the name of making something functional the prioritizes users.
Honestly this sort of climate & environment tokenism frustrates me - spending a lot of energy trying to reduce and elimate things that barely matter (or even actually make things worse by displacement), while ignoring the big ticket items. At best it's innumerate, at worst it's greenwashing.
Harry Nyquist isn't exactly an unknown engineer who doesn't have his own achievements, though - not sure why people are saying he would be fired in a modern company!
What was remarkable is how Facebook's Threads app jumped straight to being full of advertisers and hucksters - they didn't think that maybe the right way to bootstrap a social network would be to make it full of authentic conversations, at least to start with.
We're bad at onshore wind, not due to lack of possible sites, but because of a government ban on building it: https://www.theguardian.com/environment/2022/nov/29/onshore-...
Why? At what level do you define "local" - country, state, city, neighbourhood, street, literally next door, just my house?
Social media owners wanted all the benefits of superscale communities with none of the responsibilities - it isn't surprising that moderation got worse.
Reddit vaguely has a workable approach with subreddits, but it's still high variance, and they have to discourage long-lived comment threads.
I'm not going to detail challenges in medical ML - the literature can do that. But just to mention one that other people haven't that goes beyond just precision-recall: algorithms can be biased based on variance in physiology (e.g. more accurate for men), and understanding how an algorithm is biased is very important for the person interpeting the information, who should be a trained doctor.
And it doesn't matter what his credentials are, that's appeal to authority. If he thinks this should be used and trusted by people for decision making, then he should submit it to independent peer review and regulatory approval.
So you should ignore it and see a doctor regardless.
Thermometers are well understood, simple devices, and there are other complementary checks (e.g. does my forehead feel hot) if they fail.
This project might lead to people thinking they're in the clear and not seek appropriate medical treatment, or be overtreated due to an error. You should always talk to a qualified doctor if you're concerned about your health, and not use projects like these for decision making.
What peer review or regulatory approval process has this been through? Seems pretty irresponsible -- there are many notorious pitfalls encountered with ML for medical imaging. You shouldn't play with people's lives.
The best thing to do these days is use an AlphaFold prediction, if there exists a confident one, as a starting position for refinement.
There's a responsibility under refugee conventions to take on people needing asylum. The conflation of them with illegal immigration is quite ridiculous. The UK is a rich country (~6th highest in the world), we should shoulder some of the (small) burden that we're asked to - many other, poorer, countries take on much higher refugee burdens.
Just in case you live on another universe.
Or any country that isn't the US?
I'm guessing here that ASA is 'American Statistical Association', rather than say, the UK's 'Advertising Standards Authority'?
Funny how when confronted with an article that gives a clear, unambiguous demonstration of how patriarchy negatively affects women, you come up with two weak articles that supposedly demonstrate cases where women were advantaged (they weren't). From that, I'm led to believe that maybe the opposite actually isn't more common than I think.
Maybe you don't actually think that the issue in the Tokyo medical school was such a big deal after all?
Could you elaborate why you think it looks like a fractional FT?
If you solve Schrödinger's equation with periodic boundary conditions on the potential what you get are Bloch functions, which are the product of a phasor and a periodic function.
If it meant that, I'd expect it to be:
"The pound won’t be around for much longer"
so the original sentence is sort of a pun. Maybe "around" is a local use of English.
That video is basically nonsense and has nothing to do with string theory or physics.
It's not obvious that these designs have anything to do with deep learning; it doesn't appear to be mentioned in any of the linked articles.
That's true in our reference frame, with some error bar, but astronomers rarely make the distinction when talking about observing events, so I always find it odd when laymen insist on making it.
I guess the distinction is rarely made by professionals because
- It requires defining a reference frame.
- It makes no difference to the causality.
- Time is a function of distance, which usually has large error bars on it, so it's better to discuss the distance instead.
- Time as a function of distance gets more complex on cosmological scales.
On the last point, you can get the exciting realisation that, due to the expansion of space, objects actually start getting larger the further away they get beyond a certain distance.
As it turns out, TensorFlow's optimizers don't have a way to manually terminate training before convergence.
I don't know how you determined this, but the optimizer minimize op definitely only does one step, equivalent to doing the gradient update yourself.
That's really amazing.
In terms of utility to effort ratio, I think it's right to start with national legislatures. However, I hope there will eventually be a dataset that's as deep as Political Graveyard for the entire planet, even if just to satisfy my inner data nerd :)
Agree with not getting involved in feature creep. There are hundreds of failed civic tech projects, mostly because they tried to do too much. Users do want "positions" or "issues" etc. but those are really hard to find / build good data for. That's the next level!
epoch_100: Stick to making an awesome data source and interface to solve the one problem you've picked :)
In the UK mySociety does similar things, their WriteToThem services gives a drill-down on the politicians for a given postcode. You'd probably be really interested in their http://www.everypolitician.org/ project, which is trying to do this for the entire planet -- but generally working from the top down.
I'm involved with https://democracyclub.org.uk/, which works particularly on developing really good election data. epoch_100, reach out to one of us if you want to chat :)
You're taxed like a resident by the US. A friend had their tax-free (in the UK) PhD stipend taxed by the US government. Other people can find themselves owing significant taxes on property sales in the UK, despite the transaction having absolutely nothing to do with the US government [1].
There are exemptions which are pretty low and clearly don't cover the above examples. Plus you still have to do the overly onerous reporting of all your bank accounts. Ultimately the winners are the accountants.
It's not justified at all by services that the US does for overseas citizens, many countries do that without needing a draconian tax system for non-resident citizens.
I think it sort of comes down to Americans believing that you must be up to something dodgy if you want to live in another country :)
[1] http://www.theatlantic.com/international/archive/2015/02/bor...
constituent comments for the bills
In my experience, sites that do this sort of thing usually fail. They're ghost towns, and no-one cares what an unrepresentative sample of internet people think about bills.
I'm not based in the US, but I think the other points are covered by sites like https://www.govtrack.us/. Open Congress was another, but was shut down when the Sunlight Foundation went a bit haywire.