adw

15 ms·

211.
by adw 4y ago
The relevant metric is the literature - adoption is driven by new models, and most new models in academia are being implemented in PyTorch or JAX. (In particular, TensorFlow 2 is nowhere to be seen.) The new grads joining the workforce are
212.
by adw 4y ago
The ML is made of, and expressed in terms of, math. There’s no way round that if you want to train models; you need first-year undergrad statistics and linear algebra. In other words, just as you do with any language, you need to understand
213.
by adw 4y ago
Another interesting question might be if it’s affected anyone’s job application behavior.
214.
by adw 4y ago
Nah, I was just wrong :-) I know a lot of big touring productions drive lighting and patch changes off the same clock (usually a laptop which provides time code, possibly backing tracks, and possibly patch changes for effects and synths), a
215.
by adw 4y ago
https://twitter.com/adambroach/status/1582854151196721152 I love what MIDI has given us, but MIDI devices are not designed for adversarial environments. I absolutely would want this to be mandatory opt-in (per ses
216.
by adw 4y ago
https://www.youtube.com/watch?v=IDn7ZDcx9w0 is great (in the same vein). No-one uses MIDI 2.0 yet, we don’t need it! MPE ( https://www.midi.org/midi-articles/midi-polyphonic-expressio... ) is only just b
217.
by adw 4y ago
Dave Smith, may he rest in peace, and Ikutaro Kakehashi deserve all of the flowers. https://www.grammy.com/news/technical-grammy-award-ikutaro-k... Dave was a brilliant designer, both analogue and digital, but not so g
218.
by adw 4y ago
This happens in all infrastructure, particularly power and water; https://en.wikipedia.org/wiki/TV_pickup
219.
by adw 4y ago
The relevant skills have more in common with the early 2000s supercomputing community than the mod scene. We are all data-parallel distributed now.
220.
by adw 4y ago
It also speaks to culture. Decisions individuals make in large organisations are, on average, downstream of institutional culture, so if a large organisation is responsible for a lot of bad decisions then the leaders are responsible for the
221.
by adw 4y ago
Lichess/chess.com are doing pretty well! (For that matter, Wordle.) “Mobile games“ is so broad as to be meaningless as an umbrella term now.
222.
by adw 4y ago
Ground beef - all meat, really - is a luxury food.
223.
by adw 4y ago
TFA covers this (extended security coverage from a third party provider).
224.
by adw 4y ago
10k doesn’t cover search costs. What about health insurance, for example?
225.
by adw 4y ago
10x more than academia. I’m a PhD in mineral physics turned Big Four MLE via startups and I earn about 10x what I would be on if I had stayed on the academic track.
226.
by adw 5y ago
As a physicist who spent a decade in academia, including a PhD where all the new work was done in Python, it absolutely won in some fields by beating - or rather, by conveniently wrapping - Fortran. (In particular, that’s how things have go
227.
by adw 5y ago
Fortran interop (f2py in particular) was a significant factor, and as soon as you get one thing (in this case LAPACK and BLAS bindings) it snowballs. Also, Python is significantly more initially familiar for informal programmers and that’s
228.
by adw 5y ago
Oh for sure. I blame David Cameron (ha ha only serious, he was senior at Carlton which merged with Granada to form modern ITV plc)
229.
by adw 5y ago
The ITV network still isn’t a single company! STV (Scotland) is not owned by ITV plc.
230.
by adw 5y ago
He’s British, so that’d require some dramatic geopolitical developments!
231.
by adw 5y ago
Type 1 diabetes, rheumatoid arthritis, Crohn's disease, ulcerative colitis, lupus, multiple sclerosis; a whole grab-bag of fun. These are also difficult diseases to treat – often involving biologics (large molecules administered by inj
232.
by adw 5y ago
One intuition is that you can generate pairs which you know to be the “same thing” (a single example under heavy augmentation) and ensure they’re close in representation space whereas mismatched pairs are maximized in distance. That’s a lab
233.
by adw 5y ago
I'm a long time out of the game (PhD in solid state physics in the early 2000s), but the key phrase here is "phase transition". She's making an argument along the lines of: * MOND is a field-like hypothesis which explain
234.
by adw 6y ago
> Locking down Australia would be like locking down one of our less densely populated states This is very wrong. Australia is very, very dense, where people live : it just has a lot of absolutely uninhabited land . Functionally Austral
235.
by adw 6y ago
Lockdowns, social distancing and masking – done consistently – do work. Even the half-measures lockdowns the US has gone through has stopped the entire medical system from collapsing, and at times it has been very close . In NYC in the f
236.
by adw 6y ago
I am not the right sort of doctor, and you should ask a medical professional. My knowledge comes from very out-of-date experience working in a computational drug discovery lab, which gives one a very narrow view of a very small number o
237.
by adw 6y ago
You'd probably be interested in https://www.thelancet.com/journals/lanrhe/article/PIIS2665-9... This paper points out that patients on infliximab (trade name Remicade: immune system modulator used by peo
238.
by adw 6y ago
> generative grammar/ symbolic approaches were pretty much left behind by NN methods Which is the same thing as hand-engineered feature stacks being left behind in vision problems, really. The story in every field is more or less &q
239.
by adw 6y ago
Standard distance measure in most atomic-scale condensed-matter fields. Certainly inorganic crystallography/materials science/condensed matter physics.
240.
by adw 6y ago
Carl Djerassi was of course himself a world-class drug researcher: he has a good claim to being the most important figure in the development of oral contraceptives for women.

More ›