byard

radiologist, former condensed matter physicist

[byard @byardedwards]

@Thiaagofernands @hsalis Yes. Tried a bunch of options and this kit (rigorous protocol especially) was best.

[byard @byardedwards]

@CJVogel_OTF I get all the focus on Arch, but we haven't been able to run the ball all year either. I think a lot more is wrong than Arch not being able to find the open receiver.

[byard @byardedwards]

@DavidePiffer @cremieuxrecueil clearly going to have a massive impact, given the overlap of physics and mathematics. However, I think the speedup in physics will be welcomed, not seen as a threat as we're seeing with math.

[byard @byardedwards]

Not an AI expert but the pace of progress is amazing and takes a brave person to confidently say "no it won't". However. I want to see it solve things like: High Tc superconductivity. No one thinks it requires physics beyond the Standard Model, just have to figure out the mechanism of how it emerges. We already have BCS for low-temp, tons and tons of high Tc theorizing, lots of experimental data. When AIs are better at theorizing than people, we should see rapid progress. And given the # of physicists at AI labs I'm sure this is being worked on.

Worst case it will speed research like this up markedly, b/c obviously they're great at math (and much mathematical physics). But will AI be able to say: "Ok, here are the reasonable (theory) options. Here are experiments needed to decide between them." etc. And once done, will they be able to say "here's a probable room temp superconductor with high critical current" etc (assuming such exists)

[byard @byardedwards]

@AAMortazavi pretty sure this varies by person. some people get a 95% quality answer lightning-fast but never get beyond that. some can grind along to a better answer, but not as good at quick answers. "your first guess is best" is not correct for everyone.

[byard @byardedwards]

@zach_barnett probably a grey area-- if it weren't, TN wouldn't have tried it. No point doing something obviously illegal, that's guaranteed to fail. we'll get clarification, and if it's legal everyone will try it, if not it's done

[byard @byardedwards]

@NFL_DF didn't the announcers say he overthrew the receiver or something like that? looks perfect

[byard @byardedwards]

@thrashman10 dude what about Biles??? He is amazing. All over the field, non-stop motor.

[byard @byardedwards]

@Obi_Kush @TexasisCursed 💯 A few feet shorter and the safety is in play. I hope that's a 90/10 ball for Coleman vs a 5'11" corner who is behind him (and most of the 10 are incomplete, not INT). Was a shot worth taking to a tall elite receiver.

[byard @byardedwards]

@CJVogel_OTF how did Biles rank? I think he may be best player on the team (although several reasonable choices)

[byard @byardedwards]

@CJVogel_OTF they have got to handle the blitz better. don't know how much is o-line/Arch/Sark, but they are going to get blitzed all day until they figure something out. the quick pass to Brown in vacated spot was a rare success

[byard @byardedwards]

@hp_claire44679 @NashTalksTexas completely agree. Coleman is 6'3" that corner is 5'11". given how talented Coleman is, you expect him to go up and get that. at worst expect an incompletion.

[byard @byardedwards]

@MsSummerblossom @AlexHarding7 @Basecamp_Res @AnthropicAI was intended as sarcasm. AI is super impressive in many ways, and hard to predict 5 years out with such rapid progress. Already speeding things up. But a lot of hype. Protein structure is not "solved" by AI, for example https://t.co/KSDGqf63RD

[byard @byardedwards]

@iskander Yes I think so. Advances that would've taken 100 years may now take 10, or 2, etc. And what that looks like is hard to get your head around. What I want to see is physics/bio progress. First, explain high Tc superconductivity. If we really have ASI that should be easy.

[byard @byardedwards]

@tesamorelin @avi_eisen nice. a similar picture showing "at least one is a boy born either M or T" might clarify how the double counting has this go from 13/27 for any one day to 1/3 if you don't specify the day

[byard @byardedwards]

@avi_eisen I would say: 13/27*(% not twins)+1*(% ID twins)+1/3*(% fraternal twins). This ignores, among other small corrections, the possibility that one twin is delivered just before midnight, and the other just after. But for $650k they'd probably want that too ...

[byard @byardedwards]

@avi_eisen fun question. I hope I would first ask "can I ignore the possibility that they are twins, b/c I don't know the % of all births that are (a) identical twins and (b) fraternal twins and that affects the result". If they said no, include that in your answer /1

[byard @byardedwards]

@curiouswavefn Coming from a physics background, I was shocked to hear top biologists say "in our hands" when discussing their research results. But then you learn. This is an important problem, but very difficult. Biology is probably the most complex thing in the known universe. Good article.

[byard @byardedwards]

@IMAO_ Yes! If you’re seeing it from below” (yellow in front) it’s ccw. If you’re seeing it from above (blue in front) it’s also ccw