Peter Finin

Lots of other universities and schools put tremendous attention to their football rivalries, too. But football is fundamentally a game about an infantry squad, not combined arms, air superiority, or any of the other stuff that defines US military might.

US military power is fundamentally based on ships, tanks, missiles, planes, helicopters, satellites, and a lot of other industrial machinery. But the annual Army-Navy football game is still A Thing, and plays a big role in how two of our armed services compete to display their martial prowess.

I'm excited to hear that BindCraft2 sounds like it's coming along well. Any plans for when this might be publicly available for other people to use?

The subsequent blowback and resistance after that may or may not be as widespread and severe as if they try to block all or most voting nationally.

screwing up voting in a few dozen strategically chosen precincts in swing districts that are looking close. That could potentially be enough to swing control of Congress in the midterms.

Their possible paths to stealing the election depend heavily on how close it's fixing to be. If they're on course for a massive blow out loss, then they probably don't have the manpower to meaningfully swing the outcome at the polling locations, as you say. But they have a shot at meaningfully

collaboration with the archeologist Patrick McGovern based on analysis of residues found on Iron Age drinkware from the Midas Tumulus in Turkey. It's much different than most modern beers, but really good, and it sounds like it's just the kind of thing you've been wishing for.

@patrickwyman.bsky.social
I just listened to your mailbag episode. You mentioned an interest in trying ancient beers. If you haven't already tasted it, you should try to get your hands on the beer Midas Touch, which is made (all too rarely) by the brewery Dogfish Head. It was designed in

It seems reasonably likely that the fluent Spanish speaking agent is himself Hispanic. I'm continuously astounded at the confidence that so many people have that the leopards are not going to eat _their_ faces.

I had had a higher opinion of Taubenberger than that.

UAB is a big contributor to Alabama's economy, and typically brings in many hundreds of millions in NIH research funding each year.

This is really interesting stuff! Is part of the problem that AF2 was trained on primarily well-folded proteins, so it has a bias towards predicting that proteins fold well, even when they don't?

A lot of times the explanation is crippling depression.

it may put them in a position that further incentivizes even crazier behavior as we approach the midterms.

TL;DR - The political headwinds make gerrymandering more complicated for Republicans. For those of us who'd like to see the backside of the current crop of MAGA GOP, this may seem appealing. But if they misjudge, which they are currently structurally inclined to do,

Unfortunately, I think this increased uncertainty may translate into increasingly erratic and desperate actions from a variety of actors as we approach the midterms.

who surrounds himself with yes-men, who often demands that his subordinates take a maximalist position even when it is risky or outright self-defeating, though, there is a ton of room for disaster and there's a lot more uncertainty about the range of electrical outcomes.

In a situation where you are good at weighing lots of complicated information, and rationally assessing the risks and tradeoffs, this is already hard. In a situation where everything is being directed by a relatively unintelligent and senile leader with little willingness to consider details,

If there's a realistic product of a big electrical swing against you, this complicates the gerrymandering calculations a lot. If you dig too greedily and too deep, you risk a political catastrophe. So you have to balance the potential gain against an acceptable amount of risk.

So there is a risk that if you misjudge, perhaps because there's a big swing in public opinion, you actually set yourself up for an electoral blowout, and lose far more seats than you would have with more naturally drawn districts.

Gerrymandering really only works when you can predict voting reasonably well. The whole premise is that you arrange to win as many seats as possible by as small a margin as possible.

you try to arrange that every district has the same breakdown and you take 100% of the seats.

If you're expecting a majority of the votes in the state, the goal is to evenly distribute the vote mix over all the districts, even if they naturally tend to be geographically clumped up. So if you're expecting the votes to go 55% - 45% statewide,

This makes your opponent waste a bunch of votes and lets you plan a bunch of districts where you'll get 50% + a little bit. So you get more seats with fewer wasted votes, and get get a majority of the seats with a minority of the votes.

The general idea in gerrymandering is that any vote for your side in excess of 50.1% is wasted. If your side is actually going to get a minority of votes, you want to pack as many votes against you into one (or a few) district where you'll lose in a massive blowout.

Any specific plans for when you'll make Protein Hunter publicly available?

This looks really cool! I'm just a simple wet-lab scientist, but I've been using some in silico designed binders to support my research. There are a few targets that BindCraft hasn't been able to generate a binder against, and which I'd like to try Protein Hunter.

Orban came back into power fifteen years ago, at which time he was about thirty years younger than Trump is now. All other questions aside, simple actuarial tables suggest that Trump does not have that much time left.

An MIT surveyor came upon the gates of hell. He looked the Devil in the eye and said, "You're looking well." The Devil looked right back at him and said, "Why visit me? You've been to hell already' cause you went to MIT."

Data so good you don't even need to look at the p value.