2026-03-23T18:25:04.188Z
thanks github copilot

the commit in question has a somewhat large commit, about a hundred lines of code modified, only one of those is a print statement, which was removed, and said neither "Hello" nor "Goodbye"
2026-03-23T21:41:22.274Z
Bo Burnham's Welcome to the Internet is outdated because you can't tweet a racial slur anymore. Not because you're not allowed to use racial slurs, but because it's just called "posting" now.
2026-03-24T22:32:47.001Z

or something like that
2026-03-25T03:23:15.809Z
you're telling me I can buy an air pump and nothing at the same time for just $23 more?

2026-03-25T20:22:54.608Z
top FRC teams coming into Miami Valley Regional next weekend and if they've qualified for Worlds already

ah, wrong image

looks like it'll be a crazy fight at the top
there's almost no way 4028 doesn't qualify for worlds at this event, 3 teams who haven't qualified would need to outperform them
2026-03-25T20:26:14.722Z
there's also going to be 5 teams at the event who haven't competed yet, they're mostly not that strong of teams historically but there's always a chance one of them breaks into the top
also, only a single rookie team, so they should automatically get the top rookie award, whatever they're calling it this year
2026-03-25T20:28:41.200Z
that's a side effect of there being 0 rookie teams from Ohio this year, which is rather unusual
2026-03-26T14:43:31.565Z
I nerd-sniped myself last night with this question:
Two players alternate secretly choosing attack or defend. (Player 1 chooses first, then player 2 chooses, and so on). If you choose attack twice in a row, you win. But if you choose attack and your opponent chose attack last turn, you win immediately. Does this game have a Nash equilibrium, and if so, what is it? (or if there are multiple, find all of them)
this seems like such a simple game that it would have to be analyzed already but I haven't found it yet
2026-03-26T14:45:00.650Z
these kinds of simplistic games with perfect information are well-analyzed, but this one has almost no information
2026-03-26T14:47:08.805Z
I believe any optimal strategy has to be probabilistic, with three relevant probabilities:
* The probability that P1 attacks on their first turn
* The probability that P1 attacks on any other turn
* The probability that P2 attacks on any turn
With the exception of P1's first turn, all turns seem to be equivalent to one another from each player's perspective
2026-03-26T14:50:09.525Z
I was running a Python simulation to iterate over the search space for general outcomes, but I just realized that I included (0,0) as a possibility to test and that's why it was taking so long
2026-03-26T14:51:45.426Z

2026-03-26T14:52:53.110Z
so in this chart:
color scale is outcome (white: A always wins, black: B always wins)
x-axis is B's probability of attacking, 0-1
y-axis is A's probability of attaching, 0-1 (o is on top for some reason)
2026-03-26T14:55:52.139Z
Takeaway: B should always attack. A should attack with probability .5. This is a Nash equilibrium under these assumptions, and B usually wins. However, this doesn't account for A's ability to change strategy on the first turn. If they know that B will always attack, they should never attack on the first turn, and then attack with probability 1 on the second turn. This will guarantee A the win, and force B to change strategy.
(sorry I changed from P1 and P2 to A and B. I've been using them interchangeably in my head)
2026-03-26T19:29:38.070Z
so some interesting results:
if you know your opponent will always attack with a certain constant probability on each of their turns, no matter what that probability is, you should:
* hold on your first turn if you're A
* attack on every other turn
(implicit here is that both players will attack if they've attacked on the previous turn. I've always been assuming that ftr. maybe I should simplify the setup somehow so that's not necessary)
but, critically, if you're P2 and you know that's P1's strategy, you should now hold on your first turn and attack every other turn
but, if P1 knows that's P2's strategy, P1 should attack on every turn
and if P2 knows that's P1's strategy, P2 should attack on every turn
and if P1 knows that's P2's strategy, then they should go back to the first strategy in this message
so basically it's rock paper scissors
except, if you believe your opponent will pick one of the strategies in this loop, is their a strategy that you can choose that will beat them all on average?
regardless, there is almost certainly no Nash equilibrium based on this analysis
2026-03-26T19:37:41.580Z
alternate formulation:
Two players alternate secretly choosing to attack or defend. Once the first player to choose attack does so, if the next player also chooses attack, they win; otherwise, the player who first chose to attack wins.
2026-03-26T20:06:03.097Z
Here's another formulation:
Designate two players, one as A and one as B. They each secretly choose a counting number. A wins if B's number is larger than or exactly one less than A's number.
2026-03-30T18:54:21.606Z
Whoever decides that the Streamdeck Neo would have its main buttons be actual buttons but the "change page" buttons be capacitive so they have zero activation force is dumb lol
2026-04-01T23:37:41.720Z
now that's a satellite https://commons.wikimedia.org/wiki/File:TDRS_Satellite.tif
2026-04-02T19:02:53.699Z
Zstandard is probably the best piece of software to come out of Meta
React might be the most influential, but definitely not the best
2026-04-03T16:15:15.272Z
So many great photos out of Artemis II, this is one I hadn't seen before though https://commons.wikimedia.org/wiki/File:Artemis_II_Launches_Through_Earth%E2%80%99s_Atmosphere_(CIRA_2026-04-01_-_labels).webm
2026-04-03T17:53:48.922Z
Temporal is so good that it's ruined my ability to work with other date-time systems
like, what do you mean Duration isn't a first-class type in your API? You want me to use an Interval instead? get real
2026-04-05T19:16:31.928Z
if today is Easter, when's Eastest?
2026-04-06T20:35:58.891Z
Reddit should have a "worst" sort option
you can sort by most upvotes, most balanced up/downvotes, we need a most downvotes as well
2026-04-07T21:56:36.767Z
I'm sure this commenter means in the sense of like, shifting international priorities towards space exploration

2026-04-07T21:57:56.286Z
but I read it in the sense of like, some mass hypnotism where the planet's population becomes inexplicably affixed to thinking about the stars, to the point where no productive activity back on Earth can continue
2026-04-07T23:50:33.027Z
I don't think that's accurate

ah that's the wrong version
well, whatever, pretend "September 1st, 2007" and "(to be announced)" are highlgihted