This is really cool -- the fifth busy beaver number has been verified to be 47 176 870.
I love busy beaver numbers because they're both an interesting pastime and also define the boundary of what is knowable in mathematics.
https://www.quantamagazine.org/amateur-mathematicians-find-fifth-busy-beaver-turing-machine-20240702
I enjoyed this talk on pixel fonts very much: https://youtu.be/SDI8ubVZi7w
what are some things that you find confusing about the command line? Mostly interested in answers from people who use the (unix) command line but still don't feel very comfortable with it.
(thinking of writing about using the command line interactively but I'm unsure about whether that would actually be helpful)
Upgraded my cable modem last week, which was a significant improvement to my internet. Should’ve done that way earlier!
The specific technologies in distributed systems may change rapidly, but I think the fundamentals are evergreen. We are going to be retrying, queueing, caching, load shedding, and sharding until the end of time.
UBI has been tried experimentally enough times that the only time a headline should say there are “surprising results” is if it doesn’t work. https://mastodon.cloud/@slashdot/112700889501338009
@nat Although an interesting twist on this topic at Labs is the swing towards “outcome > output” mantra that became predominant there, which I had never been very happy with for similar reasons as results- vs. process-oriented thinking. Too many outcomes weren’t ones that the team necessarily had control over, and I find it distasteful to measure against goals of that nature.
@nat My own intro to results-oriented thinking came from poker, rather than M:tG, which I think is both a game that more people know, as well as an even more stark demonstration of the importance of that type of thinking. (IIRC, it was perhaps Phil Gordon’s Little Green Book, but the date on that seems wrong, since I thought I was introduced to this earlier than 2005.)
Can't believe Little Bobby Tables is all grown up and has had their first kid, Ignore All Previous Instructions
Periodic reminder that NIST does not approve of expiring passwords.
https://pages.nist.gov/800-63-3/sp800-63b.html#memsecretver
> Verifiers SHOULD NOT require memorized secrets to be changed arbitrarily (e.g., periodically). However, verifiers SHALL force a change if there is evidence of compromise of the authenticator.
a funny tidbit from her talk, which comes from a study that was written up in a CCS '23 paper: people who trusted the AI to produce secure code were unlikely to have produced secure code, and people who did not trust it were likely to have produced secure code (see fig 2 from https://arxiv.org/pdf/2211.03622)
Managing Your Mac Menu Bar: A Roundup of My Favorite Bartender Alternatives
I understand that Applause has attempted damage control, but trust is something that you can’t get back that easily after such shady behavior.
Finding myself doing TypeScript type shenanigans at work, and I’m pretty sure this is me reaching peak midwit senior engineer in TS, so hopefully I’ll be over this hump soon and find simpler ways to attain the same confidence without jumping through type system hoops.
Or at least I’d settle for being able to more easily understand (or express?) the type signatures that I’m writing.