ireneista@irenes.space ("Irenes (many)") wrote:
basically we want to do some, uh, decision support to know we're not boxing ourselves into a corner with a long compile time that there won't be a way to optimize
ireneista@irenes.space ("Irenes (many)") wrote:
basically we want to do some, uh, decision support to know we're not boxing ourselves into a corner with a long compile time that there won't be a way to optimize
ireneista@irenes.space ("Irenes (many)") wrote:
yeah thirteen and a half minutes
ireneista@irenes.space ("Irenes (many)") wrote:
but the thing is that the work it's doing doesn't feel as if it should be that much, except for the part where it uses an array instead of a proper data structure for the metadata annotations. so we're going to comment out various parts of the code and see what that does to the runtime.
ireneista@irenes.space ("Irenes (many)") wrote:
we'll know how much more when this run finishes. we think about ten minutes but we've never written it down before.
ireneista@irenes.space ("Irenes (many)") wrote:
a regular build of the compiler takes about half a second, which feels ludicrously high to us but it's probably fine. a full build of the hex dump version takes, uh, more than that.
ireneista@irenes.space ("Irenes (many)") wrote:
.... but, sigh, before we do that we remembered that we want to do some measurements to more carefully attribute where all the computation time is going.
ireneista@irenes.space ("Irenes (many)") wrote:
it also really helps to flesh out our notion of what it means for these code transforms to, themselves, be transformed, so that's cool
jscalzi@threads.net ("John Scalzi") wrote:
I've typed "the end" and also it's the day the manuscript is due
RE: https://www.threads.com/@elmontkristiansen/post/DdSSWemjy9Y
ireneista@irenes.space ("Irenes (many)") wrote:
basically, we think it's both necessary and as clean as such a thing can ever be.
ireneista@irenes.space ("Irenes (many)") wrote:
hmm, okay. time for today's work on Evocation.
when we left off yesterday, we thought through a mechanism for the hex transform to add specific behaviors to the label transform when the label transform runs inside the hex transform. it's kind of a conceptually messy thing to do, but we have now slept on it and are in touch with our feelings.
dysfun@treehouse.systems ("gaytabase") wrote:
i can't believe they went back in time and named a physicist after walter white
jscalzi@threads.net ("John Scalzi") wrote:
Mild_shock.gif
https://www.npr.org/2026/09/14/nx-s1-5962190/supreme-court-mail-in-voting-trump
Boosted by pixxl:
gildilinie@beige.party ("Gildilinie Gremlin 🏳️⚧️") wrote:
i really need people to understand that when they scraped the whole web and everything that's in print
that is a fucking LOT of data. yes it has many examples of whatever you're asking from it. and it still sucks
Boosted by pixxl:
gildilinie@beige.party ("Gildilinie Gremlin 🏳️⚧️") wrote:
also im so sick of hearing like "could AI do this if it wasn't smart"
doesn't matter what "this" is the answer is always "yes there's plenty of examples in the training data (which is all the data ever data'd)"
Boosted by glyph ("Glyph"):
mattblaze@federate.social ("Matt Blaze") wrote:
Supreme Court ruling has dropped. Injunction stands! @heidilifeldman
Boosted by glyph ("Glyph"):
glyph ("Glyph") wrote:
a lot of misallocation of resources could be stopped if more people understood the difference between "performing a rigorous thought experiment" and "fantasizing about some cool cyberpunk shit"
Boosted by glyph ("Glyph"):
maxleibman@beige.party ("Max Leibman") wrote:
Petition to amend the DSM-5 to include “Started but failed to complete online form to register for ADHD assessment, 3-5 times over 6 months” as a diagnostic criterion for ADHD.
Boosted by glyph ("Glyph"):
mcc wrote:
"A human is reading it, so you don't have the privacy of using a computer. But a computer is writing the responses, so you don't have the benefits of a human reply." This is the worst of all possible configurations.
Boosted by glyph ("Glyph"):
mcc wrote:
RE: https://hachyderm.io/@evacide/117270680462821088
When I criticize "artificial intelligence", I usually focus on the problem that it is not really intelligent. However there is another, actually important problem artificial intelligence has, which is that frequently it is not really artificial
Boosted by glyph ("Glyph"):
mhoye@cosocial.ca wrote:
I don't think it needs to be AI specifically. Liability for software would benefit everyone in the country.
Boosted by cstanhope@social.coop ("Your weary 'net denizen"):
MaddieM4@raphus.social wrote:
A file is a big container of Maybes. Is a file a region of disk blocks that can be mutably edited in place? Maybe. Is a file a logical sequence of bytes which might be entirely in-memory and unpersisted to disk? Maybe. Is it an archive which needs to be hash-checked to confirm no corruption happened to it? I dunno man, maybe!
And then you need to think about the fact that some files support rewinding/skipping with fseek, but other "files" are network streams or `/dev/urandom` or UNIX pipes.
See, the problem with "everything is a file" is that the abstraction comes back up worse in your esophagus later, turned around: "a file is everything." Files are, in a specific sense, a violation of the UNIX principle, ironically enough: they do every job, and just kind of okay. In fact they're pretty bad on any filesystem with less than 5 developer-years invested. And you all know me: I am unusually willing to let difficulty cliffs tell me when something is a bad idea.
dysfun@treehouse.systems ("gaytabase") wrote:
rubs eyes no i read it correctly after all
jscalzi@threads.net ("John Scalzi") wrote:
Only tangentially related to the events of the day, but only now do I realize that my brain considered Hozier and Mackelmore to be interchangable and has mixed them up on a regular basis
brennan@social.lol ("Brennan Kenneth Brown") wrote:
I, CEO: How Medium.com went from being anti-AI to pro-AI
https://brennan.day/i-ceo-how-medium-com-went-from-being-anti-ai-to-pro-ai/
Boosted by ireneista@irenes.space ("Irenes (many)"):
xgranade@wandering.shop ("Cassandra has had an adventure") wrote:
@discatte @atax1a @ireneista Software needs far more unions, both of the labor and discriminated kinds.
Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
uglyreykjavik.bsky.social@bsky.brid.gy ("Ugly Reykjavik") wrote:
This summer I took photos of Toppstöðin. Sadly, the building caught fire today. Toppstöðin was a backup power station built in 1946. Recently, it was sold with plans to turn it into a recreation centre.
#Iceland #Reykjavik #photography #streetphotography #abandoned #decay #door #window #history
firepoet@tech.lgbt ("Stephen 🌈 (he/him)") wrote:
It’s tough being a deontologist in these times, especially surrounded by Utilitarians.
firepoet@tech.lgbt ("Stephen 🌈 (he/him)") wrote:
RE: https://mastodon.energy/@catalystcoop/117271156071094129
Wait what? Solving climate change and using LLMs to do it? Ugh. I can’t figure out whether to block you or not. I think I’ll have to lean towards Kant’s view instead of Mill’s.
Would I wish use of LLMs as a universal rule? No. Therefore, goodbye. 😥
Boosted by ireneista@irenes.space ("Irenes (many)"):
discatte@digipres.club ("(🖬🖫). ₍^. .^₎⟆") wrote:
@atax1a @ireneista unions? NEVER, how can I rest and vest!!!!11!!!
Boosted by ireneista@irenes.space ("Irenes (many)"):
ireneista@irenes.space ("Irenes (many)") wrote:
@atax1a yes!!!!!! yes, we need fucking class solidarity