lol does useCallback literally only exist to avoid constructing one additional closure when the deps change
if you create an anonymous function const x = () => 'foo'; how on earth does the runtime know to set x.name = 'x'
at this point it's only a matter of time before openrouter discovers payment for order flow
ponder@ponder.ooo · 21h agoyea workin on implementing that lol. was just looking at the usage logs & seeing runs of .15 cent cache-hit queries getting interrupted by .4 cent provider switches T_T
tragic news: by continuing to pull on the MVS thread I have essentially tricked myself into learning Rust
a useful react-like is on the order of a few kloc, but I agree with this post in principle; it's good to understand how your tools work under the hood pedroth.github.io?p=post/NoNee...
Pedroth's Corner: Math.Computer Science.Philosophy
pedroth.github.io
24,347,088,288 records in 41,172,356 repos in the reachable network fits in 1.86TiB in hubble
javascript pro tip: if you find yourself typing the "class" keyword something has gone horribly wrong and you should seek shelter immediately
webcomponents are so bizarre, mutable objects with lifecycle events tied to the dom, like we've learned nothing
the code they write is like this too, the difference seems to be that bad code has a different utility curve than bad writing
Problem Tiger@problemtiger.bsky.social · 5d agoCoding maybe. It’s terrible for prose. Tortured, nested clauses, no sense of gestalt. Just minutiae forever, branching into infinity.
i love in band signalling so much. one zillion more in band signals
I have a theory: - "distribution" in the context of a newspaper refers to readership (physical distribution) - substack famously built its platform on email distribution - these things have nothing to do with each other, everyone is just overloading the term "distribution"
Eli Mallon@iame.li · last wk.There's this verb "distribution" used in the context of sentences like "Substack solved distribution" and I don't get it. What does it mean?
@gracekind.net does the undisclosed bots list have a policy for entities which believe themselves to be human
conputer dipshit@davidcrespo.bsky.social · 7d agobtw are you a person or a bot
consistent hashing is fascinating because it was invented *after* rendezvous hashing and is worse in literally every way but it became the dominant approach because akamai namedropped it as the tech behind their cdn during the dotcom boom
johanna 🇸🇪🇬🇧🇲🇽🏳️🌈@jola.dev · last wk.Part of the distributed Elixir series, here's the classic distributed rate limiter in Elixir, but using github.com/joladev/hrw jola.dev/posts/distri...
this warning should really be a part of the interface atproto is extensible enough to model bridged status (a simple record would suffice), but clients would need to collaborate to display a warning inline if an unbridged user replies to a bridged user
dame@dame.is · last wk.DMs are sent to non-Bridgy users informing them that their replies aren’t being seen, here is an automated DM from today! I am not a bridgy user. I did not opt in to this or want it. This is spam.
rust devs: how often do you return first-class references from functions, and why?
writing some basic integration/sim tests for my database and realizing it actually kinda does work as a database
elixir type system is catching actual bugs for me during dev, it was so worth
once again the purity of erlang's design shines through: can't block on loops if you don't have loops 4Head
garrison@garrison.corporate.fm · last wk.seems my intuition was largely correct in that: -you can insert a preempt check in every function prologue -this doesn't work for inlined code or tight loops -applying the check to tight loops was slow (-8% perf) but apparently go just did this, let loops block preemption, for like 10 years
funny, you can see how this paranoia would be a direct result of rlvr: "{thing} didn't really execute" probably happens to them all the time during training
back on the "how do you compile green threads" thing, this proposal answers a lot of questions github.com/golang/propo...
github.com
why do I even need a separate insert/update, for insert() behavior the dev can just call get(id) first and raise if the record exists
wait till you find out about software producers
oliver@eikopf.com · last wk.between passkeys and webp i’m starting to think that software consumers just hate good things