Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
anuramat
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
7 ms
·
91.
▲
by
anuramat
4mo ago
again, I don't really see how 1 and 2 are rust specific; compared to c -- sure, but it's seems unfair: the type of rust software that needs a bunch of random dependencies usually wouldn't even exist in c in the first place; i
92.
▲
by
anuramat
4mo ago
only if you don't care about the output and half the features; otherwise you can just use "mv" to compile your documents to .txt
93.
▲
by
anuramat
4mo ago
but then you're limited to an undocumented subset of pandoc markdown
94.
▲
by
anuramat
4mo ago
> austral thanks, can't believe the idea isn't more mainstream -- I've never thought dependency safety could be a thing
95.
▲
by
anuramat
4mo ago
I tried using markdown+pandoc for my notes for a while, but I couldn't figure out even the most basic things, mostly because of the dozens of incompatible flavors; in no particular order: - formatting math blocks is mostly not a thing;
96.
▲
by
anuramat
4mo ago
> supply chain risk how is rust special in this regard?
97.
▲
by
anuramat
4mo ago
I wonder what languages people are using; I imagine smaller models would be decent at bash/python but significantly worse at something like rust
98.
▲
by
anuramat
4mo ago
> what most likely happened sounds super biased UPD just look up "anthropic david sacks" > epic are you unironically happy about what is happening right now?
99.
▲
by
anuramat
4mo ago
billionaires are all evil, we get it; do you have anything relevant to say?
100.
▲
by
anuramat
4mo ago
"if I fail at hacking you, what's the big deal? if I succeed, why does your security suck?"
101.
▲
by
anuramat
4mo ago
does it hurt your feelings, or what exactly is the problem? in my experience, fable is (was) less sycophantic, so when I'm brainstorming something, I get more pushback against bad ideas; "politeness" is exactly what I don
102.
▲
by
anuramat
5mo ago
so, AR? how is AI relevant here? also, since when is CLI outdated?
103.
▲
by
anuramat
5mo ago
still, far more effective than "NEVER FUCKING GUESS"
104.
▲
by
anuramat
5mo ago
fyi you can re-enable it with `{ "showClearContextOnPlanAccept": true }` in ${CLAUDE_CONFIG_DIR:-$HOME/.claude}/settings.json also find `"disableAutoMode": "disable"` useful, since I'm typically
105.
▲
by
anuramat
5mo ago
in my experience, "undocumented tricks" break as often as documented features like when they removed "clear context and execute plan" option after releasing 1M opus because "context window is not a problem anymore&q
106.
▲
by
anuramat
5mo ago
I'd say "plagiarism machine" implies that
107.
▲
by
anuramat
5mo ago
you know you can set the temperature to zero and get exactly that, right?
108.
▲
by
anuramat
5mo ago
why would anthropic do that?
109.
▲
by
anuramat
5mo ago
> wrong 10% of the time that doesn't sound nearly as bad as you think it does; I don't see how ethics are relevant here either, unless oil is also somehow a scam regardless, one must be delusional to deny the fact that it'
110.
▲
by
anuramat
5mo ago
you think it should be deterministic?
111.
▲
by
anuramat
5mo ago
now I know what to call it, thanks
112.
▲
by
anuramat
5mo ago
so, hashline?
113.
▲
by
anuramat
5mo ago
was wondering myself, just tried comparing to petnames crate -- gets you about 2 tokens per word on average not that anyone should ever care; typos in random-looking ids are very real but already covered by human readable ids besides, this
114.
▲
by
anuramat
5mo ago
I think their point was that the project is complex, with the implicit assumption that the complexity is to a large degree inherent. Even if it's mostly accidental, and the code is overengineered slop (which it is), the system being ab
115.
▲
by
anuramat
6mo ago
There are actually multiple ways to completely eliminate the cognitive load of any sort of activity, so please, don't settle on half measures
116.
▲
by
anuramat
6mo ago
vicode -- TUI coding agent written in Rust, with tabs/subagents running in worktrees on top of fuse-overlayfs: create/fork tabs to work on multiple features/implementations, while subagents work in parallel without conflicts;
117.
▲
by
anuramat
6mo ago
you forgot to replace the em dashes :^)
118.
▲
by
anuramat
6mo ago
I'm pretty sure an average machine will always be less capable than a datacenter, the rest depends on your definition of super capable
119.
▲
by
anuramat
6mo ago
makes sense in a philosophical debate or when you're talking to your confused grandparents, but does anyone on hn not know how LLMs work, at least on the level of "tokens, matrices, data, sgd"? otherwise, that reminder must
120.
▲
by
anuramat
6mo ago
"almost as good as opus at writing python/js/... when given a spec" might be enough for a lot of people, especially if its 10x cheaper
More ›