Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
8note
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
51 ms
·
901.
▲
by
8note
8mo ago
claude-code added the /insights command which might tell you what you are doing wrong, using your history. from the basics, did you actually tell it that you want those things? its not a mind reader. did you use plan mode? did you ask
902.
▲
by
8note
8mo ago
the barrier to entry for a lot of playing around is getting a working scaffold to be able to run all your testing from id expect it could pick you out a breadboard, a micro, some actuators and sensors, along with get a code deploy and run h
903.
▲
by
8note
8mo ago
why not print onto paper and put it in a safety deposit box? your X/Y resolution isnt going to be fantastic woth filament, but your Z is gonna really suck. you could instead print(emboss?) like a barcode on some paper, and encase that
904.
▲
by
8note
8mo ago
the annoying part is that with tool calls, a lot of those hours is time spent on netowrk round trips. over long periods of time, checklists are the biggest thing, so the LLM can track whats already done and whats left. after a compact, it c
905.
▲
by
8note
8mo ago
its a lot more expensive than the US properly controlling what weapons are leaving its borders. rather than arming the cartels to fight against the mexican government, thr US could just... not
906.
▲
by
8note
8mo ago
i dont see how that would change the ultimate "money grants too much power" if the government exerts less democratic power, money will still exert too much capitalist power
907.
▲
by
8note
8mo ago
what you are doing is largely a free text=> structured api call and back, more than anything else. ML related stuff isnt going to matter a ton since for most cases an LLM inference is you making an API call web scraping is probably the m
908.
▲
by
8note
8mo ago
i think yesterday it ate the whole context window in one thinking call. i bet in a week itll eat the whole 5hour throttle in one call too:P
909.
▲
by
8note
8mo ago
CAD is a 3D modeling tool. its not the only one, and other tools are better suited for something like an apple. you can still post-process it to get it right sized for a printer
910.
▲
by
8note
8mo ago
id guess for 3d printing. not the best tool for the job of making something organically shaped, but maybe they also wanted to run some aerodynamics tests on it?
911.
▲
by
8note
8mo ago
a lot of the why is encoded elsewhere in mechanical engineering at least - the tables, the formulas, textbook problems, engineering reports. one of the challenges to making a good data set might be around bad designs and why they failed. if
912.
▲
by
8note
8mo ago
does claude code let me control whats output when? verbose i think puts it on the TUI and i cant particularly grep or sed on the TUI
913.
▲
by
8note
8mo ago
its really popular, and having claude code makes it really easy to sumbit issues and make wild guesses about issues. i wonder how many are real and not duplicates. i doubt github is well setup to handle said volume
914.
▲
by
8note
8mo ago
as a problem, it means you need a ralph loop on top of it, if you want it to finish a problem without it waiting on a checkpoint
915.
▲
by
8note
8mo ago
famous last words before quantum physics hit
916.
▲
by
8note
8mo ago
the big problem with tools like that is that the agent can read all the existing code and docs when asked, but it cant read your mind. its always going to do better if you give it a stronger description of the problem, and give it some more
917.
▲
by
8note
8mo ago
do you have the right expectations? rather than new stuff for everyone to use, the future could easily be everyone building their own bespoke tools for their own problems.
918.
▲
by
8note
8mo ago
you can also make your way along the south coast out to the ocean, and quickly get out of civilization. the larger "Victoria" is quickly growing westward
919.
▲
by
8note
8mo ago
its been gorgeous in calgary for the past couple weeks. winter calgary has always been bright and sunny, but with this climate change, its proper spring with +15C or so
920.
▲
by
8note
8mo ago
there's also other options - your own parents, and other new parents and their parents. "it takes a village" is an old saying that isnt going anywhere
921.
▲
by
8note
8mo ago
I'd have guessed that teenagers dont have enough places where theyre unsupervised or otherwise surveiled such that they can have sex without thinking about the possible results of said sex. "its the advertising" could be anot
922.
▲
by
8note
8mo ago
i just imagine it as the swap between "human watching agent while it runs" vs "agent runs for a long time, tells the user over human interfaces when its done" eg. sends a slack. or something like gemini deep research. an
923.
▲
by
8note
8mo ago
how does zulip as a technical solution solve the ultimate problems with discord and gmail and so on - EU chat control, US porn control, etc? arent you as a zulip instance owner going to have to implement all the same stuff as discord?
924.
▲
by
8note
8mo ago
you could also use some code styling agent scripts that make todo lists of everywhere where there's bad architecture, and have it run through fixing those issues until its to your liking. theyre reasomable audit tools for finding issue
925.
▲
by
8note
8mo ago
its still not truly global but that seems like a bit pie in the sky. people still do useful work without a global view, and there's still a human in the loop witth the same ole amount of global view as they ever had.
926.
▲
by
8note
8mo ago
i just got finished making myself a stylus based cad app and a bit of web app for doing layout so i think im well setup for a leather-working and embroidering setup for tbe next while. just about finished making my sister a new wallet using
927.
▲
by
8note
9mo ago
its a useful way to solve those problems, but i dont think that means its the only way?
928.
▲
by
8note
9mo ago
i dont think theres anything that makes it essentiall that llms are non-deterministic though if you rewrote the math to be all fixed point precision on big ints, i think you would still get the useful LLM results? if somebody really wanted
929.
▲
by
8note
9mo ago
this sounds like a major benefit. i dont want my uber driver bragging anout how theyre going to shoot me before i get out of the car
930.
▲
by
8note
9mo ago
you can just watch the limit on the claude usage settings view. itd be nice to know how much the session context window applies wrt token caching, but disabling all those skills and stopping sending a screenshot every couple messages gets t
More ›