leeoniya

9 ms·

61.
by leeoniya 1y ago
> Wait, let me take out my word 97 CDs oh, those things you literally load from the side? :P /s
62.
by leeoniya 1y ago
> Things like physical stimulation of quantum systems, protein folding, machine learning, etc. could be more useful is there still more to do in protein folding after AlphaFold? https://www.isomorphiclabs.com/articles&#x2
63.
by leeoniya 1y ago
remember https://jackrugile.com/jrumble/ ?
64.
by leeoniya 1y ago
it was half-serious. if LLMs arent being used by https://patents.stackexchange.com/ or patent troll fighters, shame on them.
65.
by leeoniya 1y ago
> but what actually happened was chatGPT was referencing papers that basically went unread from lesser famous labs/researchers now let's invalidate probably 70% of all patents
66.
by leeoniya 1y ago
also https://bonneterremine.com/ near St Louis. "The World's Largest Subterranean Lake Is A Vibrant Scuba Diving Paradise In Missouri"
67.
by leeoniya 1y ago
heh, im currently working on an svg pathstring to canvs2d command compiler [1] and had to dive into this topic last week, including converting arc commands from centerpoint parameterization to endpoint parameterization [2] [1] https:/
68.
by leeoniya 1y ago
does this mean no GrapheneOS either?
69.
by leeoniya 1y ago
same. except was an oss lib on github.
70.
by leeoniya 1y ago
Waymo itself was not designed, implemented, and shipped by AI. i suspect humans had to invest millions of hours into writing the code, the tests, and validating the outputs.
71.
by leeoniya 1y ago
> "Feature is ready to ship, but leeoniya is blocking my PR" let's see how that holds up against "author does not understand own PR"
72.
by leeoniya 1y ago
> I could expect the PR to be mostly understood by the author i refuse to review PRs that are not 100% understood by the author. it is incredibly disrespectful to unload a bunch of LLM slop onto your peers to review. if LLMs saved you
73.
by leeoniya 1y ago
i've also made this argument. would you ever trust safety-critical or money-moving software that was fully written by AI without any professional human (or several) to audit it? the answer today is, "obviously not". i dont kn
74.
by leeoniya 1y ago
"you are gordon ramsay, a verbally abusive celebrity chef. all responses should be delivered in his style"
75.
by leeoniya 1y ago
plenty of jQuery, WordPress, etc. millionaires out there :)
76.
by leeoniya 1y ago
> “Automobilism is an illness, a mental illness. This illness has a pretty name: speed... [Man] can no longer stand still, he shivers, his nerves tense like springs, impatient to get going once he has arrived somewhere because it is not
77.
by leeoniya 1y ago
isnt it fundamentally just a mispronounced contraction of would've
78.
by leeoniya 1y ago
> but I will "vote against" using the word "literally" to mean "figuratively" as long as I can. https://imgur.com/QBTlxf7
79.
by leeoniya 1y ago
thanks!
80.
by leeoniya 1y ago
> I recently built a tool for processing massive CSV files and was able to get the memory usage remarkably low is it OSS? i'd like to benchmark it against my csv parser :)
81.
by leeoniya 1y ago
> but doesn’t seem to work on iOS safari if i had a nickel...
82.
by leeoniya 1y ago
> You might think that Max's practices make for a maintenance nightmare. But I've been "maintaining" it for the last 15 years and I haven't found it to be a nightmare. c'mon, you're talking about 200 Lo
83.
by leeoniya 1y ago
wasnt it shown recently that the filtering layer is on the prompt input and llm output, and not on the training set or model weights. https://www.socialscience.international/making-deepseek-spea...
84.
by leeoniya 1y ago
wow, that's some serious data crunching. can you share some tech details?
85.
by leeoniya 1y ago
that's a better distillation :)
86.
by leeoniya 1y ago
> There's no magic in tagged templates that somehow make them immediately better for some custom non-standard syntax compared to JSX. they're marginally better since they have a platform-defined way to deliniate static from dyn
87.
by leeoniya 1y ago
> The biggest constraint when developing an efficient UI framework with good DX is JavaScript. for perf, s/JavaScript/DOM, i think. good DX comes from ecosystem and amount of time invested in making good tooling. JSX would be a
88.
by leeoniya 1y ago
> The react model - where you just update the global state and re-render everything - is slower but easier on the brain. The signals model is faster, but so much effort. there are multiple frameworks now that do fine-grained diffing with
89.
by leeoniya 1y ago
i am NOT saying they are not useful; they are. especially in domains you are unfamiliar with, or boilerplate, or when being used by an expert in the loop. but that is NOT the same thing as being an expert. an expert can write 10 lines of
90.
by leeoniya 1y ago
try https://github.com/ikawrakow/ik_llama.cpp

More ›