cchance

28 ms·

541.
by cchance 3y ago
Its another case of journalists knowing 0 about the product they're actaully talking about, FSD = Autosteer for them
542.
by cchance 3y ago
I mean what they're forcing tesla to do is annoy their drivers more, not actually change anything, mostly because tesla is actually better than most cars when it actually comes to safety.
543.
by cchance 3y ago
I don't have FSD speed over bumps anymore, however i wish it was adjustable as its gota pretty hard set at i think 15mph for speedbumps which is just a tad faster than what my wife feels comfortable with
544.
by cchance 3y ago
yep its best to have the ago for say hours but once it gets past a month or so its better to have dates
545.
by cchance 3y ago
white space is good, but this is too much white space
546.
by cchance 3y ago
So spin up a local llm with openai compatible api, and selfhost the inbox zero yourself lol it's opensource
547.
by cchance 3y ago
I mean it says you can self host it yourself... and its on github and opensource, and if it's using openai api, you can always use a self hosted api that's compatible with openai api to run mistral or llama to run it all locally.
548.
by cchance 3y ago
It's opensource and you can likely point it toward a standard openai compatible local llm for the inference :S It literally says "if you'd like you can self host it yourself"
549.
by cchance 3y ago
I mean... when you look at it like that, at least it's got the error checking integrated as well... tho the fact it pulls in is-number is fucking hysterical lol
550.
by cchance 3y ago
Not gonna lie, i really thought he was gonna end it with... now lets try 64bit, and leave us hanging
551.
by cchance 3y ago
Ya i feel like this issue is people think an LLM will someday "wake up" no, LLM's will just be multimodal and developed to use tools, and a software ecosystem around it will end up using the LLM to reason how to execute, basi
552.
by cchance 3y ago
Very nice, would be cool to have a little i next to each spot to explain what each thing is for newer users (batch size, etc)
553.
by cchance 3y ago
impl trait returns is so cool
554.
by cchance 3y ago
No ones thinking a 7b-70b LLM is going to be an AGI lol, a 700b-1T llm likely gets pretty damn close especially with some of the newer attention concepts.
555.
by cchance 3y ago
With which model? theirs like 6 of them
556.
by cchance 3y ago
holy shit perplexity is fast, even with the bigger models
557.
by cchance 3y ago
This has 0 to do with AI issues, and more to do with companies rushing things and likely buying the shittiest versions of things to save money, honestly im still laughing that people order at the window and fight to get the order across i h
558.
by cchance 3y ago
so ... recreate jwt?
559.
by cchance 3y ago
Gotta say i had to check wtf a nibble was lol
560.
by cchance 3y ago
Not sure on specific numbers but based on the picture he posted of it in a can he would have probably done better with metal netting from a window screen or metal chicken wire it destroys wifi signal pretty damn effectively
561.
by cchance 3y ago
it's called efficient Sam and it appears to be onpar or better than fastsam but did I miss a memory or speed comparison?
562.
by cchance 3y ago
He commented in another post that they use pooling so you don't really pay the cold start penalty as often as you'd think so maybe not a issue?
563.
by cchance 3y ago
I do find it a bit dirty to use better prompt techniques and compare them in a chart like that
564.
by cchance 3y ago
And Gemini ultra isn't available still as bard is Gemini pro
565.
by cchance 3y ago
Only pro apparently which is not as good as ultra, ultras the one that actually beats got4 by a hair
566.
by cchance 3y ago
Allocating is not spending, dear god don't clickbait topics like that
567.
by cchance 3y ago
i mean he got HN hugged
568.
by cchance 3y ago
Your literally talking about WhatsApp and Telegram as if they aren't their own segregated microcosms you can't message from whatsapp to telegram for instance so obviously they are their own segregated users by platform, its just a
569.
by cchance 3y ago
Its less about the color and more about the encryption + full res images and videos, and reactions working cleanly.
570.
by cchance 3y ago
From what i've seen every passkey implementation requires the device + a password or touchid or faceid to actually unlock the passkey authenticator... thats 2 things... the device key vault + your face

More ›