Restaldt

joined 2 years ago
[–] [email protected] 4 points 2 years ago

Don't worry bro itll all burn up on reentry

As if people forgot the law of conservation of mass

[–] [email protected] 28 points 2 years ago* (last edited 2 years ago)

Yes Dr. Ms.The Monarch

[–] [email protected] 6 points 2 years ago (1 children)

Got any strong opinions on trains or planes

[–] [email protected] 2 points 2 years ago* (last edited 2 years ago)

Facts.

Even if you could design your own model... How do you acquire a dataset even a fraction of the size those pretrained models from the corps.

Then how do you train the model in a reasonable time. Other than relying on cloud computing which leads to the same problem of only corps can play this game properly right now.

I designed and collected/labeled the data for a relatively small deep CNN for my masters thesis and training it on 60000 images was taking over a dozen hours (this was 5 years ago at this point so that part may be misremembered) on a 1080ti.

[–] [email protected] 2 points 2 years ago* (last edited 2 years ago)

50¢? Whoa now you arent trying to send us down the road of commies now are ya?

Better cut that raise in half at least

[–] [email protected] 5 points 2 years ago

Hey did you know the specific heat of water is 4.184 j/g°c because i learned that once in a physics class more than a decade ago and i still remember.

I am not a physicist and the memorization of that fact does not help me. Thanks brain.

[–] [email protected] 3 points 2 years ago (2 children)

I have... feelings about LLMs being the big thing in AI/ml right now.... because its really not much new. Maybe the transformer model kind of but ultimately LLMs are massive supervised learning neural nets trained on obscene amounts of data. And then other models use that pretrained "foundational model" to work and just tune their parameters. Which is why prompt engineer is becoming a thing.

Corpos are playing by the book here and trying to extinguish any competition before it begins by having people rely on their "foundation" models instead of innovating their own solutions

How many tutorials can you find for implementing LLM NLP tasks that dont include "import this model from X company" id wager its only maybe 33%

[–] [email protected] 5 points 2 years ago* (last edited 2 years ago)

Yeah they didn't particularly explain anything or present much of the data from the paper the article was written about.

They posed several questions that may be helpful for depressed people but my tldr for the article was people with major depression disorder may be subject to some kind of suvivorship bias for bad memories vs good ones.

Basically depressed people may be less likely to notice/appreciate the good times and more likely to ruminate/get stuck on the bad times creating a viscious cycle

Possibly because they overuse distraction as a coping mechanism leading to them being more distracted during good times

[–] [email protected] 1 points 2 years ago* (last edited 2 years ago)

Whats the number and locations of your associates?

[–] [email protected] 2 points 2 years ago

You bring up a good point of the Christo-fascist movement.

If all the jimbobs organize as zealots that could certainly lead to civil war with very messy borders

[–] [email protected] 3 points 2 years ago (1 children)

Ive never understood how the same crowd that spouts not your keys not your crypto would ever trust any password manager they havent personally read the source code for/compiled/self hosted.

Not your server not your safe/secure password

[–] [email protected] 19 points 2 years ago

Damn, Isreal really saw what happened in fallujah and went "Ooh write that down write that down"

view more: ‹ prev next ›