pehdrew.bsky.social
うさぎを追え。
43 posts
14 followers
60 following
Getting Started
Active Commenter
comment in response to
post
I recommend watching this video.
www.youtube.com/watch?v=_0h_...
comment in response to
post
can you link a paper?
comment in response to
post
I haven't see such proof, sure LLMs have lots of limitations, but they are very capable, like seen by GPT-4o for image generation, if they are able to add a system 1 and system 2 ability, they could become even more powerful, I'm not sure if you can or can't achieve AGI by doing this!
comment in response to
post
let's hope that Intel finally releases some good GPUs soon.
comment in response to
post
paper: arxiv.org/pdf/2402.03804
comment in response to
post
I've seen people using relu^2.
app.primeintellect.ai/speedrun/nan...
comment in response to
post
he left the tensorflow team, he said about working on other projects, this is nothing serious.
comment in response to
post
they got him! 😞
comment in response to
post
Deep Neural Nets. DNNs are technically AI, connectionist. people use AI to talk about NPC in games that use simple rules, AI is a broad term, should AGI come this would end the confusion.
comment in response to
post
.... . .-.. .-.. ---
comment in response to
post
careful with some of these words pal, you're running to close to the sun.
comment in response to
post
Agree I. statement this with
comment in response to
post
What's your overall view on AI in healthcare?
Some people see it negatively, which I find hard to understand—even considering that capitalism often prioritizes profit!
bsky.app/profile/pehd...
comment in response to
post
ignore the english mistakes, I haven't updated in a while.
comment in response to
post
he's does this from time to time. I expect him to delete some of tinygrad's line soon as well. Marie Kondo would approve I think.
comment in response to
post
I think this more about doing thingsearlier than late, if I have to edit an video tomorrow, why not today, still I'm not editing the batch of videos that i have today! 😌
comment in response to
post
I'm 4d tensor, be well as well.
comment in response to
post
often harmful is exaggerated though, we use "AI" in many for many things, already, it's only after the GenAI that people got a negative of it.
comment in response to
post
I disagree that it'll be a net negative of the average person, AI will improve material conditions, faster disease detection will save many lifes, this is just one of the possible positive outcomes.
comment in response to
post
I think the odds of achieving AGI in 5 years are low, not impossible, but extremely low. now AI sure will have many negative effects, but it'll have plenty of positive as well. and the positive are huge.
comment in response to
post
you gotta understand, it looks like someone on a open platform complaing the platform being open. it's confusing.
Data is valuable, that's why I'm here.
comment in response to
post
ok, things are clearer now! But I don't see we(humans) will ever stop progressing with tech, AI is the next step, it's a moonshot, yes, will it pay out, we don't know. But are people afraid of what of these scenarios: 1: AGI is achieved or 2: AI will suck for the next 20 years.
comment in response to
post
are we talking about email still or LLMs now? I think these two are totally different things.
comment in response to
post
I'd like to hear more!
comment in response to
post
I predict that they don't like von neomann as well.
comment in response to
post
X devs? that does ring the bell.
comment in response to
post
As long as Microsoft is backing OpenAI, money won't be a problem. They're are willing to spent tons of money, Project Stargate will be huge.
comment in response to
post
have you heard of a loss leader? sorry, I need to go into the real-world soon.
comment in response to
post
who are we here? plenty of people like LLMs. ChatGPT is the 8th most accessed website, not even 2 years old yet. MLLMs are truly powerful. The tech still under development, plenty of juicy yet to be squeezed.
comment in response to
post
I'm not sure, I am? is this real? wait, someone is talking about LLMs? are LLMs just Deep Neural Networks, yes, but why the user thinks it's different from any other type of classifier, it even uses the same loss function! I'm surprised. Should I be?
comment in response to
post
welcome to the ship pal! it's gonna be a numpy road!
comment in response to
post
You can't scare me with this Gestapo crap. I know my rights. I want my phone call.
comment in response to
post
I'm going to hang up this phone and then I'm going to show these people what you don't want them to see. I'm going to show them a world without you. A world without rules and controls without borders or boundaries A world where anything is possible. Where we go from there is a choice I leave to you.
comment in response to
post
I agree, Deep Learning should be the term used, DNNs for short.
comment in response to
post
this website surely will have lots of data for toxicity datasets lol
on the bright side it'll improve sentiment analysis models.