Unable to load image

Reccomend me some podcasts on AI and ML

I realize I don't understand anything about either subject and the oreilly book is like $80

I want stuff that is fairly technical if possible. Ebook recs would be good as well

12
Jump in the discussion.

No email address required.

The issue about informational content about anything AI related is the gap between technical depth and accessibility.

Do you have a background in this field? Do you know why only attention matters? Can you tell the difference between a CNN/RNN? Does it even matter? - Or do you just wanna keep up with latest developments? If that's the case I'd just recommend reading HN whenever a new 'groundbreaking' breaktheough is achieved.

In case you already have some knowledge, Edan Meyer does good videos on selected topics in AI. He doesn't go into any actual technical detail but without a very solid math background one would be lost either way.

I quite like his video on Stable Diffusion which includes an explanation and some actual hands-on examples.

Jump in the discussion.

No email address required.

I want more than a fascile understanding of AI and ML and even though I wont understand most of it I want to know what goes on in an LLM. Like how even though I'm a brainlet ai still kind of understand how a computer works down to the silicone level.

I got into some of it, how the model processes nodes, and stuff about weighting and it's all v interesting.

Jump in the discussion.

No email address required.

Then I'd recommend the following.

I love this book: https://www.deeplearningbook.org

This book can be useful for a variety of readers, but we wrote it with two target audiences in mind. One of these target audiences is university students (undergraduate or graduate) learning about machine learning, including those who are beginning a career in deep learning and artificial intelligence research. The other target audience is software engineers who do not have a machine learning or statistics background but want to rapidly acquire one and begin using deep learning intheir product or platform.

It uses the following architecture. Feel free to skip a chapter if you think it isn't worth the time:

https://i.rdrama.net/images/17013128305592773.webp

Besides that, Microsoft offers good courses for ML Beginners and AI Beginners. Chiobu linked that below.

Good luck.

Jump in the discussion.

No email address required.

What's the likelihood that this branch of AI hits another brick wall?

I guess the math is advanced enough you can totally apply to other fields even if this peters out?


Follower of Christ :marseyandjesus: Tech lover, IT Admin, heckin pupper lover and occasionally troll. I hold back feelings or opinions, right or wrong because I dislike conflict.

Jump in the discussion.

No email address required.

In all honesty after some cursory research I think this version of ML will change computing. I think within a decade optimized LLMs will be doing most of the high level computing inside everyday devices. The applications are endless, and in lots of cases programs that now exceed millions of lines of code could ostensibly be replaced with an LLM

Jump in the discussion.

No email address required.

I think people who say it's :marseysal: are r-slurred for sure, but I am curious what the limitations looks like w.r.t. "hallucinations". Not from an "ethics" BS perspective but from a "if X is supposed to do Y but doesn't do it reliably, that's really fricking annoying" from a user perspective..


Follower of Christ :marseyandjesus: Tech lover, IT Admin, heckin pupper lover and occasionally troll. I hold back feelings or opinions, right or wrong because I dislike conflict.

Jump in the discussion.

No email address required.

Near future: they design very controlled datasets and scrutinize the output. End user LLMs will be restricted wrt outputs and inputs.

Long term: they figure out how the neural network actually works and stop it from hallucinating

Jump in the discussion.

No email address required.

:#marseycomradehandshake:

Jump in the discussion.

No email address required.

Awesome, thanks

Jump in the discussion.

No email address required.

Link copied to clipboard
Action successful!
Error, please refresh the page and try again.