Video

Question: based on the diagram, the user's prompt goes to the llm first, and then it's the llm who "asks" for the retrieval data? I've seen other videos and for what I understood, the user's prompt goes first to the search and retrieve system before it finally goes enriched or "augmented" to the LLM...

This chat gave me the final push to join my first competition! I’m a diamond league league of legends, tft, valorant and starcraft 2 player and I want to try to get to a similar level with data science! thanks for this chat!

Very interesting. It sounds as if Google is taking a dense vector representation for a chat rather than for a token and compressing it through an encoder from a trained auto-encoder for these long chat representations. Then this representation is continuously updated and always passed as context. I've been working on a hierarchical chat history stored in documents processed by RAG, but it does seem to me that some sort of sparse representation is also needed. I was thinking about graphical representations, like an HTM as Jeff Hawkins says may be in cortical columns, but this may be better. I still think the RAG idea is important, however. I've been using gemma:2b since it's very fast for multiple passes.

Could you send one copy, im first one asking you❤

It’s been 1 year, what’s the state of the art for translation?

you guys are rockstars! Cheers to a healthy positive future for humanity 💫

I think I just discovered for myself that my microfiber sheets were causing sleep disturbances. I mean no matter what I did, nothing worked. I watched hundreds of videos, read plenty of articles and NIH studies. But this one change, changing from microfiber sheets to 100% cotton made a dramatic difference in sleep quality. My guess is the toxic chemicals they use to make that cheap plastic was harming me, and disrupting REM sleep. I only found out because sleeping on my awkward couch was oddly better than sleeping on my serta king mattress with memory foam padding. Didn't make sense at first. I skimped on cheap sheets. Never again. I don't care what the cost is for organic, I'll never sleep on poison another day in my life. Hope someone reads this and finds the information as useful as I did.

Who’s this goof?

So 4 is more accurate but 4o is faster?

Having learned a lot from Towards AI articles and courses, I am eagerly awaiting the eBook edition. Congratulations!

I guess this explains why Skynet failed. Don't kill off human data labelers.

This was a waste of time. No disrespect but this is 95% just her talking about herself and nothing actionable or tangible.

Treat AI generated code like the intern hacked it together from Stackoverflow and bits from random projects on Github, because that is what it is.

Your observations are awesome and spot on. This is very helpful. Thank you.

Bravo, Louis-François!!! Fier de ton travail même si on ne se connait pas! Une chaince qu'on a le MILA....

Congratulations is the ebook or kindle edition available?

best explanation

Get the book now (leave us a honest review!): amzn.to/4bqYU9b

Hello sir ,I just finished high school and I am exited about AI and mathematic behind AI but I don't think a Phd is possible economically and I Don't think I will thrive in academia , I just want a non - trivial research work where I paly non trivial part not like 1 in 1000 programmers working on one problem but rather like 1 in 100 and also i am talking about research work(i don't want to direct research but be a part of it) not be a glorified programmer ,do you think that is possible with masters and if so how? I know you are busy on your pursuits but still I would really appreciate your reply

Thanks from Ukraine ❤🙏 in such a short period of time you put so many really useful information ❤️❤️❤️

Thanks! Very informative

can you do more of these where you talk about important papers?

Hey man! You seem to be very knowledge on Ai Topic, I joined your discord server and send you a friend request! I would like to chat with you about an opportunity! :)

Et cetera bien sur mon poto <3

Subbed

I agree 100% up until the stick part then i disagree 100%. People will be sheeps. TickToc kids are sad. You need more and more knowledge. We are hedding in to crazy time with segrigation being enormus. People thinking and the brain dead zombies...think snowcrash. But that is maybe just me. Best of luck great video. You succeed in trigger me ;)

I grew up without AI and I don’t know anything he said the generation will not know. AI will play a role in decreasing the education kids will receive but no matter what it comes down to teachers and the substance students are learning I learned more about genders than proper punctuation…. And it shows. i was born in 2000 and most of the lgbtq stuff didn’t get into schools until High school but that’s truly the most important years of learning before you go out into the real world but social studies was the same class for 7 years in a row all of middle and high school was the same dang that napoleon, african slave trade, and how people lived when Canada became a nation

Whos this weirdo

thanks for sharing mate

Great explanation. Thanks for this.

But it is not free when i downloaded it

Sure enough, the comment area is full of talents. I came to see the comments.

AI can do all of this

Cool, When is gonna be released

Cool man thanks. So when you change the embedding can you still process it with ChatGPT? I mean, is it only used for internal use and storage? Or can we reuse it somehow with a new prompt?

Great Content

Great thank you!

what software i will need for that pls

has any one experiences with the latency when multiple agents interact?

Thanks, very interesting and well explained!

Avec cette technique, peut-on espérer de vraies vidéos 3d ? En VR je veux dire.

These kind of details are so relevant when you really want to understand how things work. Great video!

I enjoyed this. Thanks

Nice work. Would love to see more such shorts.

Good work!

Well explained! I need to use this without becoming a coder.

This quality of CZcams shorts is really high you deserve more subscribers