Mohsin Ali | AI
Region: AE
Friday 23 January 2026 12:55:58 GMT
80131
2565
150
755
Music
Download
Comments
R32_33Ok? :
Good and old but takes time depending on the model and the machine.
2026-01-24 00:33:34
2
serafius :
What is the technical term for this approach
2026-01-26 20:55:27
1
Bingbong :
Yeah, but you’re paying probably 5 to 10 times the API cost because the agent has to call all those tools but sure it’s more accurate
2026-01-24 03:49:05
9
TheoryOfMadness :
this is far inferior to a graph right approach that uses semantic relationships
2026-03-05 14:32:17
1
alfaric1 :
u deserve a follow but as son as i follow a page its stops showing up
2026-01-23 18:29:19
2
romainjouin361 :
way too slow
2026-02-22 05:14:45
1
888yea :
would be helpful if you mention the name of it in your video🤷♂️👍
2026-01-24 09:09:33
1
LARP Lore :
Still have to crawl and still have a end index to use what’s the point
2026-01-25 14:54:33
1
Blue_Robin :
Can this be done offline?
2026-02-02 02:54:21
1
Choice Royce :
does it give you page numbers or section numbers so the user can double check the answer given?
2026-01-27 08:26:20
1
Somar :
how to use
2026-01-27 19:33:30
1
emoracle :
lets call it an "index"
2026-01-24 14:37:34
2
ramli_mounir :
thank bro💪
2026-02-01 23:58:41
1
doctorintech :
Speed? And cost?
2026-01-26 10:32:35
1
TCN :
Thanks Ali, keep updating us AI Tools☺️
2026-01-24 08:12:26
1
James :
Awesome, but how do we use it?
2026-01-24 17:14:48
1
alangmartini :
how can I test it? 🙏
2026-01-24 16:10:34
1
Mark :
very smart, thanks for sharing 👍
2026-01-26 08:28:51
1
Mo | Dream Bigger Now! :
I wonder how updating the RAG data would work?
2026-01-25 02:52:07
1
nico3.14😼☮️ :
Higher upfront cost to build the tree and also higher cost to retrieve right?
2026-01-24 14:40:31
1
Grabbygui :
costs a lot more and a slower nah ?
2026-01-24 11:58:35
2
MadaParadise :
my problem with it: you use an LLM to generate the Tree. I am reading the repo and I will add other comments later
2026-01-24 09:29:53
1
adminnoy7tz :
That’s just normal chunking now. Did you think we were just breaking by character count? What year is it there?
2026-01-24 18:11:02
4
Alex Sciuto :
As someone who has made different RAG systems, traditional rag with Re-ranking is very good at capturing the users question.
2026-01-24 17:57:02
1
kabsikabs :
doesn't need chunks, but the agent research on what if it is not on chunks? if no vector de is not needed how are they persiste ?
2026-01-24 17:31:37
1
To see more videos from user @the_enterprise.ai, please go to the Tikwm
homepage.
