How AI Discovered Hidden Information in Light
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Related videos
How To Spot Covid Misinformation
Dr. Sam Bailey
5.7k views
Mom Discovers Severed Head In Daughter's Bedroom
Dr Insanity
475.0k views
Australia’s Monopoly Is Hidden in Plain Sight
ColdFusion
355.9k views
The Highest Protein, Most Anti-Inflammatory Nut in the World has Been Discovered
Thomas DeLauer
31.2k views
FBI Discovers BURN BAGS Hidden In SECRET ROOM, Russia Gate COVER UP
Timcast IRL
42.9k views
AI is Burning - 5 New Papers
Discover AI
34.3k views
RAG is Exploding: 58 NEW RAG Methods in 48 hours
Discover AI
40.8k views
Police Discover A Killer’s Hidden Torture Forest
Dr Insanity
7.6m views
The AI Reasoning Lie
Discover AI
30.5k views
Graph RAG Evolved: PathRAG (Relational Reasoning Paths)
Discover AI
31.7k views
Top Comments (6)
We got a noise classifier before GTA6❤.
Solving over exploitation and over exploration of ai, one research at a time.
I don't know what for, but you're hired for any of our future research endeavours by merit of your accent alone. Unlimited funding.
6:10 In the future I guess a benchmark may be having an AI assistant attend video calls with humans, and basically act as an employee doing tasks.
i can listen to you everyday!!!!! keep it coming make me smarter!!
Very happy about this research, if they can't grasp the root cause at dozens of reasoning loops then they're really cool paper weights. I've experienced this personally, and it's the only reason I was following world models, purely as socratic models. I think a new training approach might improve this substantially, it's not a fundamental or architectural problem
Unlock the Data Inside
Turn Videos into Knowledge
- Get FREE 10/day: transcripts, summaries, chats
- Chat with videos, export text & PDF
- $1 free API credit for RAG, chatbots & research
Free forever plan • All features unlocked
Top Comments (6)
We got a noise classifier before GTA6❤.
Solving over exploitation and over exploration of ai, one research at a time.
I don't know what for, but you're hired for any of our future research endeavours by merit of your accent alone. Unlimited funding.
6:10 In the future I guess a benchmark may be having an AI assistant attend video calls with humans, and basically act as an employee doing tasks.
i can listen to you everyday!!!!! keep it coming make me smarter!!
Very happy about this research, if they can't grasp the root cause at dozens of reasoning loops then they're really cool paper weights. I've experienced this personally, and it's the only reason I was following world models, purely as socratic models. I think a new training approach might improve this substantially, it's not a fundamental or architectural problem