meta_pixel
Tapesearch Logo
Log in
Daily Tech News Show

Yann LeCun’s World Models Raise $1 Billion - DTNS 5222

Daily Tech News Show

Tom Merritt

News, Technology

4.81.5K Ratings

🗓️ 10 March 2026

⏱️ 29 minutes

🧾️ Download transcript

Summary

Amazon is implementing new safeguards to protect against outages related to generated code, and Google is unifying its suite of Gemini integrations inside Google Drive.


Starring Tom Merritt and Jason Howell.


Links to stories found in this episode can be found here.


Hosted on Acast. See acast.com/privacy for more information.

Transcript

Click on a timestamp to play from that location

0:00.0

This is the Daily Tech News for Tuesday, March 10th, 2026.

0:09.1

We tell you what you need to know, give you the important context and help each other, understand.

0:13.5

Ah, yes, that's what we try to do.

0:15.6

And today, Jan Lacoon is betting that LLMs are dead.

0:19.8

They're a dead end.

0:23.2

They're not dead. They're a dead end. They're not dead. They're a dead end.

0:28.6

But world models. World models are the future. Time to learn a new buzzword, folks. Yeah, we'll talk all about it. I'm Jason Howell. I'm Tom Merritt. Let's start with what you need to know with the big story.

0:36.5

Advanced machine intelligence, aka.K.A. AMI. Labs, that is Jan LeCoon's new world model company, has raised $1.03 billion. So let's just say a billion dollars and some change.

0:51.2

Jan LeCoon left meta, you may remember, last November, where he was chief AI scientist

0:57.1

for a number of years.

0:58.2

He left to co-found AMI and concentrate on world models as his bet for achieving human

1:06.1

level AI.

1:07.1

He argues that LLMs alone are not going to lead by themselves to AGI, and that world models

1:14.2

are the real missing piece here.

1:16.2

Now, world models learn from video, they learn from sensor data, interactions in the real

1:22.5

world to build an internal simulation of how objects and agents and time and other real-world entities,

1:31.2

let's say, actually behave. Lacoon compares us to large language models where he says

1:39.1

they mainly predict the next word based on patterns in the text, right?

1:45.0

So think of world models as something along the lines of common sense, let's say, common sense knowledge of the world.

1:53.0

The development is directed at cause and effect in real environments.

1:58.0

That basically means it has a potential to be very impactful, not only just

2:01.9

for general AI applications, as Jan Lakun believes, but also, once we start getting into,

...

Please login to see the full transcript.

Disclaimer: The podcast and artwork embedded on this page are from Tom Merritt, and are the property of its owner and not affiliated with or endorsed by Tapesearch.

Generated transcripts are the property of Tom Merritt and are distributed freely under the Fair Use doctrine. Transcripts generated by Tapesearch are not guaranteed to be accurate.

Copyright © Tapesearch 2026.