Yeah, they open sourced it, but most of the work is done by a neural network and they didn't share the weights of that. So it's still 2/3rds a black box. And that black box is changing every month (I think it actually is changing faster than that).
@scobleizer
-
Using AI Agents and Cognitive Architecture for Automated Data Reporting
By
–
It isn’t a tool. It is @blevlabs cognitive architecture. And the agents it builds. It is way way way better than any other AI. Had it grab 5,000 posts from my lists via X API. Then it wrote this report.
-
Using an AI agent to analyze curated community post lists
By
–
I created the lists over 19 years here on X. All by hand. Improved every day. The report was done by an AI agent created by @blevlabs
. It analyzed thousands of posts from my lists (I instructed it to grab 1,000 posts from each of my AI Community and a few other lists). -
Using AI Agents for Automated Research and Analysis
By
–
It is a big trend in AI community. Everyone is setting it up and getting it to do a ton of things. I'm having my agents go through thousands of posts here on X to write the ultimate guide and report about what's going on here. Be back soon. Created by @steipete
-
New AI Model and World Simulator Releases
By
–
🤖 AI Product Launches
— Robert Scoble (@Scobleizer) 23 janvier 2026
@MiniMax__AI launches M2.1 coding model – Chinese AI company with competitive benchmarks [https://t.co/iYgbOneo8e] (from AI Investors)@odysseyml introduces Odyssey-2 Pro world simulator – generates interactive 720p simulations [https://t.co/VD9Ogq6rmg]…AI Product Launches @minimax_ai launches M2.1 coding model – Chinese AI company with competitive benchmarks [
https://
x.com/MiniMax__AI/st
atus/2003550306073039221
…] (from AI Investors) @odysseyml introduces Odyssey-2 Pro world simulator – generates interactive 720p simulations [
https://
x.com/odysseyml/stat
us/2014781263224459293
…] -
Comparing AI Capabilities in Autonomous Robotaxi Systems
By
–
Most Robotaxi rides are single riders. Zoox is too limited and too expensive to operate and doesn’t have as nice an experience. And its AI isn’t nearly as good as Tesla’s.
-
Discussion on AI system architecture and training data transparency
By
–
There really isn't an algorithm in these new AI systems. More like billions of algorithms. And, no, I can't see Elon open sourcing his dataset that goes into training these billions of algorithms.
-
World Models in Robotics Training
By
–
True. World Models is what everyone in robotics is building. Trained on video data. See my latest post.
-
Defining AGI through embodied AI capabilities
By
–
New definition of AGI: can your AI cook dinner while embodied in a robot over an open flame without setting the house on fire? Five years away. Minimum. The world models, even those from Google or @worldlabs aren't ready. Until AI can do that I won't believe AGI is here.
-
Reflecting on the black-box nature of neural network weights
By
–
Yeah, I've been reading through the source code too. It's quite interesting. What's really interesting is that a lot of it is run by neural networks that we can't see inside of; we don't know the weights, for instance, but we can see the effect. My feed is generally better today