“Fourth, profit allocated to investors and employees, including Microsoft, is capped. All residual value created above and beyond the cap will be returned to the Nonprofit for the benefit of humanity.”
AGI
-
AI Catches Up to Evolution in Visual and Speech Recognition
By
–
Well, we caught up with evolution in just a few years when it comes to visual perception, speech recognition, and some forms of reasoning and planning.
-
Understanding Digital Twins and Singularity Theory
By
–
Think I need to better understand your definition of digital twins and singularity in this case to comprehend the theory.
-
AI Assistants as Future Digital Infrastructure Smarter Than Humans
By
–
A panel on AI at the Paris Peace Forum today. Starting at 42'00, I plot the future of AI: all of our interactions with the digital world will be mediated by AI assistants that will eventually become smarter than us.
Because they will become a common infrastructure containing all -
Custom Instructions Cause AI Agent Web Browsing Regression
By
–
Yeah, I had to add a custom instruction and tell it to never browse the internet. Big regression
-
AI Existential Risk Fears Based on Flawed Ideas
By
–
The fears of AI-fueled existential risks are based on flawed ideas.
A WSJ piece by Princeton's @random_walker and @sayashk . -
Academia’s Role in AGI: Beyond Scaling Current Architectures
By
–
You might think that joining academia is like "retiring from the AI race" if you believe that making progress towards human-level AI is just a matter of scaling up current architectures and data. But then, you would be wrong.
It requires new ideas, new concepts, new principles. -
Knowledge Versus Understanding: Adaptation and Cognitive Plasticity
By
–
Knowledge is not understanding.
Reminds me of the folks who have tried wearing prism glasses that invert the world.
They get used to it after a few weeks of practice.
Then they get all confused when they remove the glasses.
But it only takes a short time to recover. -
Learning World Models Beyond PID Controller Tuning
By
–
It's a *lot* more than adjusting the coefficients of your PID controllers.
Building a model of your own dynamics is acquiring a kind of knowledge, but it's relatively simple.
However, learning a model of the world and of your possible interactions with it is fiendishly complex. -
Knowledge in Execution: Learning Without Language
By
–
The point is that there is a *huge* amount of knowledge in the execution.
Arguably, almost *all* the knowledge is in the execution.
Animals learn to execute complex actions as quickly as humans without the help of language. So whatever knowledge in contained in language must be