The reason that (2) is in question and open source AI R&D is threatened is (1). Those questions are not disconnected. If you ask "give me arguments for why turbojets and rocket engines won't kill people", I could respond by pointing to a dozen treatises on how to design and
@ylecun
-
Could Open Source SkyNet Have Prevented Terminator Takeover?
By
–
Since many AI doom scenarios sound like science fiction, let me ask this:
Could the SkyNet take-over in Terminator have happened if SkyNet had been open source? -
Defense of Open R&D in Artificial Intelligence
By
–
A defense of open R&D in AI, posted as a comment to a question by @tegmark
. -
Altman, Hassabis, Amodei Lobby for AI Regulatory Capture
By
–
Altman, Hassabis, and Amodei are the ones doing massive corporate lobbying at the moment.
They are the ones who are attempting to perform a regulatory capture of the AI industry.
You, Geoff, and Yoshua are giving ammunition to those who are lobbying for a ban on open AI R&D. If -
Backpropagation and Gradient-Based Learning: Future of AI Training
By
–
Actually, backropagation is one of the things that we know will perdure.
Certainly, learning through gradient-based optimization is here to stay.
Backpropagation is merely the way the gradient is estimated. To get rid of backprop, you would have to replace gradient-based -
World Model Optimization: Cost Function for Action Planning
By
–
An objective function C(f(x,a),y) that must be minimized with respect to a.
– x: current observation from sensors (or a prompt)
– a: imagined action sequence
– f(x,a): a model of the world that predicts the outcome of taking actions a
– C(f(x,a),y): a cost function measuring the -
Knowledge Versus Intelligence: Why LLMs Differ From Human Cognition
By
–
No human brain contains as much knowledge as Wikipedia.
But that doesn't make Wikipedia smart.
An LLM that could correctly answer any question whose answer is in Wkipedia would have more knowledge than any human.
But it still would not be nearly as smart as humans. -
Objective-Driven AI Systems: Evolution Since 2016
By
–
I've been making the case for objective-driven AI systems in talks since 2016.
But I've used the name Objective-Driven AI only recently.
E.g. here: -
Joint Embedding Predictive Architecture Evolution Since 1993
By
–
World models from video: since 2014.
Learned World Models for planning: since 2018.
Joint Embedding Architectures: since 1993, but a lot more since 2019.
Joint Embedding Predictive Architecture (JEPA) for images: since 2021.
JEPA trained from video: since earlier this year.