Title: The TRUTH About Quantum ComputingSomehow he has gotten to be the academic authority on whether quantum supremacy has been achieved, so I always check his latest opinion. He seems to be getting more confident, but it is still just a prediction. He says it may or may not be achieved. And that not to trust your RSA keys.
Date: 2026-05-13 @5:00PMAbstract: Yes, scalable quantum computing should actually work! Sooner than many expect, which will create a huge headache when it breaks the encryption currently used to protect the Internet. But no, we don't think quantum computing can do most of what the popular articles promise in AI and optimization and so forth. Come to this talk to learn about why!
And another one:
Title: Scott Aaronson - Theoretical Computer Science and AI AlignmentAlso on the subject, Dr. Bee dropped a new one today that starts:
Date: 2026-05-14 @1:00PMAbstract: I'll survey some areas where I think theoretical computer science, math, and statistics can potentially contribute to the urgent quest to align powerful AI with humane values. These areas include: the watermarking of AI outputs, mechanistic interpretability (including Paul Christiano's "No-Coincidence Principle," and succinct digests of the training process to aid interpretability), and theoretical guarantees for out-of-distribution generalization.
In Einstein’s theory of general relativity, the time an object experiences depends on its acceleration. But time in quantum physics works more intuitively – it’s a universal parameter experienced by every object in the same way. In a new paper, physicists say they want to use a special type of clock to test that difference. Let’s take a look.No, quantum physics does not say that objects can be in two places at the same time. It does not say that cats can be alive and dead at the same time.0:00 Quantum physics says that objects can be in two places at the same time. A group of physicists now says that they can also be at two times at the same time, ...
Quantum computers are often explained this way, but QM really just estimates probabilities for different place. Once you observe the object, it is in just one place. Her statement is like saying: probability theory says that a tossed coin can be heads and tails at the same time.
Aaronson is an AI enthusiast: [11:40] AI is what "I would regard as, you know, maybe the most consequential technology, uh, that humans are ever going to be building."
Update: At the other extreme, Brian Greene released a new interview with an AI skeptic. The guy says that the AI LLMs are just a stupid trick of little value.
Dear Roger,
ReplyDeleteAbout your update to this post: See my latest blog post, at my own blog.
In short: I don't like your characterization of Gary Marcus as an ``AI skeptic,'' and the `` guy'' says that the AI LLMs are ``just a stupid trick of little value.'' He doesn't actually say that, and neither do I think of the matter in that belittling way.
I can understand you Americans interchanging a ``guy'' with a ``gentleman,'' but the context does matter. It should not be missed, if you ask me.
Either in writing that update you were too much in a hurry, or your positions have some worse underpinnings, mostly [I am guessing here] to do with your own money and all, or something else which I cannot figure out.
Best,
--Ajit
In the interview, Marcus repeatedly argues that he successfully predicted that AI would not be reliable or intelligent, and that AI is being oversold, and that little progress has been made.
ReplyDelete