0:00
/
0:00
Transcript

#18 - Scott Aaronson: quantum computing, AI and AGI progress

Scott Aaronson is a theoretical computer scientist at the University of Texas at Austin, known for his pioneering work on quantum computing and computational complexity. He writes the widely read blog Shtetl-Optimized and has shaped how researchers and the public understand both the possibilities and limits of quantum technology.

We talk about the reality of quantum computing, cryptography, AI progress, large language models, and what the future might look like when these technologies converge. Topics are outlined in the timestamps below.

Watch on YouTube. Listen on Apple Podcasts, Spotify, or wherever you get your podcasts.

Support my work here and follow me on Twitter here

Timestamps

00:00 – Intro

03:25 – How computer science views quantum mechanics today

06:50 – Superconducting qubits and how quantum machines are built

10:15 – The rules of quantum probability explained

13:41 – Quantum error correction and protecting fragile states

17:06 – When quantum algorithms provide a speed-up (and when they don’t)

20:31 – Skepticism and testing the limits of quantum hype

23:56 – Why Scott is optimistic about scalable quantum computing

27:22 – Potential applications: materials, chemistry, and beyond

30:47 – Shor’s algorithm and breaking classical encryption

34:12 – Bitcoin, cryptography, and the risks of a working quantum computer

37:37 – Grover’s algorithm and the reality of search speedups

41:03 – Large language models vs hard computational problems

44:28 – What tasks AI still can’t solve (and how to test them)

47:53 – GPT-4 vs GPT-3: progress, hype, and possible limits

51:18 – How companies train and deploy models responsibly

54:44 – The pace of change since ChatGPT launched

58:09 – Power and danger: capability without aligned goals

1:01:34 – Why AI is not just another technology but a civilizational shift

Discussion about this video