## Updatez

1. On the IBM Qiskit blog, there’s an interview with me about the role of complexity theory in the early history of quantum computing. Not much new for regular readers, but I’m very happy with how it came out—thanks so much to Robert Davis and Olivia Lanes for making it happen! My only quibble is with the sketch of my face, which might create the inaccurate impression that I no longer have teeth.
2. Boaz Barak pointed me to a Twitter thread of DALL-E paintings of people using quantum computers, in the styles of many of history’s famous artists. While the motifs are unsurprising (QCs look like regular computers but glowing, or maybe like giant glowing atoms), highly recommended as another demonstration of the sort of thing DALL-E does best.
3. Dan Spielman asked me to announce that the National Academy of Sciences is seeking nominations for the Held Prize in combinatorial and discrete optimization. The deadline is October 3.
4. I’m at the NSF Workshop on Quantum Advantage and Next Steps at the University of Chicago. My talk yesterday was entitled “Verifiable Quantum Advantage: What I Hope Will Be Done” (yeah yeah, I decided to call it “advantage” rather than “supremacy” in deference to the name of the workshop). My PowerPoint slides are here. Meanwhile, this morning was the BosonSampling session. The talk by Chaoyang Lu, leader of USTC’s experimental BosonSampling effort, was punctuated by numerous silly memes and videos, as well as the following striking sentence: “only by putting the seven dragon balls together can you unlock the true quantum computational power.”
5. Gavin Leech lists and excerpts his favorite writings of mine over the past 25 years, while complaining that I spend “a lot of time rebutting fleeting manias” and “obsess[ing] over political flotsam.”

### 38 Responses to “Updatez”

1. Dan Spielman Says:

I’d like to add that the full description of the scope of the Held Prize includes “combinatorial and discrete optimization, or related parts of computer science, such as the design and analysis of algorithms and complexity theory” and that the committee would like to interpret this broadly.

2. olivia l. Says:

Thanks for talking to us Scott! love the way the article turned out !

3. James Cross Says:

“I spend “a lot of time rebutting fleeting manias” and “obsess[ing] over political flotsam.”

4. Chaoyang Lu Says:

“numerous” silly memes and videos… I have already deleted a lot of those that require heavy Chinese Kongfu novel/TV/movie allusion 🙂

Dear Scott:

Re: Illustration by Joel Russell Huffman

For a young curmudgeon like you, a toothless smile wouldn’t quite fit in, would it? … But then the point is: Does it really show?

Huffman has captured, if you ask me, the contours for the eyes and the line near the cheekbones for the smile, so astounding well… granted, he had had ample assistance of technologies, but still… . These are among the two of the most difficult things to capture, if you ask me. [I was *very* good in sketching, and also pretty good in watercolour paintings, during my teens. Ask any COEPian of the times. Esp. during the College Election Poster times. Also, for the much less significant aspect of the calligraphy for the F1 club of the Bombay people. For the latter, ask Mr. Sandeep Johri — and also ask him to show a little better respect to me, at least now!]

However, a supply of the sketch in the .SVG format, rather than in the .PNG, would have been much more desirable, because it would naturally scale much better — I mean, in either direction, and, of course, within the degree of the approximating polynomial employed for the purpose.

But yes, Huffman has quite a good portrait to show. Much like those in the Indian Express, India.

Best,
–Ajit

6. Mystery Says:

Scott,

Right now, the only place in the internet to really openly talk about incel issues is incels.is—because every other site has censored us and kicked us off. You can go on incels.is and read some of the posts there and judge for yourself whether it was a good idea for society to deny us any other platform to voice our frustrations.

My request for you is this. My trolling obviously hugely disrupted your blog and forced you to change your comment policy. Yet, you haven’t really posted yet saying “Okay, I figured out what was going on, it was actually this lonely guy…” etc. etc. I think you should post something explaining that it was an incel who trolled and disrupted your blog, and also asking the question “what does it mean in our society that an incel would come and do this to me?” In other words, what happened to your blog is a microcosm of what frustrated, hopeless incels might do to our country and our society as a whole if their frustrations are not addressed. You could use this incident on your blog to invite bigger questions about how we deal with these people who feel left out of our culture, and how to mitigate the dangers.

Scott, I urge you not to do what Mystery #6 wants. All else aside, it sets up terrible incentives.

8. Scott Says:

Grad student #7: Indeed I won’t. After what he put me through—even trolling/impersonating a second time after confessing and promising not to do it anymore—I need a break from dealing with this stuff.

9. JimV Says:

Item 5 is an excellent resource. I had to bookmark in case anybody should ask me, “What’s so great about Shtetl-Optimized?”. The QC explainer was very good too. Keep up the good work.

(Angry reaction to #6 started but deleted. It would be a waste of time.)

10. manorba Says:

JimV #9:

“Item 5 is an excellent resource. I had to bookmark in case anybody should ask me, “What’s so great about Shtetl-Optimized?”. The QC explainer was very good too. Keep up the good work.”

i would add that item 1 is also an excellent primer on complexity theory.

(Angry reaction to #6 started but deleted. It would be a waste of time.)
+1

11. fred Says:

Mystery #6

“which was trying to carve out a space on the internet where we can talk about incel issues, and get a sponsor for us (yours truly) who’s actually reputable and has status.
[…] In other words, what happened to your blog is a microcosm of what frustrated, hopeless incels might do to our country and our society as a whole if their frustrations are not addressed.”

Trolling again, and this time with threats.
Here’s a secret, bro: everyone is frustrated and hopeless, and everyone has to learn how to deal with it.
You obviously have lots of time and energy on your hands, use them to learn to be quiet and reflect, maybe you’ll eventually grow the balls/ovaries it takes to figure your path in life on your own…

12. manorba Says:

Mystery #6:

ok, this time i’ll bite 😉

Mystery #6:
“[…] In other words, what happened to your blog”

what happened to his blog? exactly, what do you think happened?
as a regular reader i just saw a blogger quickly learning to deal with trolls and resume normal operations, and discovering along the way that a lot of people cares for him. Long way for a chubby shy nerd ah…
All in all you’ve done a great service to scott. i’m not sure that was your intention tho…

Just as fred #11 told ya, learn some basic social skills, and your life will change in a heartbeat. for example, realize that people are people regardless of gender and sex (or race, religion, political ideas…). not targets, mating partners, even friends. that comes later. learn empathy. learn to put yourself in the others shoes.

btw you may have so many “sponsors” (albeit not as shiny as prof. Aaronson) if you just put some effort in understanding the world around you.

And do you know why we are telling you this? Dude, we’ve all been there. I was your stereotypical chubby shy nerd too. i still am, at 53. i just realized along the road that everybody is just trying to survive.

13. 1Zer0 Says:

fred #11

“Here’s a secret, bro: everyone is frustrated and hopeless, and everyone has to learn how to deal with it.”

Please don’t use an universal quantifier for such a statement since it certainly doesn’t apply to me ^.^

Otherwise I agree with the rest of the statement + previous comments suggesting not to waste any time on that energy vampire.

14. P. Tran Says:

Scott, what are your thoughts about the recent claim by the Honeywell team of the first error correct quantum circuit which performs better than the equivalent physical circuit (https://arxiv.org/abs/2208.01863 and https://www.quantinuum.com/pressrelease/logical-qubits-start-outperforming-physical-qubits). Is this really the first example of a logical qubit circuit operation beating a physical circuit? Curious to whether this is really a landmark thing, or just one small step in the same direction that people have already been doing.

15. fred Says:

1Zer0

Maybe you’re not frustrated enough because you haven’t realized yet how hopeless things are! 😛

But as I was saying, it’s a matter of focusing on the right things.

16. fred Says:

Hey Scott,
I think you’ll find this discussion with John Carmack on AGI pretty interesting (I’ve skipped to the relevant part).
He’s the DOOM guy, worked at META on VR (still is part-time CTO for VR there) and now works on AGI full time. At some point he said he was invited to work at OpenAI.
His take on the risks of AGI is pretty intereting, down to earth as usual.

https://youtu.be/I845O57ZSy4?t=14568

17. f3et Says:

How come #6 passed the filter ?

18. Scott Says:

f3et #17: How come your comment passed the filter? 🙂

19. Dan Staley Says:

Mystery #6: I’ll give you the benefit of the doubt that you’re not intentionally trolling here.

If you want to talk about how you can improve your confidence and be less shy around other people (because I’m willing to bet it’s not just girls you have trouble talking to), there are places on the Internet you can do so. Maybe you’ll even get decent advice here, though it’s not exactly the focus of this blog.

On the other hand, if you want to go around referring to people as “Stacies” or “Chads” or other dehumanizing terms, or you want to talk about how much you hate people you don’t even know, well, I hope you can understand why that isn’t welcome here.

20. gentzen Says:

fred #16: While listening to John Carmack and his opinion that it might be just a small idea missing why we currently cannot get AGI, the old idea that the class of computable functions is a bad model of the historical contingent way human reason came to my mind. Other models, like computing with limited advice, or computation in the limit might be better models. But those are not good models either, because they ignore computational complexity bounds. But computation in the limit is interesting, when it is used to describe the interaction between different computing agents (or the interaction between a computing agent and its environment), the structure of which might be exposed when studying (generalized?) composition of limit computable functions.

And then another thought why computation in the limit is not a good model ocurred to me. That was 2 days ago, and I had forgotten it again. So I was just lying in bed, wondering whether I want to stand-up, while wondering what that other thought had been. After I managed to recall parts of what probably was that other thought, I decided to stand-up and post it, before I forget it again: To keep thinking after you already took a decision and gave a potentially incorrect answer is a nice strategy, and avoids “last word attacks”. But if you never commit to anything after a finite number of steps, like computation in the limit seem to do, then how should anybody interacting with you ever make progress? Somehow one would need something similar to how the infinite signed-digit representation of the computable real numbers strikes a balance between commiting to part of the information given in its output, while still reserving the right to modify it and increase its precision later. This is related to why representations are important in intuitionisitic mathematics.

21. Jonathan Baxter Says:

What happens when you ask DALL-E to paint in the style of DALL-E?

22. fred Says:

Do you guys think we’ll get soon enough a version of DALL-E that’s generating coherent “videos” instead of just static pictures or is that so qualitatively different than it may take a while? (I know that some people are putting DALL-E images together to do stop animation, but that’s different)
Rather than actual videos (from one fixed point of view) we’ll probably directly skip to fully 3D animated scenes.
Generating training sets could be a big problem, but maybe we wouldn’t feed actual 3D data but collections of videos, since AI is already able to generate accurate 3D models from a set of images around a subject.
Or maybe this could be done within a virtual world, like in a 3D video game like GTA5.
The memory requirement for this seems huge, but once we have that, the system would capture the world just as humans do: as things moving in 3D space.

23. fred Says:

You’d think that once we have good self-driving cars, it shouldn’t be hard to modify their model to be able to generate a realistic animated street scene (in 3D) from some text description (which refers to things that matter to it, like cars, pedestrians, sidewalks, etc).
Because the ability to understand something means also the ability to imagine a new plausible situation.

24. Bertie Says:

Hi Scott, I have seen you generously take time to answer some rather silly Qs from time to time, so may I ask a silly one?

Is there any connection between the phenomenon of an individual photon that can be shown to have been influenced by an allowable path (say through a beam-splitter) that it didn’t take and the ability of an electrical field to know which path has least resistance in an amount of time that is faster than light could have physically traversed the literal path (say a hugely long wire that finishes up right next to it’s starting point)?
Thanks!

25. 1Zer0 Says:

fred #22

Makes me think: Compiling the source code of a 3D engine and executing it is already “putting text to (fake) 3D”
You could probably construct some sort of procedural generation engine with Lindenmeyer systems (Like SpeedTree) and variations of Perlin Noise (Like No Man’s Sky Uses) that is taking seed values generated from common language text input and constructs a large worldspace from it. Well, the first part has been achieved already but I don’t think common language to seed values has been implemented yet.

26. John Says:

Here is a video I found on Youtube of Chaoyang Lu’s talk circa 2021 about the Boson Sampling experiment:

27. fred Says:

@1ZeR0

Ah, yes, thanks, I’ve actually watched that video when it originally came out, very interesting how they generate an infinite variety or interesting and realistic worlds.

On another topic, I’m done watching the MIT Quantum Physics I series of lectures that’s on youtube (quite challenging watching QM classes when working out!).

The very last class on QC is really well done, and Pr. Allan Adams even quotes Scott. He also mentions that some people who are both intelligent and reasonable do think that a large scale QC can’t be constructed (but that was 8 years ago!).

https://youtu.be/awpnsGl08bc?t=518

28. Cris Moore Says:

I am proud to say that Pan Zhang (coauthor of https://arxiv.org/abs/2111.03011 ) is a former postdoc of mine at the Santa Fe Institute.

29. fred Says:

This is what I meant by using virtual worlds to train AI. Eventually something like DALL-E should be train with this type of 3D scene representation in order to be able to perfectly generate images, videos, or 3D scenes (in the same way human painters or human animators perceive the world).

30. fred Says:

Correction: with such training data the AI would be perceiving the world (that’s virtually generated) beyond what humans can perceive, i.e. they would learn by perceiving 3D space (+ time) all at once, not just stereoscopic projections of the world from some point of view. The annotated spacetime block is their training set.

31. OhMyGoodness Says:

fred#16

I enjoyed your linked video and agree his approach is practical and pragmatic. He notes that GAI doesn’t need to work in exactly the same way as a human brain but then returns to relying on the human brain as a model for GAI intellectual development. I remember Feynman’s comments in this regard that all the money expended on education research and child development have had no measurable positive impact on education nor child development.

I don’t see that progress towards GAI is contingent upon realizing an accurate model of the human brain and continuing to constrain AI research to the illusion of how humans do it. A brute force approach is more likely to succeed and nothing at all wrong with that. It seems to me that a lot of intellectual energy could be saved by accepting that GAI research will culminate in something novel, or akin to an alien intelligence, rather than a human in a computer.

32. OhMyGoodness Says:

It might be Homo Siliconus.

33. OhMyGoodness Says:

…although Dr Aaronson might be more in favor of Homo Alanturingis.

34. Dong Kim Says:

Can you teach a course on Quantum Complexity? Your MIT OCW course doesn’t have any video lectures.

35. OhMyGoodness Says:

The Digital Party in Denmark is running for national office in the 2023 election based on policies claimed to be developed by an AI.

https://www.trtworld.com/magazine/can-an-ai-led-danish-party-usher-in-an-age-of-algorithmic-politics-60008

A brief survey of AI successes in diagnosing brain tumors. This is another area where AI’s now outperform human specialists. To be fair to us, humans collect a huge amount of high tech data to input to these models so the AI’s shouldn’t receive all the glory. 🙂

https://wetalkscience.com/brain-tumors-and-ai/

36. OhMyGoodness Says:

The US Navy is testing unmanned AI equipped solar powered sailing vessels under their Saildrone program (small sailboats). Recently an Iranian vessel hooked a tow line on a Saildrone in international waters and made for port. The AI requested assistance and two US Navy vessels responded and the tow line was released without incident. A Chinese vessel previously captured two unmanned USN underwater drones that were using only off the shelf commercial technology.

It looks as though the conclusion is that in order for AI war vessels to operate autonomously, and without nearby support, they must be able to initiate defensive measures.

Marine facilities, both fixed (oil platforms) and vessels, suffer from the same cost structure as aircraft and spacecraft. The cost goes up dramatically if they are manned. If manned then HVAC, food storage and prep, potable water, sewage, sleeping areas, rec areas, etc must be provided and that increases size and cost dramatically. There are large economic incentives to de-man marine operations as fully as possible and so another especially attractive niche for autonomous AI operation.

37. OhMyGoodness Says:

One other comment. Maybe there is some error in my logic but I see no way the universe is operated algorithmically, or could be simulated by a classical device, not because of quantum events but because there is no indication that the physical world is constructed of rational numbers. To the limits of metrology physical constants and values are irrational or transcendental. From the circumference and area of a circle to the hypotenuse of a triangle to the universal gravitational constant to the mass of an electron the values are not rational. In operation there is no indication the universe rounds off so infinite value accuracy must be used or at some time it will be discovered that there is in fact round off error manifest in the operation of the universe. Granted that accepted values have ppm to ppt accuracy but this accuracy is insufficient to maintain symmetry through the tens of billions of years of time evolution of the entire universe to date.

38. OhMyGoodness Says:

One of the John Carter novels (published around 1910) includes the greatest inventor of Barsoom (Mars), Fal Sivas. Fal cut open the skulls of many people to study their brains. He concluded that brains functioned mechanically (not chemically) and built an artificial mechanical brain about the size of a human brain that was capable of autonomous operation. He used these brains to build other brains and built an army that was autonomous but under his control.

It surprised me how close many of these ideas are to current thoughts (ex the mechanical function part). The self replication idea actually pre dates von Neumann.

You can use rich HTML in comments! You can also use basic TeX, by enclosing it within  for displayed equations or  for inline equations.

Comment Policies:

1. All comments are placed in moderation and reviewed prior to appearing.
2. You'll also be sent a verification email to the email address you provided.
YOU MUST CLICK THE LINK IN YOUR VERIFICATION EMAIL BEFORE YOUR COMMENT CAN APPEAR. WHY IS THIS BOLD, UNDERLINED, ALL-CAPS, AND IN RED? BECAUSE PEOPLE ARE STILL FORGETTING TO DO IT.
3. Comments can be left in moderation for any reason, but in particular, for ad-hominem attacks, hatred of groups of people, snide and patronizing tone, trollishness, disingenuousness, or presumptuousness (e.g., linking to a long paper or article and challenging me to respond to it).
4. Even when no individual comment violates policy, when there are dozens of comments from a single commenter hammering home the same few themes, and the commenter shows no interest in changing their views or learning from anyone else, the commenter will receive a warning followed by a 3-month ban.
5. Whenever I'm in doubt, I'll forward comments to Shtetl-Optimized Committee of Guardians, and respect SOCG's judgments on whether those comments should appear.
6. I sometimes accidentally miss perfectly reasonable comments in the moderation queue, or they get caught in the spam filter. If you feel this may have been the case with your comment, shoot me an email.