iiinova — where your digital self keeps evolving

Aram Harrow @harrowing.bsky.social · 06/06/2026

The Erdos unit-distance conjecture disproof was achieved using the publicly available GPT 5.5 Pro by Yichen Huang. github.com/yichenhuang/...

github.com

GitHub - yichenhuang/unit-distance

Contribute to yichenhuang/unit-distance development by creating an account on GitHub.

071

Aram Harrow @harrowing.bsky.social · 26/11/2025

You could likewise say that the US is just 5% of world population so what's the big deal if immigration to the US is restricted? Doesn't affect many people. The point is that there are big gains to migration, either between or within countries, and the main obstacles are policy choices.

000

Aram Harrow @harrowing.bsky.social · 26/11/2025

Network effects would cause these places to grow if they could, so stopping their growth hurts the US overall. Chicago grew from 30k to 1.1M over 40 years. We don't do that any more. Also the issue isn't preferences for cities, it's that jobs are created in places with housing supply constraints.

110

Aram Harrow @harrowing.bsky.social · 24/11/2025

Life expectancy in 1950s America was lower than in India today. On the other hand, we didn't have social media, so it's kind of a wash.

060

Aram Harrow @harrowing.bsky.social · 10/10/2025

Like Scott, I used chatgpt for a proof. It gave me something convincing but wrong. Still, it led me to a 1964 theorem about medians of sums of Bernoullis. And to fast algorithms for computing many power sums at once. Below was my first attempt. See if you can find the bug.

050

Aram Harrow @harrowing.bsky.social · 10/10/2025

Excited to post my fun collaboration with Angus Lowe and Freek Witteveen on how to truncate better using randomness. scirate.com/arxiv/2510.0...

1201

Aram Harrow @harrowing.bsky.social · 08/10/2025

The word "macroscopic" always confused me. This is like a W state, not a cat state, right? Do you know of any good discussion of this point?

200

Aram Harrow @harrowing.bsky.social · 15/09/2025

The money pays for research. If those contracts stop, the research stops. Yes private donations and industrial funding will still exist but the total amount of research will go down.

020

Aram Harrow @harrowing.bsky.social · 02/06/2025

Apple has to pay college-educated employees more. And colleges give training that's more general and more externally legible than you'd get from a company, thus raising the employees' market value more. So I see your point but it's not at all clear that this is a "subsidy".

010

Aram Harrow @harrowing.bsky.social · 02/06/2025

Endowments are used more for research than undergrad education.

010

Aram Harrow @harrowing.bsky.social · 20/05/2025

Thanks Earl! I'm glad you enjoyed it. We really found qDRIFT and your later paper with Berta+Wan inspirational and we think there's a lot of room for more use of those ideas still.

020

Aram Harrow @harrowing.bsky.social · 01/04/2025

"why" is always tough because there's no objective answer but the dirac plate etc are meant to illustrate that the fundamental group of SO(3) is Z_2, suggesting its universal cover should map to SO(3) via a 2-1 map. see also math.stackexchange.com/questions/11...

math.stackexchange.com

Universal covering of $SO(3,\mathbb{R})$

How do you prove that the universal covering of $SO(3, \mathbb{R})$ is $S^3$ ? Or equivalently, that it is diffeomorphic to $P_3\mathbb{R}$ ? Thank you for your answers.

000

Aram Harrow @harrowing.bsky.social · 24/03/2025

School closures were hard on kids, not only for test scores. I'm guessing you haven't talked to many parents about this topic? Maybe a necessary evil, but certainly not a trivial cost.

000

Aram Harrow @harrowing.bsky.social · 17/03/2025

It's even more impressive without the !.

140

Aram Harrow @harrowing.bsky.social · 12/03/2025

That's what all physicists hope for more than anything else.

170

Aram Harrow @harrowing.bsky.social · 05/03/2025

Selling exceptions is lucrative.

050

Aram Harrow @harrowing.bsky.social · 22/02/2025

Yeah, it is definitely cooler than its symmetric cousin.

110

Aram Harrow @harrowing.bsky.social · 22/02/2025

I don't find this claim too offensive. Yes they're no better at sending cbits but they're more powerful in other ways. "Richer information" would be better than "more information" but also raises more questions.

040

Aram Harrow @harrowing.bsky.social · 05/01/2025

craphound.com/category/una... It's more DRM than AI, but then again "AI" in most of these examples just means "algorithms."

craphound.com

Unauthorized Bread Audiobook | Cory Doctorow's craphound.com

120

Aram Harrow @harrowing.bsky.social · 05/01/2025

There is of course a Cory Doctorow short story on this theme.

100