Yao Yue 岳峣

2.2K Followers
387 Following
447 Posts
Curiosity, compassion, skepticism, optimism.
LocationSan Francisco
Blueskyhttps://bsky.app/profile/yaoyue.org
Hubhttps://yaoyue.org
Workhttps://iop.systems
My 7yo quite enjoy reading the physical newspaper by the Onion. Not sure if I should be concerned.

The current and projected impact of AI and formalization on the practice of mathematics is analogous to the impact that the automobile had on the evolution of cities.

Before the introduction of the automobile, city streets were narrow and optimized for humans, horses, and carriages. When cars, buses, and trams were introduced, they were undoubtedly faster and more powerful than any prior form of transport; but they would clog the roads and crowd out pedestrians.

Over time, new roads, railways, and freeways were built for the exclusive use of mechanized vehicles, enabling rapid and efficient long-distance travel; but this came at the cost of urban sprawl, the degradation or destruction of once-walkable communities, traffic congestion, and significant environmental impacts.

It was only belatedly realized that to resolve these problems, it was not sufficient to simply make automobiles faster, more powerful, or more energy efficient, or to bulldoze all the old roads and networks to make way for new ones. Thoughtful urban planning, as well as the development of social and legal rules on how to manage traffic, were necessary to allow both pedestrian and automotive transport to co-exist in a manner that retained the benefits of both. (1/5)

@irene This is a bit different from what I was thinking. I think the hardest problem is always "unknown unknown", and we see a lot of younger researchers being completely unaware of work from decades back, or different disciplines/subdomains rediscovering the same pattern in different context. Identifying this type of relevance or similarity would be awesome, and it requires more breadth than any human can achieve simply by reading on their own.

ACM Digital Library's AI summary is a stupid feature because authors probably spend the most time per-word on perfecting the Abstract. This is made even more stupid by the fact that you need to pay money to see it.

What would be a more worthwhile "premium" feature is to use LLM to generate a graph of publications using both explicit references and LLM-assisted feature extraction to greatly improve literature review and related work research.

@midendian What bird is that?
I like how lowkey-inclusive the illustration on the 737 safety sheet is—woman holding a baby, dude with a prosthetic leg, and crocs.
In some sense the Alysa Liu story is like Monsters Inc. It turns out joy is more energizing than fear.
After watching some small amount of hockey this OG I couldn’t help but noticing how the men often had missing teeth while all the women mostly retain their perfectly pearly whites. Also women’s hockey was every bit as entertaining as men’s, if not more? To my amateur eyes the women seemed to do more intricate offenses. So in short, I don’t know what the men are risking their faces for.

If you thought you saw snow or hail in the Bay Area last night, you probably saw graupel.

Graupel is a weather phenomenon where snowflakes become coated in supercooled water droplets, giving them a soft, tiny Styrofoam-ball texture. It sounds like hail, looks like snow, and is terrible for making snowmen.

https://docpop.org/2023/02/thundergraupel/

Thundergraupel - Doc Pop's Weblog

The weather has been wild in SF. On New Year’s Eve, we had the second-wettest day in San Francisco’s history. Since then, we’ve seen much more hail and lightning than we usually get, but on February 23rd, 2023, we had a graupel storm. Graupel is distinct from snow or hail and tends to occur alongside […]

Doc Pop's Weblog
@tbortels @simon I’m not disagreeing that good engineers can review code directly and find serious problems quickly. I’m just more curious about how the solution or mediation changes when we take that option mostly off the table.