Aleks Singer

@chadsonofchad Would love to have you speak at vexsummit.com on fixing private markets for the space industry!

Incredibly exciting time to be investing in the space economy

This will be a great keynote! Come to Private Markets Summit!

Karpathy's Autoresearch pushed my vibecoded Rust chess engine AI from "expert" to a top 50 grandmaster, a #311 chess engine. It ran over 70 experiments on its own and tried to hill climb to the top ELO score it could, landing at 2718!

@chrislakin @Altimor @grok Yes I love the monoculture :)

@Altimor 25% of people in SF work in tech? what? @grok

People forget what a small town San Francisco is. 800k people, only 220k of which in tech. With the average job stint lasting 2.5, that's 88k people on the market every year. OAI and Anthropic each hiring ~4k in 2026 -- 10% of the entire market, from 2 companies.

Can't wait for the Private Markets Summit!

@NathanpmYoung Great use case for privacy cards. privacy.com/virtual-card

Adobe are one of the few companies I pay but actively dislike. Sure, their products are good but their choices on subscription, copyright make me hate them.

4/ We will also keep fighting the dangerous interpretation of 18 U.S.C. §1960 that treats software development as “money transmission.” That’s why we support Storm’s appeal, the BRCA’s passage to clarify the law, and Michael Lewellen’s civil case seeking a declaratory judgment.

The global economy is held up by SF lying to NYC about how fast AI is growing and NYC lying to Abu Dhabi about how safe private credit is

@cremieuxrecueil I'm a lifelong learner. That must count.

If you say you're a student, Apple will knock another $100 off the price.

@FutureJurvetson Is the application here mostly defense tech? Or do you see this being used mainly to improve data centers?

📲 New details on ZaiNar's geolocation service — the first 5G positioning system delivering sub-10cm accuracy on existing networks. No device software or hardware needed. A new utility for carriers. Deets: zainartech.com/news/5g-killer…

From a conversation with @Ryan_Singer

Negligence and duty of care aren't different because the software is an LLM. There is nothing Claude can do that someone good with Google couldn't have done 10 years ago. Claude just does it faster.

We should retire the term "AI alignment." The existing vocabulary from software quality, consumer protection, and product liability works fine: does it do what it claims to do, does it mislead users, does it cause harm, who is liable when it fails?

@NathanpmYoung How many total people did they contract to do it? Only 5 or are there more and you're only displaying the top 5?

Metaculus contracted me to forecast against the community with the winner getting $2500.

At this point in the @Metaculus Spring cup, I am 3rd, having outperformed the community forecast on 5 of the 6 questions so far.

By the way, I will accept being-right points for the general notion that “my god, we cannot trust the government with the regulation of this technology, we will need new kinds of private institutions for this” at your leisure

@NathanpmYoung China won't slow down.

I don't know if this protest is the right way to go, but I wish that if AI CEOs wanted to slow down, they could. Hassabis has said he was open to it, Amodei says he'd like more time. It seems like coordination problem: luma.com/s0k8wvee?tk=GP…

@sodofi_ Luckily that's rare

san francisco is the land of hard work and discipline unless it’s sunny outside. then it becomes the land of partiful picnics

@yishan This is way bigger than Kinsey.

This is actually a contribution to human sexuality research on the level of Kinsey Reports. Aella has fearlessly combined a willingness to transcend taboos (merely *talking frankly* about sex is one that stops most people) with data science and coding to collect and organize what is probably the largest and most comprehensive data set on human sexuality in the world. To “academic researchers” who want to downplay this with criticisms that it doesn’t follow “proper” research data collection procedure, or that analyses are flawed, I will say this: - when the Kinsey report was first published, all the same types of criticisms were leveled. Professional statisticians condemned his sampling procedures. - other “real” academic research on sexuality today are typically based on far smaller datasets whose collection and interpretation are just as subject to (if not more so) the same questions about validity/biasing - Aella is exceedingly careful about her data and methods, and notably if challenged on them tends to take criticisms seriously and correct any errors, something I don’t see nearly as often from “real” academics.