Paul Hsieh

We need to do disconfirmation on an absolutely massive scale. I think about this all the time now. The first step is to go after scholarly misconduct, plagiarism, falsified data, and p-hacking. Then we need to investigate which scholarship rely on studies that have since been refuted, discredited, or retracted. The result will be that we discover we know a lot less than we think we do. But that is how reliable knowledge is created. AI will allow us to do this on an epic scale, and we should lean into that. One of the great discoveries in the history of the species is that our brains lie to us about how simple and clear the world appears. We are extraordinarily good at self-deception. It is only through disconfirmation that we can complete the arduous process of seeing the world as it is. We have been handed a huge opportunity to vastly improve our understanding of the world. We must not shrink from it.

Whether on foot or on bicycles, bears must stay on the trail.

More from the author: x.com/bjornlomborg/s…

“Gore’s apocalyptic climate predictions have aged poorly.” newsweek.com/data-vs-drama-…

Lots of people here on X argue with physicists about physics, lawyers about the law, etc!

@JackPrescottX LOL !!!

Inventing Salesforce from first principles:

Need the Pope to release an encyclical against Teams calls

One tomato to rule them all, one tomato to find them, one tomato to bring them all and in the darkness bind them.

I asked every major LLM to describe the other LLMs as if they were classmates at the same university. They did NOT hold back. Here's what ChatGPT, Claude, Gemini, Grok, and Perplexity said about each other: 👇

My intuition about geometry was catastrophically wrong. I never knew that, in high dimensions, spheres effectively disappear. By 100 dimensions, an inscribed sphere occupies a smaller fraction of its cube than a proton occupies of the observable universe. By 500 dimensions, the volume is smaller than what standard floating-point arithmetic can even represent. The equations are real. Here’s the wild horror hidden in higher-dimensional geometry:

Visualization of a TCE (bispecific antibody) binding to CD3 on a T-cell, triggering the release of perforin and granzymes (the dots) to kill the tumor cell. Perforin forms pores in the tumor cell membrane, allowing granzymes to enter the cell & eliminate it:

TOMORROW: The next @centermodhealth newsletter comes out. Subscribe here to make sure you don't miss it: centerformodernhealth.org/about-newslett…

The future is going to be amazing. I have a front-row seat for the AI revolution in software engineering and cybersecurity, and I can't wait to see all the wonders coming out of the AI revolution in literally everything else!

Good on United and the flight attendant on enforcing the rules!

I'm seeing a lot of tech founders on O-1s and H-1Bs ring the alarm bells over this. But they don't need to panic. H-1B and O-1 visas are dual intent visas, i.e. you are allowed to apply for a green card while you are on them. My understanding is that nothing changes for these applicants. Tourist, F-1 (student), and similar visas are temporary for non-immigrant intent. On those, you are not meant to apply for a green card, because that shows immigrant intent. That's why people on F-1 visas applying for green cards, for example, can't leave the country for a year or two while they're waiting -- they might not be re-admitted at the border. The loophole that DHS/USCIS is closing is the one where you can be on a non-immigrant visa but still apply for permanent immigration inside the country, because that immigrant intent is only evaluated at the border, not by USCIS at application-time. (Confusing, right?) On its face, limiting non-immigrant visas to not enabling permanent residency, and instead directing those applicants to the proper immigrant visa route, is not crazy. When you get a non-immigrant visa, you do certify that you don't intend to immigrate. The only technical talent that suffers from this are F-1 students on post-completion OPT, of whom there are many bright achievers in Silicon Valley. But many of those are also good fits for O-1s. As an aside, I obtained my green card on the F-1 -> EB-3 path, which is what's being challenged by this memo. I'm grateful it was straight-forward at the time because I never got an H-1B visa in the lottery (30% chance), but this was also not an ideal process. Three years of uncertainty. In my view, what closing this loophole does is force the conversation on the need for skilled immigration visas -- a better, faster, higher-certainty pathway from a strong university to the right to work that doesn't have you competing in a stacked lottery. The Twitter reaction is alarmist, but I think the policy reality is more benign.

Some time ago mathematicians proposed that the first thing we should share with alien intelligence is this image:

A commercial for the Minas Tirith LEGO set. Whoever directed it knows what it’s like to be a Lord of the Rings fan.

“A blind tasting revolutionised the wine world 50 years ago” economist.com/culture/2026/0…

Trolley problem, California high speed rail edition