I'm excited to share that I will be starting my lab at NTU Singapore. It is bittersweet to be moving halfway around the world. On one hand, I'm excited for a new adventure. I have spent my entire adult life chasing this dream, and it's finally happening. On the other hand, I still love America and believe in all of the good that it stands for. My parents moved here to give me the gift of the American Dream and I never thought I would experience what it's like to be an immigrant to a new country myself. I can only say that life has a way of offering blessings in disguise, and I’m grateful for this opportunity in every way.

Amazing work from Ophir Shalem’s lab. I also want to emphasize how valuable it was that they preprinted this work two years ago. It has been a major boon for my lab and many others. This -- more than a solitary Anthropic wet lab -- is what it takes to accelerate scientific discovery: every lab builds on the results of others. Without this preprint, my lab would be months to years behind where we are today. Congratulations, and I’m sorry peer review took so long!

There's an interesting paradox here. The best model for Alzheimer's would be a human brain that we can grow in a dish. But the closer our model recapitulates a real brain, the more we venture into the sci-fi dystopia of disembodied brains in a vat. The only reason we can justify organoid work is that we argue 1) it's just a clump of neurons, it's not a real brain, but 2) wait it's actually good for disease research, which TBD. Once you start xenotransplanting these organoids into apallial mice, then we enter a moral gray area where we we might need to get a second opinion from the Effective Altruists 😉

Biology is like a beautiful tech tree. New knowledge and new cures are unlocked by dozens of dependencies. This is why it's hard even for an all seeing AI to make instant progress in our field. If we want an equivalent to the Millennium Problems in biology, they should be beautiful, universal conjectures. "DNA is the genetic material" was perhaps the ultimate example, proven in parts by Avery et al. (1944) → Hershey and Chase (1952) → Watson and Crick (1953). Everyone should read The Eighth Day of Creation to relive that story. Nothing quite so elegant remains - some truths of Nature can only discovered once. But among my personal favorites: 1. What were the functions of the first polypeptides that were made in an RNA world 2. What genetic innovations enabled the evolution of human intelligence? 3. Do other forms of life exist in our solar system or in our galaxy? If so, do they use DNA or other types of nucleic acids and amino acids? You may notice that these are more questions than conjectures. Indeed, to have a hypothesis means we're already halfway there. In this regard, math and biology may differ.

Waymo is going to be a massive success in Singapore and I can't wait. All they have to do to be better than the local Grab drivers is to not tap the brakes every 3 seconds for no reason whatsoever.

I hope you all realize that 40 years from now, when we're all old and graying and enjoying a life of abundance, there will still be a new generation of AI researchers trying to convince us that GPT-42 is going to make us all go extinct

I know everyone wishes there was some responsible adult who's ready to hit the pause button once some major AI disruption happens. The big boys at the CIA and the pentagon. Bernie Sanders holds an anti-datacenter rally. But this is a government that has spent the last decade doing precisely zero to reign in the harms of social media. Gambling is legal across the land. This is a kick the can down the road system of government, especially when it comes to hard decisions that impact massive business interests. We'll fix social security sooner than we'll hit pause on AI development, mark my words.

Why is it your view that the Senate, which hardly ever does anything, which will suddenly spring into action here? DoW already had their fight with Ant, and everyone in Silicon Valley has radically miscalibrated how the world works if it's a bunch of staffers that hold the fate of the world in their hands

The frontier labs can't work on everything, so they're really only automating themselves out of a job. For the rest of us, AI delivers enormous consumer surplus as we use it to dive into the nitty gritty details of our everyday work.

This paper is honestly just a litany of complaints, assembled by committee. There are a lot of serious people working on progress studies and how to encourage risk taking in science, from @mattsclancy to @salonium to @RuxandraTeslo The reality is that like most postdocs, I was given a large degree of leeway to craft my own project. My PI sugar daddy funded it all. Now that I have my own lab, I've got another 7 years of runway to work on literally anything I want. With a team of scientists under my supervision. All in all, that's 10-15 years to make a mark and do something worthy of the opportunity I've been given. If I don't succeed, well, skill issue. To me, the most inspiring stories of risk taking in science are stories like Brian Kobilka and Shinya Yamanaka. Kobilka gave up HHMI funding to work on crystallizing GPCRs, ran out of money twice, even offered to cut his own salary. Everyone told him it couldn't be done, pharma had poured billions into the effort and failed. His first generation of postdocs working on the project also failed. But then the next group succeeded. For Shinya, the seminal iPSC paper is a two author paper. There's no collaboration, zero interdisciplinary team, and they were by no means a well resourced lab compared to the other mES powerhouses at the time. The honest reality is that the extreme career pressures of the academic track drive postdocs in particular to be risk averse. Every MIT/Harvard postdoc is asking themselves: what's the surest route to a CNS publication so that I can get that dream job? If you take on a risky project, a bold hypothesis, you're well aware that you're betting your career on that paradigm shifting hypothesis being right. And so the careerist postdocs flock to labs that are willing to spend a $x00k on a scRNA-seq atlas, knowing that going hypothesis free is the surest route to publication. It's a shame because the postdoc corps is the lifeblood of scientific research, the age at which young scientists are at the height of their creativity, skill, and determination. If there's one thing to fix, it is this. As academics, we can do better than to coddle young scientists (ECRs I guess we're called), to lament the unsurmountable obstacles we face, the systemic pressures, the crappy administration, alas! It is not so. We must demand better from ourselves. Such that we are worthy of the public funding that supports us.