Common tech jobs described as cabals of mesoamerican wizards

Is the cell really a machine?

AI scaling might hit a plateau soon as we run out of data. Large language model skepticism, from the same author.

Organ problems? No problem! An Israeli company wants to grow embryous made out of stem cells, to eventually produce organ replacements.

How much quantum computing skepticism should one harbor?

Scott Alexander reviews What We Owe The Future

Plasma dilution as an anti-aging treatment continues to work in humans

Nick Lane on cancer and aging

How NEPA works

On how the Rockefeller Foundation funded biology

Nat Friedman has a list of AI-enabled companies for people to work on. I am not terribly excited about them (I'm not the only one!). The one that does seem interesting, automating software engineering, is one that maybe he's directly working on? On chatbots I'll say that they won't work for reasons that have little to do with LLMs (As we found out at, people just don't like chatbots if they can have more context on a screen; see also here). I also see strong usecases for conceptual design, but nothing too revolutionary yet. Another idea that I find exciting that's maybe less cool is using LLMs to find vulnerabilities in code.

Now this seems more exciting: How to build GPT3 for science

For the first time, it has been possible to get ribosomes to make ribosomes in vitro

Yann LeCun on the limits of large language models (tacit knowledge!)

Trevor Klee: Biology has analogies, not paradigms

Notes on FROs from Sabrina Singh

Maybe non-nutritive sweeteners (like sucralose) are worse than we thought; though there also seems to be lots of individual variation

A totalizing theory of biomedical progress

Midjourney, DALLE2, and Stable Diffusion compared

Biotech lab space costs

Spatial technologies of the future, by Zack Chiang

ARIA, UK's new innovation agency announces their leadership team