1/4 One of the weirdest things I learned researching this @opinion column was that the more language Chat GPT learns, the better it gets at solving linear equations. bloomberg.com/opinion/articl… 2/4 Here's what Harvards @AndrewLBeam said about this. “I think the best way to think about it is that solving systems of linear equations is a special case of being able to reason about a large amount of text data in some sense.”

1/ This @WSJ column implying vaccines fuel bad new variants such as #XBB.1.5 cries out for a re-analysis. I'm not talking about my own competing view but one informed by conversations with the A-team on evolution: @roby_bhatt @LubanLab and @jbloom_lab wsj.com/articles/are-v… 2/ It's correct that the main selective pressure driving evolution of the virus is our immunity, and the only variants to spread now will be ones that can get past it.....

1/ Nearly a year before @DLeonhardt quoted @mtosterholm on the big mystery of Covid waves, I quoted @mgmgomes1 in @bopinion raising the possibility that human heterogeneity was a factor in the unexpected fall of waves: bloomberg.com/opinion/articl… 2/ Her insight might help explain why the US winter wave plummeted in January, long before vaccines were widespread enough to play a role. And why the big wave plummeted in India and now cases are falling in the UK. We should be looking for the sources of heterogeneity.

In my latest episode of Follow the Science, I question today’s ubiquitous outrage about Covid-19 misinformation. How do we know what counts as misinformation with a new disease and so much still being worked out?
open.spotify.com/episode/1gNftq… Physician and science communicator @RogerSeheult and I talk about drugs. For a while, hydroxychloroquine looked promising enough to test in humans. When Trump touted Seheult was censored by Youtube for talking about it.