Binding Sites

Happy Monday

This last two weeks felt both short and long; Memorial Day was right in the middle, but we have had a whole Bio AI panel since the last issue! Time: what even is it?

I will be in NYC for the Jefferies Global Healthcare conference this week, and for Fathom’s first post-Series A board meeting on Friday. Not many words from me this week, I’m afraid, but I hope to have more to share after the (many) scheduled meetings in NYC and the rumination time in the air. Until then, enjoy learning about pandas etc.

Binding Sites

1. What It’s Like to Be a Panda Besides being an essay replete with beautiful prose, this served as a good reminder that evolution can lead to bizarre dead ends just as it can lead to honed, fit-for-purpose traits.
2. Data Center Discontent, Understanding the Opposition, Fixing the Problem Ben Thompson notes an interesting contrast to globalization with the opposition to AI datacenters: the people who believe they will be affected by labor displacement from AI actually have a say this time, and are responding by blocking data center construction. The solution he proposes is to do what we’re best at here in the U S of A and pay people off: to get the data center proposal approved, the operator could agree to pay every person in the town hosting the facility an annual stipend that collectively amounts to a negligible percentage of the total revenue. This works as long as demand for compute remains essentially unlimited, which is showing no signs of changing.
3. Betting on biotech: Prediction markets set sights on clinical trials Well, can’t see any way in which this could go wrong.
4. Do we need more ‘aluminum-standard’ evidence? In an industry that lives and dies by data, it’s important to remember that some situations call for making a decision based on the data you have, not the data you could theoretically get. Failing to act while waiting for yet more data is a decision in itself.