Opens profile photo
Follow
Chris Said
@Chris_Said
Data Scientist at Stitch Fix. Formerly: Opendoor, Twitter, Facebook, neuroscience.
New Yorkchris-said.ioJoined August 2008

Chris Said’s Tweets

On the absurd GOP rush to abandon the successful legacy of Operation Warp Speed, a project that validated early conservative arguments about drug development.
Quote Tweet
The debate over vaccines may be purely political for Trump and DeSantis, says @mattyglesias, but it has long-term implications for US science policy trib.al/5MyLnek
26
353
Love to see it: Republicans and Democrats agreeing that we need to reform environmental procedure laws (NEPA) to build transmission lines faster and mine more critical minerals like lithium.
Image
Image
Image
8
224
Kelsey is completely right about this.
Quote Tweet
I think a lot of people look into the case that AI will definitely kill us, find it's not airtight, and end up believing something like"AI is 10% likely to kill us" but not treating that like it's itself an enormously big deal that would if true need to be society's top priority.
Show this thread
2
Twitter is doing that thing again where it keeps reverting me back to the algorithmic timeline. I wish last year's Elon would talk to this year's Elon!
Quote Tweet
Chronological tweets seem much better than what “the algorithm” suggests
Show this thread
1
4
No, yeah, they're rich. They're just confused because so much of their income is spent on convenience or positional goods.
Quote Tweet
“If you think of a middle-aged professional couple .. each making ~$200K a year, filing a joint tax return, .. paying through the nose for rent or maintenance or a mortgage, you’re probably not going to describe their lifestyle as ‘rich.’” Uh yes I am. nytimes.com/2023/01/16/opi
23
290
Congress is unwilling to fund pandemic preparedness directly so it’s kind of nice that the Defense Department just decided to take this up on its own.
Quote Tweet
Can we take future pandemic threats off the table? @josh_wingrove and I join The Big Take @podcasts to discuss how the US is preparing (or not) for the next pandemic. Listen up!🎧podcasts.apple.com/us/podcast/how
Show this thread
3
Trying not to get in culture war stuff, but this is creepy and bad.
Quote Tweet
"UC Berkeley’s rubric for evaluating DEI contributions, which is used by universities around the country, dictates a low score for a candidate who professes a desire to 'treat everyone the same.'" thefp.com/p/how-dei-is-s
11
TIL that while white noise draws its name from white light, brown noise gets its name from Robert Brown, the guy who discovered Brownian motion.
1
12
Regardless of whether rapid tests led to a moderate drop in infections vs a moderate increase in freedom, they were obviously useful and it was unconscionable that the FDA resisted them for so long.
3
Show this thread
My guess is that in America, most of the benefit of rapid testing wasn't a reduction in infections, but rather the freedom to do more stuff.
Quote Tweet
In technical terms, rapid tests expand the Pareto frontier of freedom and safety. Hopefully they'll give us more freedom and more safety. But if they only give us more freedom, that's good too!
Show this thread
Image
Image
1
Show this thread
My guess is that the true reduction in infections in Liverpool was lower than 43%. In America, which didn't have such well-organized testing programs, the benefits were probably even smaller. But any drop >5% is an enormous win.
1
Show this thread
But while it’s clear that rapid testing wouldn’t eliminate transmission, the more interesting question is whether it had a moderate effect (e.g. 30% reduction) or no effect (0% reduction).
2
Show this thread
On the website, we briefly made some claims that rapid testing could “end the pandemic”. I felt those were obviously aspirational at the time, as they relied on implausible assumptions of daily testing. I assumed this was common knowledge.
1
Show this thread
First Belgium, now France. Mandated CO2 monitoring in indoor spaces. What gets measured gets fixed.
Quote Tweet
And its done. New legislation passed in 🇫🇷 that foresees mandatory CO2 monitoring of indoor spaces. Starting with schools, daycare, restaurants. Bravo 👏 twitter.com/scientistandre
Show this thread
29
Extraordinary that there is now a thing called Protein Language Models.
Quote Tweet
You can use a language model to achieve state-of-the art drug-target interaction predictions. 1. Represent the drug molecule as a sequence of characters 2. Represent the protein as a sequence of amino acids 3. Embed in shared space 4. Dot product biorxiv.org/content/10.110
Image
7
Re-upping this. And obviously I’m especially interested in input from medical professionals!
Quote Tweet
Question for an upcoming blog post: What is an example of a very basic and impactful medical question that could be easily answered with an RCT, but no such RCT has been done?
Show this thread
Zany idea this inspired: bounties for groups that run trials that answer these questions that would be enormously important to payers (whether private insurance or gov). Need to incentivize private production of public goods!
Quote Tweet
Should you go to the doctor for that problem? We really don't know, and we as a profession do little to find out. New Years resolution?? open.substack.com/pub/sensibleme
2
4
Show this thread