New RAND report on an important (and messy) question: When should we actually worry about AI being used to design a pathogen? What’s plausible now vs. near-term vs. later?
(1/12)
I helped convene two expert Delphi panels in AI + Bio to weigh in. Bottom line (near term): Through ~2027, AI is mostly an accelerator for already-skilled actors—not an autonomous designer of novel pathogens. Useful, yes. Independent, no. And significant uncertainty remains about the slope of progress.
(2/12)
Full report: rand.org/pubs/research_…

I want to flag this amazing paper, "Preferences for Prevention: People Assume Expensive Problems Have Expensive Solutions," by @TalbotMAndrews and @ryanbq

onlinelibrary.wiley.com/doi/abs/10.111…

Thread (1/6) @TalbotMAndrews @ryanbq OK, to start, the title really does tell you the conclusion, and it's an important one. The piece it leaves out is the mechanism, and the way they showed that it happens.

Fantastic opening statement from US Undersecretary @Jenkinsbd of @USAmbCD this morning about the need to support the inexpensive but critical work of the #1972BWC at the #BWCRevCon today, and need to see its work expanded. (I unfortunately missed the first statement of the day, by the UK delegation. I understand that they mentioned the Russian invasion of Ukraine, but no point of order was raised.)

Happy to report that nothing surprising or exciting has happened in the last hour an a half of statements- just states reiterating importance of BWC and need for new widely discussed ideas proposed in the 29 non-Russian working papers. meetings.unoda.org/bwc-revcon/bio…

#1972BWC #BWCRevCon ...and I spoke too soon!

The chair just said that Algeria will have the last statement, because 2 countries have asked for a right-of-reply before we end for the day. (I wonder who the second one was?)

#1972BWC #BWCRevCon

Estonia was speaking about Russian disinformation. Russia is now raising a point of order saying that this is not on the current topic, and per rule 16.2 asked the chair to stop the provocations and unrelated opinions of the Estonian delegation. #BWCRevCon #1972BWC The chair now says that they rule that states should be sure to follow the rule, but in order to allow states to complete their statements and tie them into the topic, Estonia may continue and finish making their point.

Estonia says it has already finished.

#BWCRevCon #1972BWC

Very excited to be here for @JHSPH_CHS's Global Forum on Scientific Advances Important to the BWC side event here at the #1972BWC Review Conference in Geneva. @JHSPH_CHS This follows up on their (interrupted) annual BWC side event, last held in 2019 - centerforhealthsecurity.org/our-work/event…

They are recording this event for their podcast, hosted by @ggoronvall, in case anyone is worried that they will miss out.

People interested in reducing biorisk seem to be super excited about 222nm light to kill pathogens. I’m also really excited - but it’s (unfortunately) probably a decade or more away from widespread usage. Let me explain. Before I begin, caveat lector: I’m not an expert in this area, and this is just the outcome of my initial review and outreach to experts. And I’d be thrilled for someone to convince me I’m too pessimistic. But I see two and a half problems.

The recent (inexplicable and irresponsible) decision by the NIH to fund more of Peter Daszak's dangerous bat-virus discovery work brings up the obvious, and easily answered, question; is this safe?

No, it's not, in two different ways.

Thread, 1/6

https://twitter.com/WilliamAEden/status/1576387289336336386

So there are two parts to this question. The first is if the collection can be done safely.

The second, of course, is whether actually gathering the data is useful or safe, but as @KEsvelt has explained, the answer is no, absolutely not.

https://twitter.com/kesvelt/status/1498409798903209996

I've been thinking about how people change the world for the better for quite a while. Turns out it's hard, and the world is complex, but more critically, most people aren't trying. And if they care about the world, and want it to be better, that's a shame. (1/25) What do you want to do? (And what are you actually optimizing for? 80000hours.org/2022/06/know-w… @lxrjl)

Most people don't try to change the world. Some people lack resources to take risks, or don't want to risk failure. This is understandable. But others just never thought about it.

1) We never had evidence that vaccines don't prevent transmission.

2) We now have evidence that they do help prevent transmission.

3) The science community's "caution" about conclusions plus media's scaremongering with "we don't know if..." is going to kill people. Again.

https://twitter.com/KateGrabowski/status/1359323992990101507

We're setting ourselves up for failure.

Don't think that we can change public opinion on a dime. We can't.

And the US isn't the only place in the world. This messaging is killing people in Israel already, where we're already seeing vaccine hesitancy as the key issue.

https://twitter.com/EpiEllie/status/1359352230021910532

The medical and scientific community is concerned about damaging trust in vaccines and in science by taking risks responding to COVID-19.

Instead, they damaged trust by failing to do so, avoiding risk and being slow and conservative to react.

(Thread.) First, it seems scientists are worried that "providing the public with information about uncertainty would increase distrust in science and scientific institutions, as well as cause panic and confusion."

I recently published a paper outlining risks of technological fragility, as they affect the long-term future, titled "The Fragile World Hypothesis" - sciencedirect.com/science/articl…

Go read it! Or, if you don't have the time, here's a thread summarizing the main ideas. As an introduction, technology is fragile - we build very complex system that need ongoing maintenance just to remain functional. This is clearly true of computer systems, but I'll start with a more basic example - bridges.

One critical failure of (mis-applied) empirical / scientific thinking is over-reliance on evidence. Evidence is critical. It's just not enough - we can easily be misled by data.

Here's a a couple dozen tweets on how this happens, and what to do about it. #tweetstorm tl;dr; In general, be wary of over-reliance on evidence, since it is inevitably conditioned on many assumptions or factors, some of which change.
That overreliance on data is especially bad with small n, or in complex domains, or in domains with non-predictive theory.

A short monologue on creativity versus the need for concrete and quantifiable metrics that align with actual goals.
cc: @HelmanDaniel

To start, creativity is indispensible for many types of work, and it needs to be understood to design useful measurements. @HelmanDaniel Creativity qua creativity isn't useful, however - in fact, it's usually just the opposite.

So creativity is important to understand, but NOT to measure or reward. We want to allow creativity when it is useful, not measure a proxy for being creative.

I said last week that Brandon Tuckfield's (@MormonThinker) article on @Quillette about attraction inequality in online dating makes unjustified conclusions based on bad data and worse analysis. Here's the story - quillette.com/2019/03/12/att… @MormonThinker @Quillette First, a disclaimer; Life isn't fair. I'm not arguing about whether some people get disproportionately more attention in online dating than others, or if that's good or bad. I'm just explaining that the article is bad statistics, bad journalist, and bad social science.