Joe @twitskeptic@qoto.org

@ct_bergstrom Not relevant to your main point, but serially participating in large consortium projects (human genome, 1000 genomes, T2T) is a great way to inflate your h-index.

15 very highly cited papers over 20 years that significantly change your specialty? Not so much.

@ct_bergstrom @seldoncrisis

He repudiated the way he framed his views but not the actual views.

First he tap danced around the main and repugnant idea that Black people are inherently less intelligent than white people ("it's not my area of expertise and and I'm not interested") and then attempted to distract by wandering off into a topic (eugenics) which wasn't referenced in the original offensive context.

But he gave to a lot of charities, so I guess that makes up for it.

@ct_bergstrom https://twitter.com/kareem_carr/status/1613325029848092672

If you've ever spent any time in Ann Arbor (or even if you haven't) this is a fun, if marginally condescending, article if you're a certain age. https://www.nytimes.com/2023/01/12/style/ann-arbor-geezer-happy-hour.html

@ct_bergstrom Damn, not good news.

@ct_bergstrom Looks like a publicity stunt that's right out of the Zuckerberg playbook of violating user trust, announcing a new "feature" based on the violation, temporarily backing out due to bad publicity, but then leveraging the attention to raise money or get even more publicity for his company.

@ct_bergstrom @JamesGleick @emilymbender Summary of the RL from the paper.

@ct_bergstrom @JamesGleick @emilymbender One subtlety that's lost in ChapGPT discussions is that it really is a LLM that's *then* fine tuned with human knowledge through reinforcement learning. This paper from OpenAI explains the process. https://arxiv.org/pdf/2203.02155.pdf

So, yes, ChapGPT does generate random bullshit but not quite as much as a strict LLM would be expected to generate. Not sure if @emilymbender has addressed this.

@ct_bergstrom He lost me at the specious analogy between biological and artificial neural networks. Just as a starter, I don't think that there's back propagation in human brains.

@epiellie FWIW, Trevor Bedford's data shows pretty consistent XBB 1.5 growth rates in several countries. https://github.com/blab/rt-from-frequency-dynamics/tree/master/results/xbb15-ba2320

@ct_bergstrom There's also a requirement that less risky approaches for answering the given scientific question should be considered.

@ct_bergstrom FWIW, the HHS departmental review requires representatives from a number of different specialties including public health, biosafety and biosecurity. Seems a lot more rigorous than your standard grant review. Is it good enough? No idea.

@ct_bergstrom
RE GOF: I admit that (as a non-expert) I'm uncomfortable with some GOF research. But I can also see how the risk-reward equation may occasionally tilt towards the reward side under certain circumstances.

So two questions:
1) Are you aware of any troubling new research funded under the NIH guidelines?
2) What would you change in the HHS guidelines, which does require department level review? https://www.phe.gov/s3/dualuse/Documents/p3co.pdf

@ct_bergstrom All that's missing is the NFT.

@ct_bergstrom We started this a few months ago per your advice and it's really rewarding! Initially it was just a pair of crows but now we have about 5 or so crows who patiently line up for their daily peanut feast.

More recently they've started following us on our walks, which seems to attract other crows. The word is out! Only downside is I feel guilty when I don't have any peanuts with me.

@ct_bergstrom This seems like yet another example of regulatory failure. The FAA/DOT could issue minimum customer service regulations that would prevent this kind of bad behavior (like the tarmac wait rules), but, unfortunately, regulatory capture make this unlikely.

@ct_bergstrom I guess she's talking about Ebola viremia, but I vaguely recall that there's an association between covid mortality and covid viremia. And, of course, covid is much more transmissible than Ebola, which requires contact with precious bodily fluids.