Alex Crits-Christoph
@acritschristoph.bsky.social
Computational microbiologist
I like to post about: microbial genomics, microbial ecology, evolution, micro+plant biotechnology, climate, symbiosis, virology, ag, sci publishing and policy
I like to post about: microbial genomics, microbial ecology, evolution, micro+plant biotechnology, climate, symbiosis, virology, ag, sci publishing and policy
Longer term this technology must be open, decentralized, and decorporatized. I think driverless cars as a public good ot driverless car independent co-ops are good future models, and possible as the tech gets cheaper and universal.
But the tech itself is good even if it is currently corporate!
But the tech itself is good even if it is currently corporate!
December 2, 2025 at 5:05 PM
Longer term this technology must be open, decentralized, and decorporatized. I think driverless cars as a public good ot driverless car independent co-ops are good future models, and possible as the tech gets cheaper and universal.
But the tech itself is good even if it is currently corporate!
But the tech itself is good even if it is currently corporate!
Reposted by Alex Crits-Christoph
The worst nightmare I have is not simply that it happens, but that everyone sees it happen, and the question is not simply whether it is unfair (we all agree), but whether I handled it right: should I have said less? Chosen words more carefully? Been more paranoid about who I was speaking to?
December 1, 2025 at 9:59 PM
The worst nightmare I have is not simply that it happens, but that everyone sees it happen, and the question is not simply whether it is unfair (we all agree), but whether I handled it right: should I have said less? Chosen words more carefully? Been more paranoid about who I was speaking to?
and "but democracy is wrong sometimes" actually isn't a good argument at all, because individuals are wrong a lot more of the time
December 1, 2025 at 10:09 PM
and "but democracy is wrong sometimes" actually isn't a good argument at all, because individuals are wrong a lot more of the time
Totally disagree! Science is and should be democratic. What data we collect and use, how we analyze that data, what we infer from it, and what studies are deemed methodologically sound are in fact opinions.
The only way to consistently get at that accurately is consensus opinions of experts.
The only way to consistently get at that accurately is consensus opinions of experts.
December 1, 2025 at 10:09 PM
Totally disagree! Science is and should be democratic. What data we collect and use, how we analyze that data, what we infer from it, and what studies are deemed methodologically sound are in fact opinions.
The only way to consistently get at that accurately is consensus opinions of experts.
The only way to consistently get at that accurately is consensus opinions of experts.
Reposted by Alex Crits-Christoph
And as a blast from that terrible past, the motivations are sadly often backed by interests that reward this behavior with power and money. I still think we have not done enough to train scientists to counter the doubt machine... www.nature.com/articles/d41...
Scientists: don’t feed the doubt machine
From climate to COVID, naivety about how science is hijacked promotes more of the same.
www.nature.com
November 30, 2025 at 5:50 PM
And as a blast from that terrible past, the motivations are sadly often backed by interests that reward this behavior with power and money. I still think we have not done enough to train scientists to counter the doubt machine... www.nature.com/articles/d41...
So for those who are skeptical of the use of e.g. JIF as a quality metric, the challenge will be to work towards *alternative and competitive* metrics that quickly convey a less biased and more accurate correlate of article quality. Without that, AI will force folks to rely on the metrics that exist
November 29, 2025 at 11:04 PM
So for those who are skeptical of the use of e.g. JIF as a quality metric, the challenge will be to work towards *alternative and competitive* metrics that quickly convey a less biased and more accurate correlate of article quality. Without that, AI will force folks to rely on the metrics that exist
Ahh!
(1) yes in my opinion literate+initially-convincing writing was "rate-limiting" for a lot of bad science
(2) I think high JIF journals will successfully (relatively) filter AI-slop, while preprint servers and low-JIF journals will be less successful - that's been the trend so far.
(1) yes in my opinion literate+initially-convincing writing was "rate-limiting" for a lot of bad science
(2) I think high JIF journals will successfully (relatively) filter AI-slop, while preprint servers and low-JIF journals will be less successful - that's been the trend so far.
November 29, 2025 at 11:04 PM
Ahh!
(1) yes in my opinion literate+initially-convincing writing was "rate-limiting" for a lot of bad science
(2) I think high JIF journals will successfully (relatively) filter AI-slop, while preprint servers and low-JIF journals will be less successful - that's been the trend so far.
(1) yes in my opinion literate+initially-convincing writing was "rate-limiting" for a lot of bad science
(2) I think high JIF journals will successfully (relatively) filter AI-slop, while preprint servers and low-JIF journals will be less successful - that's been the trend so far.
Hmm must admit, I can't say I follow your point(s) here!
Our point here is that with more noise out there, people will default to reading and trusting papers in high JIF journals and from labs and institutions they know, unless we build new systems of communicating the quality of research
Our point here is that with more noise out there, people will default to reading and trusting papers in high JIF journals and from labs and institutions they know, unless we build new systems of communicating the quality of research
November 29, 2025 at 2:30 PM
Hmm must admit, I can't say I follow your point(s) here!
Our point here is that with more noise out there, people will default to reading and trusting papers in high JIF journals and from labs and institutions they know, unless we build new systems of communicating the quality of research
Our point here is that with more noise out there, people will default to reading and trusting papers in high JIF journals and from labs and institutions they know, unless we build new systems of communicating the quality of research
So the issue is that AI has made it easier for people to create works that appear convicing but aren't. This means that there's more noise, and it takes longer to dissect. So people will have to lean on traditional signals: who they know, author institutions,l, and journal name- now more than ever.
November 28, 2025 at 9:28 PM
So the issue is that AI has made it easier for people to create works that appear convicing but aren't. This means that there's more noise, and it takes longer to dissect. So people will have to lean on traditional signals: who they know, author institutions,l, and journal name- now more than ever.