Mel Andrews
@bayesianboy.bsky.social
I’m not like the other Bayesians. I’m different.
Thinks about philosophy of science, AI ethics, machine learning, models, & metascience. postdoc @ Princeton.
Thinks about philosophy of science, AI ethics, machine learning, models, & metascience. postdoc @ Princeton.
Pinned
Mel Andrews
@bayesianboy.bsky.social
· Nov 17
The reanimation of pseudoscience in machine learning and its ethical repercussions
The present perspective outlines how epistemically baseless and ethically pernicious paradigms are recycled back into the scientific literature via ma…
www.sciencedirect.com
A thread of my papers. The first was recently published with Cell Press, coauthored with Andy Smart and Abeba Birhane. It tackles the connections between epistemic deficiencies in deployed ML and ethical consequences. www.sciencedirect.com/science/arti...
Reposted by Mel Andrews
my crew vs the gov't
February 5, 2026 at 3:33 AM
my crew vs the gov't
BREAKING: Trump administration vows to put stop to sexual deviancy and trans agenda in organic macromolecules.
February 5, 2026 at 3:30 AM
BREAKING: Trump administration vows to put stop to sexual deviancy and trans agenda in organic macromolecules.
my cats are useless. my cats are not mousers. they have never managed to kill anything. their most successful hunting feat to date involved drooling on a sleepy bug. today they happened upon a nest of baby mice.
February 1, 2026 at 10:17 PM
my cats are useless. my cats are not mousers. they have never managed to kill anything. their most successful hunting feat to date involved drooling on a sleepy bug. today they happened upon a nest of baby mice.
It intrigues me that “philistine,” (n. uncouth or uncultured), means this in virtue of the ancient association between the Palestinian people and indigeneity. The term was first used by the Israelites to refer to the non-semitic locals (Philistina).
January 31, 2026 at 6:49 PM
It intrigues me that “philistine,” (n. uncouth or uncultured), means this in virtue of the ancient association between the Palestinian people and indigeneity. The term was first used by the Israelites to refer to the non-semitic locals (Philistina).
Reposted by Mel Andrews
Every time I hear some genius bragging about how fast they are learning to code/write/make art by using AI instead of their own brain, I think of Jenko in 21 Jump Street on the drug that makes him think he’s figured out all of chemistry but the whiteboard is just a whole bunch of 4s
January 31, 2026 at 1:27 AM
Every time I hear some genius bragging about how fast they are learning to code/write/make art by using AI instead of their own brain, I think of Jenko in 21 Jump Street on the drug that makes him think he’s figured out all of chemistry but the whiteboard is just a whole bunch of 4s
Reposted by Mel Andrews
Incredible that they published it.
It’s entirely credible and aligns with what actual linguists, neuroscientists, psychologists, computer scientists, etc. have been saying all along.
It’s entirely credible and aligns with what actual linguists, neuroscientists, psychologists, computer scientists, etc. have been saying all along.
A new study from Anthropic finds that gains in coding efficiency when relying on AI assistance did did not meet statistical significance; AI use noticeably degraded programmers’ understanding of what they were doing. Incredible.
The latest from Anthropic: using Anthropic's products makes you worse at your job
January 31, 2026 at 12:34 AM
Incredible that they published it.
It’s entirely credible and aligns with what actual linguists, neuroscientists, psychologists, computer scientists, etc. have been saying all along.
It’s entirely credible and aligns with what actual linguists, neuroscientists, psychologists, computer scientists, etc. have been saying all along.
using chatbots in debugging but only for the placebo effect to make me feel like I’m outsourcing labor and force me to read my own code and realize I’ve double-initialized something:
Kernighan's Law
«Everyone knows that debugging is twice as hard as writing a program in the first place. So if you’re as clever as you can be when you write it, how will you ever debug it?»
-- Brian Kernighan, 1974
«Everyone knows that debugging is twice as hard as writing a program in the first place. So if you’re as clever as you can be when you write it, how will you ever debug it?»
-- Brian Kernighan, 1974
January 31, 2026 at 12:42 AM
using chatbots in debugging but only for the placebo effect to make me feel like I’m outsourcing labor and force me to read my own code and realize I’ve double-initialized something:
Reposted by Mel Andrews
I've got to give Anthropic credit for willing to publish this specific study, yet they're also posting unhinged bullshit like "we don't really know if LLMs have feelings".
A new study from Anthropic finds that gains in coding efficiency when relying on AI assistance did did not meet statistical significance; AI use noticeably degraded programmers’ understanding of what they were doing. Incredible.
The latest from Anthropic: using Anthropic's products makes you worse at your job
January 31, 2026 at 12:14 AM
I've got to give Anthropic credit for willing to publish this specific study, yet they're also posting unhinged bullshit like "we don't really know if LLMs have feelings".
Reposted by Mel Andrews
They did the meme!
January 31, 2026 at 12:13 AM
They did the meme!
A new study from Anthropic finds that gains in coding efficiency when relying on AI assistance did did not meet statistical significance; AI use noticeably degraded programmers’ understanding of what they were doing. Incredible.
The latest from Anthropic: using Anthropic's products makes you worse at your job
How AI assistance impacts the formation of coding skills
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
www.anthropic.com
January 30, 2026 at 11:47 PM
A new study from Anthropic finds that gains in coding efficiency when relying on AI assistance did did not meet statistical significance; AI use noticeably degraded programmers’ understanding of what they were doing. Incredible.
Important update. Google’s YouTube has not only erased thousands of videos of human rights violations in Palestine, it has now blocked livestreaming from Al Jazeera within the country. If you have not already divested from the platform, now is the time to do so.
January 30, 2026 at 10:50 PM
Important update. Google’s YouTube has not only erased thousands of videos of human rights violations in Palestine, it has now blocked livestreaming from Al Jazeera within the country. If you have not already divested from the platform, now is the time to do so.
Reposted by Mel Andrews
Now, I am become drivel, the nonsenser of words
I was put on this earth to be the progenitor of senseless drivel. Look what you have done, you bastards! Now I am forced to rail against the senseless drivel, as entire civilizations throw themselves against the rocks of senseless drivel machines.
January 29, 2026 at 7:45 PM
Now, I am become drivel, the nonsenser of words
well this seems fine and normal
January 29, 2026 at 4:48 AM
well this seems fine and normal
Reposted by Mel Andrews
God, I wish I never became the victor. Now I have all this fucking history to write.
January 28, 2026 at 8:37 PM
God, I wish I never became the victor. Now I have all this fucking history to write.
I was put on this earth to be the progenitor of senseless drivel. Look what you have done, you bastards! Now I am forced to rail against the senseless drivel, as entire civilizations throw themselves against the rocks of senseless drivel machines.
January 28, 2026 at 5:33 AM
I was put on this earth to be the progenitor of senseless drivel. Look what you have done, you bastards! Now I am forced to rail against the senseless drivel, as entire civilizations throw themselves against the rocks of senseless drivel machines.
...if only there were a way to integrate a scholar’s top productivity software with the choking to death on the linguistic vomit of the internet software... hang on! I’ve got just the thing!
So initial experiments with Open AI's vibe-coding science tool Prism are going about as well as expected.
January 28, 2026 at 5:16 AM
...if only there were a way to integrate a scholar’s top productivity software with the choking to death on the linguistic vomit of the internet software... hang on! I’ve got just the thing!
If you, as a scientist, cannot be bothered to engage in the intellectual work of science, please quit your job and leave it to someone with skill and integrity.
“The idea is to put ChatGPT front and center inside software that scientists use to write up their work in much the same way that chatbots are now embedded into popular programming editors.
It’s vibe coding, but for science.”
It’s vibe coding, but for science.”
OpenAI’s latest product lets you vibe code science
Prism is a ChatGPT-powered text editor that automates much of the work involved in writing scientific papers.
www.technologyreview.com
January 28, 2026 at 5:10 AM
If you, as a scientist, cannot be bothered to engage in the intellectual work of science, please quit your job and leave it to someone with skill and integrity.
If you happen to live in America at this historical juncture, the bad news is that you are in danger. The good news is that there are people in your city, in your neighborhood, on your block getting organized to protect each other against that danger. Find them.
January 28, 2026 at 4:02 AM
If you happen to live in America at this historical juncture, the bad news is that you are in danger. The good news is that there are people in your city, in your neighborhood, on your block getting organized to protect each other against that danger. Find them.
Reposted by Mel Andrews
The "ban whistles" stuff is how you know whistles are working incredibly well
January 28, 2026 at 2:10 AM
The "ban whistles" stuff is how you know whistles are working incredibly well
Reposted by Mel Andrews
My "I don't! believe in Phrenology" tee shirt is raising a lot of questions already answered by my tee shirt.
In response to announcement that he will be a new CBS contributor, Andrew Huberman tweeted “I’ve always gone issue-by-issue on health & science, spoken to experts w/a range of takes & presented where I see the center of (data) mass pointed. I’ll do the same with @CBSNews”. In Sept he posted this:
January 28, 2026 at 2:38 AM
My "I don't! believe in Phrenology" tee shirt is raising a lot of questions already answered by my tee shirt.
I confess that I still struggle to believe that there are people out there who cannot tell the difference between genuine photographs and AI-generated images.
January 28, 2026 at 12:19 AM
I confess that I still struggle to believe that there are people out there who cannot tell the difference between genuine photographs and AI-generated images.
Good for them, I guess, but what is da tat centers? @theguardian.com
January 26, 2026 at 7:11 PM
Good for them, I guess, but what is da tat centers? @theguardian.com
Chatbots have been an incredible loss for scientists and an incredible win for the people who send scientists 5,000 word emails about anti-gravity engines and CIA microchips.
January 24, 2026 at 10:31 PM
Chatbots have been an incredible loss for scientists and an incredible win for the people who send scientists 5,000 word emails about anti-gravity engines and CIA microchips.
I often hear, in my line of work, that large language models have surpassed the capabilities of “mere stochastic parrots.” Yet few are willing to attribute to them anything like “understanding.” What, in your view, is the best recent appraisal of the limitations of LLMs?
January 22, 2026 at 7:04 PM
I often hear, in my line of work, that large language models have surpassed the capabilities of “mere stochastic parrots.” Yet few are willing to attribute to them anything like “understanding.” What, in your view, is the best recent appraisal of the limitations of LLMs?
attn fellow faculty: when students hand in AI-generated papers, any of you try eating them?
January 21, 2026 at 12:34 AM
attn fellow faculty: when students hand in AI-generated papers, any of you try eating them?