Alexander Schmidt-Lebuhn
@anschmidtlebuhn.bsky.social
Botanist, taxonomist, phylogeneticist.
Pinned
Phylogenomics and genetic admixture of the invasive fireweed complex (Senecio inaequidens–Senecio madagascariensis) at a global scale - Biological Invasions
The southern African Senecio inaequidens–S. madagascariensis complex, (‘fireweed complex’), contains several species that have established as weeds outside of their native ranges. Senecio madagascarie...
doi.org
We have expanded our target capture dataset of the Senecio inaequidens - S. madagascariensis complex to a global scale. More clarity on the areas of origin of invasive madagascariensis, but also raising new questions.
doi.org/10.1007/s105...
@biolinvasions.bsky.social
#asteraceae #compositae #weeds
doi.org/10.1007/s105...
@biolinvasions.bsky.social
#asteraceae #compositae #weeds
Reposted by Alexander Schmidt-Lebuhn
If this is what they do to white American citizens in front of everyone, imagine what happens to non-white people behind closed doors.
January 24, 2026 at 7:34 PM
If this is what they do to white American citizens in front of everyone, imagine what happens to non-white people behind closed doors.
This is so strange. The author thinks that it is normal and good behaviour for OpenAI to make available a plagarism bot and for himself to use the plagiarism bot for proposal writing, but that it is unexpected and terrible if clicking the don't store my data button leads to data being deleted.
January 23, 2026 at 1:09 PM
This is so strange. The author thinks that it is normal and good behaviour for OpenAI to make available a plagarism bot and for himself to use the plagiarism bot for proposal writing, but that it is unexpected and terrible if clicking the don't store my data button leads to data being deleted.
So, doing it alone is serious work, and AI Employees are toys that appeal to small children. That is what this advertisement tries to say, right?
yeah, just out of interest, how many people choose the pen on the right for real work or art? See a lot of them in professional workplaces, do you?
January 22, 2026 at 10:35 PM
So, doing it alone is serious work, and AI Employees are toys that appeal to small children. That is what this advertisement tries to say, right?
Reposted by Alexander Schmidt-Lebuhn
Our paper on the mysterious Devonian organism Prototaxites has now finally been published! See the paper here (www.science.org/doi/10.1126/...) and our explainer thread below!
Prototaxites reconstruction by Matt Humpage
Prototaxites reconstruction by Matt Humpage
January 21, 2026 at 7:25 PM
Our paper on the mysterious Devonian organism Prototaxites has now finally been published! See the paper here (www.science.org/doi/10.1126/...) and our explainer thread below!
Prototaxites reconstruction by Matt Humpage
Prototaxites reconstruction by Matt Humpage
I remember seeing The Matrix in cinema with a friend, and we were both laughing through most of it because we could just not take it seriously. The slow-mo trying so hard to be cool. The perpetuum mobile. People dying when you pull the metaverse plug, as if they had a soul that got disconnected.
For the kids of today, The Matrix is probably foundational to the point of being dusty. This is not a complaint -- The Matrix fuckin' BANGS -- just a reminder that time passes. It's not that kids today can't/won't see SF films before then, just that their eye on them will be very different.
Teaching a Science Fiction CW class today, and mentioned Kubrick's 2001 to illustrate a point. Out of 18, only 1 had seen it (and he wasn't super clear about it). To these—bright, creative, engaged!—20yos this was just some antique movie.
January 21, 2026 at 9:38 PM
I remember seeing The Matrix in cinema with a friend, and we were both laughing through most of it because we could just not take it seriously. The slow-mo trying so hard to be cool. The perpetuum mobile. People dying when you pull the metaverse plug, as if they had a soul that got disconnected.
The equivalent in science is first having abstracts to summarise the content of research papers, journals then inventing 'highlights' to summarise them again, and now AI summarising them in a third and less reliable way. Has everybody gone insane?
Gmail is now asking if I’d like to have two sentence emails summarized
Holy fuck I do not want an ai summary of my email thread at the top of the page by default. 1) I know how to read, and 2) I would rather spend my time READING THE ACTUAL FUCKING EMAIL
January 21, 2026 at 8:56 PM
The equivalent in science is first having abstracts to summarise the content of research papers, journals then inventing 'highlights' to summarise them again, and now AI summarising them in a third and less reliable way. Has everybody gone insane?
Interesting patterns here, but I see a few potential caveats. First, this all seems to be based on species numbers, not local biomass. Second, the cleaned dataset is 2815 taxa, a small fraction of angiosperm diversity, and IIUC, only fleshy-fruited taxa even when analysing flower colour?
🎨 From the upcoming #AJB Special Issue: “Paradigm Shifts in #Flower Color." 🎨🌼
Does the abiotic environment influence the distribution of flower and fruit colors?
New research by Agnes Dellinger, Leah Meier, @iochromaland.bsky.social & Miranda Sinnott-Armstrong
doi.org/10.1002/ajb2... #botany
Does the abiotic environment influence the distribution of flower and fruit colors?
New research by Agnes Dellinger, Leah Meier, @iochromaland.bsky.social & Miranda Sinnott-Armstrong
doi.org/10.1002/ajb2... #botany
January 20, 2026 at 9:53 PM
Interesting patterns here, but I see a few potential caveats. First, this all seems to be based on species numbers, not local biomass. Second, the cleaned dataset is 2815 taxa, a small fraction of angiosperm diversity, and IIUC, only fleshy-fruited taxa even when analysing flower colour?
The post is the same as the title of the linked article, which is the same as its subtitle:
The post is the same as the title of the linked article, which is the same as its subtitle:
The post is the same as the title of the linked article, which is the same as its subtitle
The post is the same as the title of the linked article, which is the same as its subtitle:
The post is the same as the title of the linked article, which is the same as its subtitle
Scientists mimicking the Big Bang accidentally turn lead into gold
Scientists mimicking the Big Bang accidentally turn lead into gold
Scientists mimicking the Big Bang accidentally turn lead into gold
www.independent.co.uk
January 20, 2026 at 8:47 PM
The post is the same as the title of the linked article, which is the same as its subtitle:
The post is the same as the title of the linked article, which is the same as its subtitle:
The post is the same as the title of the linked article, which is the same as its subtitle
The post is the same as the title of the linked article, which is the same as its subtitle:
The post is the same as the title of the linked article, which is the same as its subtitle
There is a discourse cycle right now from vibe coders explaining loudly that all those who claim LLMs don't work are deluded. I know very few people who claim that LLMs are useless in coding. But the vibe coders are carefully side-stepping the real issues: LLMs are built on massive IP theft; ...
January 19, 2026 at 9:47 PM
There is a discourse cycle right now from vibe coders explaining loudly that all those who claim LLMs don't work are deluded. I know very few people who claim that LLMs are useless in coding. But the vibe coders are carefully side-stepping the real issues: LLMs are built on massive IP theft; ...
This medium post is like reading the diary of somebody being driven mad by some kind of Lovecraftian horror. I did not expect them to say that their method prioritises "did it get done" over "does it run" non-ironically, as the benefit of the new approach, as opposed to an indictment of AI-coding.
> “It will be like kubernetes, but for agents,” I said.
This is quite possibly the most cursed sentence I've read all day, and I've been reading up on the Greenland crisis so, y'know, that's an honest-to-god accomplishment.
From steve-yegge.medium.com/welcome-to-g...
This is quite possibly the most cursed sentence I've read all day, and I've been reading up on the Greenland crisis so, y'know, that's an honest-to-god accomplishment.
From steve-yegge.medium.com/welcome-to-g...
Welcome to Gas Town
Happy New Year, and Welcome to Gas Town!
steve-yegge.medium.com
January 19, 2026 at 12:30 AM
This medium post is like reading the diary of somebody being driven mad by some kind of Lovecraftian horror. I did not expect them to say that their method prioritises "did it get done" over "does it run" non-ironically, as the benefit of the new approach, as opposed to an indictment of AI-coding.
Always find it fascinating how megafauna survived several cycles of warm ages and ice ages over two million years and then all rapidly went extinct everywhere from "rapid environmental change" when modern humans with stuff like spears or stone axes were around. That is such an enormous coincidence!
Woolly Rhino was not experiencing population decline prior to its rapid extinction ~14,000 years ago, based on DNA recovered from mummified wolf stomach in Siberian permafrost. Instead, extinction likely due to rapid environmental change.
🧪🌎🧬🖥️ #EvoSky #Genomics
pmc.ncbi.nlm.nih.gov/articles/PMC...
🧪🌎🧬🖥️ #EvoSky #Genomics
pmc.ncbi.nlm.nih.gov/articles/PMC...
Genome Shows no Recent Inbreeding in Near-Extinction Woolly Rhinoceros Sample Found in Ancient Wolf's Stomach
Using temporarily spaced high-coverage ancient genomes, we can assess population decline prior to extinction. However, finding suitable ancient remains for recovering this type of data is challenging. Here, we sequenced a high-coverage genome from ...
pmc.ncbi.nlm.nih.gov
January 18, 2026 at 12:39 AM
Always find it fascinating how megafauna survived several cycles of warm ages and ice ages over two million years and then all rapidly went extinct everywhere from "rapid environmental change" when modern humans with stuff like spears or stone axes were around. That is such an enormous coincidence!
Reposted by Alexander Schmidt-Lebuhn
Using an LLM for research is equivalent to using a hammer to exchange a light bulb.
It is a problem of our times, however, that research is now used to mean everything from a scientist running a double-blind experiment to somebody typing "invertmecin cure cancer?" into the Youtube search bar.
It is a problem of our times, however, that research is now used to mean everything from a scientist running a double-blind experiment to somebody typing "invertmecin cure cancer?" into the Youtube search bar.
RT if you've never had a tab of ChatGPT open 'just to do quick research'
January 17, 2026 at 10:30 PM
Using an LLM for research is equivalent to using a hammer to exchange a light bulb.
It is a problem of our times, however, that research is now used to mean everything from a scientist running a double-blind experiment to somebody typing "invertmecin cure cancer?" into the Youtube search bar.
It is a problem of our times, however, that research is now used to mean everything from a scientist running a double-blind experiment to somebody typing "invertmecin cure cancer?" into the Youtube search bar.
Reposted by Alexander Schmidt-Lebuhn
If we buy the theory that AI will keep improving, we also have to buy that a large part of the software that produces money today will be worth nothing. This is already happening for apps, and also for this community of "micro-saas" you mentioned. Why would I pay for something I can make for free.
January 15, 2026 at 7:58 PM
If we buy the theory that AI will keep improving, we also have to buy that a large part of the software that produces money today will be worth nothing. This is already happening for apps, and also for this community of "micro-saas" you mentioned. Why would I pay for something I can make for free.
Reposted by Alexander Schmidt-Lebuhn
This is what I keep banging on to the AI hustlers about. WHERE IS YOUR MOAT. Like if you can make a SAAS in a weekend do you not think everyone else will have the same idea
January 15, 2026 at 6:14 PM
This is what I keep banging on to the AI hustlers about. WHERE IS YOUR MOAT. Like if you can make a SAAS in a weekend do you not think everyone else will have the same idea
Reposted by Alexander Schmidt-Lebuhn
The AI evangelists have gotten excited about the revolutionary potential of Claude Code.
Even if they're right about the capabilities, they really haven't thought through what comes next.
open.substack.com/pub/davekarp...
Even if they're right about the capabilities, they really haven't thought through what comes next.
open.substack.com/pub/davekarp...
What comes next, if Claude Code is as good as people say.
We know how this turns out. First comes the novelty, then comes the corrosion.
open.substack.com
January 15, 2026 at 6:03 PM
The AI evangelists have gotten excited about the revolutionary potential of Claude Code.
Even if they're right about the capabilities, they really haven't thought through what comes next.
open.substack.com/pub/davekarp...
Even if they're right about the capabilities, they really haven't thought through what comes next.
open.substack.com/pub/davekarp...
The way modern authoritarian states work is not by canceling elections but by making it effectively impossible for the opposition to win. Cancellation isn't the risk; stuff like candidates being prosecuted or ICE harassing non-white voters at polling stations seems more plausible.
the question to ask about this is, okay, he wants to cancel the midterms. how does he get the VA state board of elections to cancel the midterms? how does he get the georgia board of elections to do it? how does he convince republican house members to quit their jobs and give up their paychecks?
Trump says a lot of deranged shit, but, per this Reuters article — and the threats of invoking the Insurrection Act in Minnesota this morning — he is very clearly exploring how to cancel the midterms.
www.reuters.com/world/us/fiv...
www.reuters.com/world/us/fiv...
January 15, 2026 at 5:42 PM
The way modern authoritarian states work is not by canceling elections but by making it effectively impossible for the opposition to win. Cancellation isn't the risk; stuff like candidates being prosecuted or ICE harassing non-white voters at polling stations seems more plausible.
I remain mystified why meeting software like Zoom doesn't automatically equalise the volume across different meeting participants. Maybe that is my ignorance as a non-programmer showing, but that seems both eminently doable and an obvious improvement to the user experience?
January 14, 2026 at 8:02 PM
I remain mystified why meeting software like Zoom doesn't automatically equalise the volume across different meeting participants. Maybe that is my ignorance as a non-programmer showing, but that seems both eminently doable and an obvious improvement to the user experience?
Reposted by Alexander Schmidt-Lebuhn
Most popular stated moments for when people realized Elon was a fraud:
1. The “pedo guy” incident
2. Anything Mars related
3. When he started to claim he founded Tesla
1. The “pedo guy” incident
2. Anything Mars related
3. When he started to claim he founded Tesla
January 14, 2026 at 7:29 PM
Most popular stated moments for when people realized Elon was a fraud:
1. The “pedo guy” incident
2. Anything Mars related
3. When he started to claim he founded Tesla
1. The “pedo guy” incident
2. Anything Mars related
3. When he started to claim he founded Tesla
Ozothamnus stirlingii with bull ant sitting on the capitulescence. Photographed two weeks ago in Namadgi National Park, Australian Capital Territory.
#asteraceae #compositae
#asteraceae #compositae
January 14, 2026 at 5:42 PM
Ozothamnus stirlingii with bull ant sitting on the capitulescence. Photographed two weeks ago in Namadgi National Park, Australian Capital Territory.
#asteraceae #compositae
#asteraceae #compositae
Many interesting aspects about this piece, from a researcher submitting a manuscript to a journal without ever checking its publication fee (srsly?) to Latin grammar fail. But the underlying problem is real:
www.the-scientist.com/rising-publi...
www.the-scientist.com/rising-publi...
Rising Publication Costs Strain Researchers
Open access publishing has led to researchers paying thousands of dollars to publish their work, limiting funds for research and leaving scientists with hard choices.
www.the-scientist.com
January 13, 2026 at 11:53 PM
Many interesting aspects about this piece, from a researcher submitting a manuscript to a journal without ever checking its publication fee (srsly?) to Latin grammar fail. But the underlying problem is real:
www.the-scientist.com/rising-publi...
www.the-scientist.com/rising-publi...
Assuming for a moment that there are infinity resources on the planet for continuing to grow population to infinity, have we considered free child care, generous parental leave, and regulations that ensure parents' careers do not suffer when they take parental leave? You know, progressive policies.
The US and other countries face aging, shrinking populations. Conservatives have shaped debate over the issue. Some liberals say it's time for progressives to weigh in. n.pr/3NH3yso
As birthrates tumble, some progressives say the left needs to offer ideas and solutions
The US and other countries face aging, shrinking populations. Conservatives have shaped debate over the issue. Some liberals say it's time for progressives to weigh in.
n.pr
January 13, 2026 at 9:40 PM
Assuming for a moment that there are infinity resources on the planet for continuing to grow population to infinity, have we considered free child care, generous parental leave, and regulations that ensure parents' careers do not suffer when they take parental leave? You know, progressive policies.
My own take here is that it is impossible to justify every belief we hold from first principles. The real problem is that most people (1) do not have a good sense of what is plausible and (2) prioritise protecting ego or groupthink over aligning their beliefs with potentially unpleasant reality.
You don't know anything unless you know HOW you know it: that your knowledge is based in actual evidence. Most Americans are walking around with heads full of shit they think they know that's just 100% deepfried bullshit from Hollywood or your cousin's mom's friend's story or whatever.
January 13, 2026 at 9:33 PM
My own take here is that it is impossible to justify every belief we hold from first principles. The real problem is that most people (1) do not have a good sense of what is plausible and (2) prioritise protecting ego or groupthink over aligning their beliefs with potentially unpleasant reality.
Reposted by Alexander Schmidt-Lebuhn
just imagining this "tools make discovery" frame applied to historic scientific achievements
"Bread discovers miracle mold"
"Kite finds lightning is actually electricity"
"Bread discovers miracle mold"
"Kite finds lightning is actually electricity"
Scientists have used artificial intelligence to create an enzyme that can eat one of the toughest plastics on Earth. The enzyme breaks polyurethane down into reusable chemicals in just 12 hours at 50°C, turning it back into raw materials. Truly circular recycling. buff.ly/oUxRjjl #ShareGoodNewsToo
Neural network finds an enzyme that can break down polyurethane
Given a dozen hours, the enzyme can turn a foam pad into reusable chemicals.
buff.ly
January 13, 2026 at 4:03 PM
just imagining this "tools make discovery" frame applied to historic scientific achievements
"Bread discovers miracle mold"
"Kite finds lightning is actually electricity"
"Bread discovers miracle mold"
"Kite finds lightning is actually electricity"
I thought for some time now it might be good to collect my experiences with genAI in one thread, so why not now. Major caveat: I am not willing to pay a pro subscription to the plagiarism machine, so people can always say I should have used Claude Poem v16 or whatever and it would have worked.
There is a class of AI critics on this website, typically academics, that speak about the tools like it is early 2024.
I'm not asking you to start liking AI. But you sound ridiculous because you're describing things that no one experiences anymore. Get updated.
I'm not asking you to start liking AI. But you sound ridiculous because you're describing things that no one experiences anymore. Get updated.
January 12, 2026 at 3:51 AM
I thought for some time now it might be good to collect my experiences with genAI in one thread, so why not now. Major caveat: I am not willing to pay a pro subscription to the plagiarism machine, so people can always say I should have used Claude Poem v16 or whatever and it would have worked.