What's great about this is that one of them will absolutely nail it, and you're impressed, and then it refreshes after a minute and the same model will be laughably bad, and this is 100% my professional experience with LLMs.

It would be nice if bad things were regulated, and if regulations against bad things were enforced, but apparently "bad things not happening" is not something most voters find that important these days.

The Youtube ad selection algorithm in its wisdom decides to serve me, a cyclist Luddite researcher, SUVs, AI-generated vampire fiction, crypto trading apps, and payment systems for small businesses. I am half-convinced that advertising is a fraud committed against advertisers.

I hope, but am not optimistic, that people understand the difference between (a) I could not do my job without this database. and (b) I could not do my job without the hallucinating, plagiarising, text and image generator.

If I ever believed this, you couldn't drag this admission from me

Look at that, the genus I did my PhD about also occurs in Chile! dx.doi.org/10.4067/s071... #minthostachys

What gets me about alt text is that unless the image contains text or is a visual gag, it generally shouldn't even be necessary, because the post itself should be informative. Like, if you post an image of a bug, building, machine, or person, just put its name into the actual post under it.

Okay, but the problem is that if the "AI" here is an LLM, you cannot trust any of its outputs without double-checking the source yourself. I hope it is just a non-probabilistic search function that they call "AI" to sound fancy instead of actually a probabilistic word-guessing model.

Every ad now

Very excited to have a new publication out in the most beautiful journal in the world, Capitulum! #Compositae #Hecastocleis #Asteraceae #synantherology www.compositae.org/capitulum_04...

This makes sense: genAI, being built to generate stuff, replaces jobs in art and writing by generating slop art and slop writing. Many of us find its outputs repulsive, but companies don't seem to care.

Need to bookmark this to show people what good reporting looks like. Explaining the sources of information. Explaining uncertainty and how the information cold be wrong. Explaining what it means. Clear and concise.

It might be true that the world needs to remove ten billion tonnes of CO₂ from the atmosphere every year but it’s never going to happen for many different reasons: technical, financial, social, environmental, governance, etc. If we don’t reduce CO₂ emissions, we’re screwed. That’s it.

Saying we need massive CO₂ removal is like saying we need fusion power. We might need them but we ain’t got ‘em.

I had a classmate who was completely incapable of understanding the concept of sarcasm in writing but who became a STEM professor. Minds are fascinatingly varied. (Which also shows how silly it is to take IQ scores seriously or to believe that AI is a linear thing that you can just crank up to 11.)

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

Leaving aside the whole "future soldier" bit, which implies that they have a very different understanding of "lest we forget" than I do, why do so many people seem to think that poppies have two petals? Have they never seen a single poppy flower?

Computer vision species identification of lichens and bryophytes from biocrusts in Australian drylands New in #AppsPlantSci by Lawler, @anschmidtlebuhn.bsky.social, Cargill & Gueidan bsapubs.onlinelibrary.wiley.com/doi/full/10.... #lichen #bryophyte #biocrust #MachineLearning #ComputerVision

Yes really — the “we asked this guy. He doesn’t actually know jack shit but he’s rich” model of reportage really needs to go

Unusually, the ASBS awards two Nancy Burbidge Medals in 2025. The first is awarded to Ian Telford, in honour of his lifetime contributions to Australasian botany. He scientifically described over 100 taxa and collected over 14,000 specimens. #asbs2025

The ancients who wasted enormous resources on religious monuments while their societies imploded were much more rational than we are. At least they had the excuse of not understanding stuff like salinisation, soil erosion, or why the rivers were drying up. We do understand, yet we do stuff like this

A session of short updates begins with Peter Jobson discussing how the census of the flora of New South Wales is updated. #asbs2025

#asbs2025 continues with lightning talks. Tareg Shaldoom, who is also providing technical support for the meeting, begins with a talk about the Phebalium glandulosum complex in the citrus family. Several putative species revealed; common garden experiment to exclude ecological plasticity.

seems bad! www.theatlantic.com/technology/2...

#asbs2025 continues with a lecture by the 2024 recipient of the Nancy Burbidge Medal, Karen Wilson. (There was no conference in that year, so she is being honoured this year.) Karen is an expert in Casuarinaceae and Cyperaceae, among other groups.

The cake celebrating the founding of the Palaeobotany and Palynology Chapter of the Australasian Systematic Botany Society. #asbs2025

Even at this botanical conference, the disconnect re genAI is evident. On the one hand, overwhelming amounts of slop "art" and comments implying that AI is a magic wand you can wave at research questions. On the other, some colleagues plainly stating that genAI is GIGO, dangerous, and overhyped.

Australasian Systematic Botany Society meeting 2025 commencing in Armidale. Lots of great science to look forward to. #asbs2025

My prediction is that in the long term, research publishing will be a lot about editors knowing which authors they can trust to do serious work. But this means it will become much more difficult for early career researchers to get established without relying on patronage. Generative AI, yay.

arXiv will no longer accept review articles and position papers unless they have been accepted at a journal or a conference and complete successful peer review. This is due to being overwhelmed by a hundreds of AI generated papers a month. Yet another open submission process killed by LLMs.

Will be at a conference most of next week, and now I have to check if that holds in our field. It may not. In contrast to many of the respondents to this post, we generally show photos of study organisms, maps, phylogenetic trees, STRUCTURE plots, etc., which LLMs can't do.

Not providing the option of "no and never ask me again" should be illegal and punishable by being catapulted into an active volcano.

Okay, so this writer uses an LLM to check his writing and seems extremely proud or surprised (?) that the LLM trained months ago isn't aware of recent events. The question arises: why not use standard office software to flag errors with red underlines? Long available, much less wasteful than LLMs.

Ah, it is that piece again entirely based on pretending not to understand what 'emergence' means, only this time without a link, it seems? I could write a piece arguing that we should reject the concept of gravity if I started it by pretending that physicists see gravity as magic.

What drives me nuts about stuff like this is that anybody with any sense of scale can immediately see that this would be a drop in the ocean compared to CO2 outputs. It is the equivalent of my field organising a conference to discuss if we can understand plant evolution by studying two leaves.

Seriously, web interface designers should be forced to use their own... creations, at least for a few months. Trying to request open access publishing at the Springer Nature website just now made me want to throw the computer out of the window.

I am committing to spend ten trillion on the energy transition. Wait, are you saying that I need to actually have money to make spending commitments? How strange.

Nothing matters anymore, including that the title of the alleged article, its subtitle, and the social media post announcing it are all the same copy-pasted text. Such dedication to the craft!

Saw a headline about homeschooling popularity being on the rise, so it's time for the periodic reminder: You aren't smart enough to home school your child or children.

Based on a recent discussion about 'AI' with other scientists, here are the two things I'd really like everybody to understand. First, from what I am reading, 'hallucinations' are part of how LLMs work, so they are never going away, so you cannot use them for anything that requires reliability.

This is also just not reflective of how people do things. In my discipline, yes, the first author is generally the postdoc or grad student who led the study, and the last author is generally the supervisor or whoever got the funding. But there are other fields that simply do alphabetical order.

I remember one of the last two times I decided to give Windows a chance. Bought computer, found it needed 45 min to install updates, so I did dishes. Nearly had a heart attack when Cortana suddenly said something, bc I thought some stranger was in the room. Wiped the machine and installed Ubuntu.

This thread resonates with me a lot. When I read the Foundation novels, I kept wondering where that great scifi writer was that everybody sees in Asimov. Even if you leave aside his misogyny (which you shouldn't), he had very little imagination.

This would be obvious supervillain stuff if the robots worked out. Now to see if they will work exactly as well as the boring tunnels, full self driving, Mars colonies, and Grok becoming a superintelligence.

Bioinformaticians and their software names... surely there will be no confusion between this and RogueNaRok 😀

I for one and pretty sick of the “I voted for the Trump in my imagination” excuse. Voters are never going to study policy statements on websites; that’s not a realistic standard. But if you can’t even get the vague *gist* of what a candidate wants to do, that’s a You problem. Skill issue.

The average age of U.S. homebuyers is now 56, up from 49 last year. In 1981, the year trickledown economics began, it was 31.

Ooof. Just read that thread. I guess my attitude is that if I'm writing a grant on a topic, I should know enough about it that I don't need AI to find citations and develop a question for me. If I don't know what a specific gene/pathway does in a process, should I really be writing a grant on it?

The usual thought exercises, just for fun: If you had predicted in 2023 that Trump would demolish part of the White House, how would Sensible Centrists in the USA have treated the prediction? If Obama had done this, how would US conservatives have reacted?

Apart from the incompetence and laziness of the professor, what is so bizarre about this is the idea of some random professor having their own textbook that they teach from. In my field we had The Textbook that was co-written by several top people in the field and was used across many universities.