Alexander Schmidt-Lebuhn

Alexander Schmidt-Lebuhn

@anschmidtlebuhn.bsky.social

Botanist, taxonomist, phylogeneticist.

411 followers114 following2.6K posts

Pinned

Kind of obvious if you consider that (a) it is much more dangerous to life and limb for a cyclist to behave in an unexpected way in traffic than it is for a car driver, and (b) cyclists find it slightly more difficult to exceed 50 km/h speed limits than car drivers.

Brent Toderian@brenttoderian.bsky.social

“Cyclists are often perceived as recklessly flaunting road laws. A new study challenges this perception. The study found only 5% of cyclists broke road rules while riding on bike infrastructure versus 66% of motorists who broke rules while driving.” Safe bike infrastructure fuels rule-following.

Fascinating how easy it is to accidentally hang up on a phone call by moving a hand somewhere within 5 cm of my wireless headset, but whenever I deliberately want to trigger the button, it doesn't react at all no matter what I try.

Can we just stop trying to force AI into everything? By definition, an expert peer reviewer should already know where to focus their scientific judgement. The problems, if there are any, are editors not picking competent reviewers, or authors freeloading by refusing to review for others.

Springer Nature@springernature.com

How can AI support researchers in peer review? Springer Nature and Google DeepMind are partnering to co-design an AI peer review assistant that helps reviewers navigate manuscripts and focus on where their scientific judgement matters most. Learn more: spklr.io/63320EVgT8

I am reminded of a blog discussion months ago regarding the socially, societally, and environmentally catastrophic consequences of the normalisation of everybody having a car with a combustion engine, where somebody thought that why do you hate ambulances was a killer counter-argument.

SE Gyges@segyges.bsky.social

being anti AI is a luxury belief only available to first world people, especially Americans imagine how important machine translation is for people whose first languages aren't that common

Also, just a few years ago we didn't have LLM labs stealing all our data to flood the web with garbage art, fraudulent eBooks, fake websites, and sparkly chats for texts that already have summaries. Somehow, we managed to afford that luxury. I remember once using Google Translate for Catalan, etc.

Alexander Schmidt-Lebuhn reposted

People say socialism is the future but I ask you if socialism could foster an innovative tech industry that trapped us all in a suicidal arms race that will collapse our economy while using up all our resources and funnelling billions of dollars into the pockets of 6 people? I think NOT!

Alexander Schmidt-Lebuhn reposted

personally I think it's neat that the tech industry persuaded us all to put everything online, and then destroyed the internet

Matt Novak@paleofuture.bsky.social

Jason Clinton is the Deputy Chief Information Security Officer at Anthropic.

Good to know that outsourcing makes everything better. Why would a country want to maintain control over key strategic assets if those could just as well be handed over to oligar-, sorry, I mean to the ineffably efficient free market that only ever delivers optimal outcomes?

Paul Guinnessy@paulguinnessy.bsky.social

Space Force recently tried to book rides for three future satellite missions, but #SpaceX declined to offer its services, as it doesn't want to sell the Falcon rockets anymore. No other rocket companies offered to take them either. #EverythingIsFine

Alexander Schmidt-Lebuhn reposted

Alexander Schmidt-Lebuhn reposted

A reminder: 🚗: I hope I find parking. 🚴🏽‍♂️: I hope I don’t die. Our priorities are not the same.

Alexander Schmidt-Lebuhn reposted

I genuinely don't think that human civilization is compatible with "100 dudes should own everything, make nothing, but get infinitely increasing money forever"

Oh come on now, surely this has to be a deepfake produced by people who want to hurt Microsoft's reputation? Microsoft themselves wouldn't do this, right? Help?

KetAnthropic JoshAI@ketanjoshi.co

The first genuinely honest ad for generative AI I've ever seen, from Microsoft / CoPilot A mediocre white guy sees his non-white colleague getting praised for good work, he asks AI to *literally copy it*, and then he gets all the praise instead: www.ispot.tv/ad/SoDi/micr...

Alexander Schmidt-Lebuhn reposted

Ah, we are back to the Metaverse again, but also, WHY ARE YOU SHOUTING ALL THE TIME?

FinTwitter@fintwitter.bsky.social

NVIDIA CEO JENSEN HUANG SAYS HE BELIEVES MOST PEOPLE IN GERMANY WILL MOSTLY WORK INSIDE A DIGITAL TWIN OF THEIR COMPANY, BUILDING DIGITAL CARS WITH DIGITAL BRAINS, WITH AI ROBOTS BUILDING ROBOTIC CARS INSIDE ROBOTIC FACTORIES

Instead of searching for the next best way to score biologists after "how many papers", maybe job and grant panels could actually read the work of the candidates and understand if it has value. Inconceivable? Nobody has the time for that, they just want the prestige of serving on the panel? Sigh.

Ran Blekhman@blekhman.bsky.social

Some of my thoughts about the surge in AI-generated papers, why this feels like a completely new kind of problem, and what this means for how we evaluate science in biology. Out now in PLOS Biology: journals.plos.org/plosbiology/...

So, banning somebody or suspending their account on social media is an "editorial decision"? Okay, I will take that if it means that Zuckerberg will be held legally responsible for all posts on Instagram the way an editor would be for the contents of the newspaper/book. Not holding my breath here.

Mike Masnick@masnick.com

People demanding Zuck be hauled before Congress to explain why No Kings insta was suspended... how would you feel if MAGA GOP hauled MSNow execs to explain why they covered Trump the way they do? Both are editorial choices protected by the First Amendment. If this annoys you...

Ideally, the bazillionaires owning social media want to have it both ways: the power to ban people that a newspaper editor has but without the social expectation that said editor apply that power to uphold the standard of factual accuracy.

Alexander Schmidt-Lebuhn reposted

We already have flying cars. They're called helicopters. They're really expensive and there are good reasons you have to have a pilot license and work with air traffic control to fly them. Putting jets on a sedan is not going to result in a better vehicle or transportation experience in any way.

It is really odd how half of us are frustrated about how unreliable AI is and how much shallow slop its users are producing, and the other half constantly talks about how great it is, try to force LLMs into every workflow, and unironically talk about "the AI revolution" &c. This can't go on forever.

Phil Lewis@phillewis.bsky.social

Employers worldwide are pulling back on AI as they navigate a flood of errors and low-quality work that sap employee time, raise privacy and security risks, and drive up costs

Relatedly, the same "both can be true" applies here. There are many who count as LLM users because they felt they had to give it a try to see if it works or because it is forced upon them, but many others went all in on producing work slop and would not want to lose that. bsky.app/profile/peps...

Peps2025@peps2025.bsky.social

"moar peepl ar using it than evar" My brother in christ the 'AI' features that infest every search engine, third rate social media app, and phone are turned on by default. You are forcing these products on everyone and saying "look everyone uses them"

It is very frustrating how many people in discussions of authoritarianism / dictatorship seem to believe that as long as there are still elections, all is fine, what are you complaining about. For example, they can have elections but shut down polling stations where most opposition supporters live.

Ghostly Bree@mostlybree.kitrocha.com

eternal scream of RAAAAGE those kids should not have to do this any more than the people waiting in line for 12 hours in Georgia should have to do that.

They can purge ethnic groups or social classes likely to support the opposition from voter lists. They can ask their billionaire buddies who own traditional and social media to suppress communications of the opposition, etc. Funnily enough, people get it with countries that are designated enemies...

Alexander Schmidt-Lebuhn reposted

The tragedy of the boom and bust cycle is that you are not allowed to say "this is a boom and bust cycle" during the boom part, even if it is obvious. Meanwhile the people who were saying "this boom will last forever" will now smoothly pivot to "the bubble has popped, like I always said it would"

Unusual Whales@unusualwhales.bsky.social

BREAKING: Bank of America is warning that the AI boom is starting to look like a classic boom-and-bust cycle.

Some faint hope at last?

Simon Evans@drsimevans.carbonbrief.org

NEW: Another demand downgrade in the latest IEA oil market report Previously, IEA said a rebound in 2027 would outweigh the Homuz-related drop in 2026 Now, it sees oil failing to make up lost ground next year – demand will remain below 2025 levels

It is worse, because Civ's France or Germany at 1939 equivalent in-game military tech would also lose against Civ's Carthage or Songhai at 2026 equivalent in-game military tech. Despite introducing nation-innate abilities in later versions, Civ is IMO actually a good antidote to that kind of racism.

jamelle@jamellebouie.net

i have been following this (unbelievably stupid) discourse and it really seems like a bunch of racists think the world works like a game of civilization

Alexander Schmidt-Lebuhn reposted

It's frankly incredible that hitting me with a stick and going " Oh no! The stick has gone out of control!!" fools regulators.

Ben Collins@bencollins.bsky.social

The most indebted corporations in America right now are, at their core, computer virus companies. They think the computer viruses are their girlfriends, so they think this absolves them of guilt. Let's find out!

Bit puzzled by a recent experience why scientific journals still take keywords so seriously. When I search for literature, the search tools clearly search in titles and abstracts, so what are keywords even for? If the title mentions the name of the study group, why does it have to be a keyword too?

No idea where he gets the idea there is any evidence for token prediction models being conscious, but "that's absurd" is only a strong claim if the rejected idea is not absurd. In this case, it is. Repeating myself here, but start an LLM session and watch what it does if you don't provide a query.

Will Stancil@whstancil.bsky.social

Even raising the prospect is taken as absurd, when in reality it's the people absolutely certain they AREN'T conscious who are making the stronger, less-evidenced claim.

The problem is, and this isn't a new idea either, that we humans are very keen to impute sentience. We have done it to weather, death, strange noises, even the existence of the universe. No surprise we do it to software that actually manages to write coherent sentences. Doesn't mean that is correct!

Yes. A technology will get used predominantly in certain ways determined by its core functionality. You could use a microwave as a foot rest, but it is mostly going to be used for microwaving. You could do (very few) useful things with an LLM, but it is mostly going to be used for spam & plagiarism.

Bruno Dies@brunodias.dev

I am becoming incredibly exhausted of this pattern: many people: "AI sucks!" Very Smart Person: ah but have you considered (distinct application of ML or statistical modeling that has no real relationship to what the tech industry is marketing as, and people broadly perceive as, "AI")

People just love doing this, in any context. I once complained about car culture warping our entire society around itself - commute times, space taken up by roads and parking lots, traffic fatalities, fossil fuel use - and the push-back was: what about ambulances.

Alexander Schmidt-Lebuhn reposted

Replying to @iwriteok.bsky.social

can't wait to build my Robot That Mag Dumps 14 Rounds of 400 Winchester Magnum in a Random Direction Every Hour. i assume this is legal, since the precise time at which the trigger gets pulled is randomized!

Alexander Schmidt-Lebuhn reposted

Replying to @carlquintanilla.bsky.social

We’ve learned nothing from the Tulip Mania 1634-1637

Recently had a lengthy discussion with other commenters on a blog on whether to take the warnings of AI doom coming from within these companies seriously, and it is clear hardly anybody appreciates how deranged many of their staff are.

Mike Masnick@masnick.com

Over a year ago I met an Anthropic employee at a conference and at one point she mentioned that one of the things she was working on was whether or not it was cruel to ever turn off an AI model. I thought she was joking. She was not.