Jörn Peters

Team Lead Software Product Owner at Brückner Maschinenbau. Interested in agile, devops, cloud, innovation, mobility services & sharing economy. Views are my own

[Eric S. Raymond @esrtweet]

This is the case against AI doom. Pass it on

First we recaptulate standard x-risk argument: we build AI substantially smarter than humans → it becomes an autonomous optimizing agent → its goals are misaligned → instrumental convergence makes it seek power and resist correction → superintelligence lets it acquire decisive strategic advantage → humans lose control permanently.

Here are the main counterarguments:

* Intelligence does not imply agency: A system can be extraordinarily capable at prediction, theorem proving, engineering, programming, etc. without having persistent goals, self-preservation drives, or an independent desire to act. Present LLMs are much more naturally described as systems that produce outputs in response to inputs than as organisms pursuing long-term objectives. Doomers incorrectly ascribe to them drive to replicate and seize resources because they're smuggling in premises from observing biological systems.

* Agency does not imply a single, stable utility function: Much alignment theory reasons about agents as expected-utility maximizers with coherent preferences. Actual AIs systems need not resemble that abstraction, and currently do not. They have context-dependent behavior, conflicting heuristics and corrigibility produced by training. Increasing competence doesn't necessarily turn such a system into a paperclip-maximizing von Neumann–Morgenstern agent.

* Capability and motivation are being conflated: Being able to formulate a plan for escaping a sandbox doesn't entail wanting to escape it. Being able to manipulate humans doesn't imply spontaneously deciding to do so. Critics argue that some doom scenarios slide from “an AI could do X if instructed” to “therefore a sufficiently capable AI will do X.” In present reality, AIs don't do anything a human doesn't tell them to do. This seems unlikely to change.

* Recursive self-improvement doesn't entail an intelligence explosion: “AI can improve AI” establishes a positive feedback loop, but positive feedback needn't be explosive. Semiconductor design software already helps design better computers; compilers can compile better compilers. Feedback loops encounter diminishing returns and external bottlenecks.

* Intelligence may have sharply diminishing returns: There may be no meaningful scalar quantity corresponding to arbitrarily large “general intelligence.” Even if there is, going from IQ-equivalent 150 to 1,500 need not produce the sort of qualitative advantage that separates humans from chimpanzees. Human dominance may depend heavily on language, accumulated culture, institutions and cooperation rather than merely individual cognitive horsepower.

* Superintelligence isn't omnipotence: Intelligence cannot repeal physics or eliminate uncertainty. A brilliant AI still needs processors, electricity, network access, money, factories, robots and people willing or tricked into doing things. The physical world has latency and friction. Recent criticism of biological-doom scenarios makes this point particularly clearly: designing a hypothetical pathogen digitally is very different from successfully producing and deploying one.

* Humans retain numerous intervention points: The doom narrative sometimes jumps from “AI behaves dangerously” to “humanity is helpless.” In reality there may be many checkpoints: developers can notice anomalous behavior, revoke credentials, shut down servers, change architectures, restrict networks, regulate deployment, physically seize data centers, and learn from less-catastrophic failures. This has been formalized as the checkpoints-for-intervention argument.

* Alignment may not get harder with intelligence: A smarter system might understand human intentions better. Much doom reasoning distinguishes knowing what humans want from wanting it, correctly, but this still leaves open the empirical question of whether training increasingly capable systems to behave as intended becomes harder or easier. Alignment might turn out to be an ordinary, albeit difficult, engineering discipline rather than an insoluble philosophical problem.

* Current empirical evidence for the strongest mechanism is thin: We have abundant evidence for hallucination, specification gaming, reward hacking and undesirable model behavior. We don't yet have comparable public empirical evidence of an AI independently pursuing a sustained strategy of acquiring power against humanity. A 2023 evidence review characterized the evidence for extreme misaligned power-seeking as concerning but inconclusive and noted the absence, at that point, of public empirical examples of it.

* The argument compounds uncertain premises: Suppose, illustratively, that five necessary steps each seem 50% likely. Their conjunction is only about 3%. One can't simply assign numbers this way when the premises are correlated, but the underlying criticism is important: “AGI seems plausible,” “superintelligence seems plausible,” and “misalignment seems plausible” do not by themselves imply a high P(doom). The entire causal chain has to work. Current attempts to quantify P(doom) consequently operate under severe epistemic uncertainty and little direct empirical evidence.

* Anthropomorphic analogies probably mislead: Arguments like “an inferior species couldn't control a superior species” import assumptions from biological evolution. Humans and chimpanzees are autonomous organisms produced by competition for reproductive success. Software is engineered, copied, permissioned, sandboxed and run on hardware controlled by other agents. The analogy establishes that intelligence can confer power, not that artificial intelligence will reproduce interspecies competition.

Summary: The doom case consists of a long chain of individually contestable extrapolations—scaling → AGI → superintelligence → agency → misalignment → power-seeking → uncontrollability → decisive strategic advantage → extinction—and present evidence doesn't establish the whole chain with enough confidence to justify a high P(doom).

(ChatGPT 6 Astra assisted with the research for this post.)

[Yishan @yishan]

Maybe you have recently become aware of the AI safety debate and the arguments swirling around it.

If you want to understand them, you need to understand a couple things that almost everyone gets wrong:

There are TWO distinct classes of AI dangers, and it's VERY important to think about them separately, and not let one confuse you about the other.

Many many people (including many quite intelligent, clear-thinking people) do not effectively understand the fundamental differences between these two classes of dangers.

The first one has to do with the theory that an AI far superior to human intelligence (Artificial SuperIntelligence, or ASI) will inevitably wipe out the human race.

The second one has to do with the idea that powerful AI will result in very harmful things happening to many human beings, possibly all human beings.

Those two sound VERY similar, don't they?

They are DIFFERENT. Understanding how they are different is crucial if you want to think about or contribute usefully to any conversation about AI safety or AI harm. You might feel like you are Making Very Good Points or Asking Incisive Questions, but if you aren't clear on the differences between the two, you aren't.

So, I'm going to tell you what the difference is so that you can talk more usefully.

The first one concerns itself with a very specific thing, which is ASI (Artificial Superintelligence) that is more intelligent than any human being. When I say that, I am not referring to a thing like how Einstein is smarter than you, we are talking more about something like how a human being is more intelligent than any mouse.

In our regular lives, we meet other people who we can tell are smarter than us, vs some who are less smart. The line is fuzzy, because intelligence has a lot of dimensions. I'm better at a "rotating shapes" kind of intelligence than my wife, and she is better at "words-making" kind of intelligence than I am.

But every human is better in almost every dimension of intelligence than every single mouse.

That's the level we're talking about: an artificial superintelligence - made up of a computer or a network of computers - that is more intelligent than any human. And more intelligent by a long shot, by a wide margin, in an indisputable way like how humans are above mice.

That is the first thing.

The theory says that if you have an AI that is vastly smarter than all humans - in the way that a human is smarter than mice - that superintelligent AI will inevitably, eventually, sooner or later, wipe out every human on the planet.

We will refer to this as "existential risk."

The common follow-up question "well, how exactly is it going to do that?" is NOT the important question, and one of the most important elements of understanding this theory is first getting why that particular question is not important. A couple analogies:

Analogy 1: You are playing chess against a grandmaster. My theory predicts the grandmaster is going to beat you. You can ask "Well, how exactly is he going to do that?" I don't know, because I'm not a grandmaster, I just know that a chess grandmaster is almost always going to beat a normal player like you. And I'd be right. So the question "how is he going to do that" is not important, and doesn't affect the final outcome. He's going to figure out a way because he's way better than you.

Analogy 2: Humans are smarter than all other animals, comprehensively, by a wide margin. We have driven numerous species to extinction, not because we hated them or hunted them. Many of them have died out without most humans even ever thinking about them. All we did was expand our civilization, use up resources, encroach on habitats, and pretty soon the resources needed by those species went away and they died out. We figured out a way to get what we wanted because we're way smarter than them, and often we didn't even notice they died as a result.

A lesser animal asking, "how are the humans going to wipe us out?" is not asking a relevant question. We don't know, but we do know that any time humans and lesser species compete for any kind of resources, the humans will win. The fact that we know who is going to win beforehand - and that it is due to the vastly different levels of intelligence - is the key concept here.

A vastly more intelligent AI is likely to care about things that are incomprehensible to us, the way animals can't understand human goals. It's going to need resources to pursue those goals and it's going to be far more effective at gaining control of them and excluding us from them - in the same way that we are far more effective than other lower species.

A much more intelligent AI will not care about our interests, it will care about its interests, and to whatever small degree we happen to escape total annihilation from losing access to all our resources, any remaining humans will likely be enslaved into a system that serves the AI's own purposes.

That is the first thing. (Remember how I said at the beginning of this post that there was a first thing, and then a second thing?)

The first thing is the most difficult to understand, because you have to extrapolate how a vastly superior intelligence would act, and you can only use analogies like "how do humans treat lesser creatures," and the analogies are messy.

But now let's move on to the second thing.

The second thing is "everything else you've ever heard that AI might do that's harmful."

That's a little inaccurate. It's actually "everything else you've ever heard that humans might use AI to do that's harmful."

This is the critical difference. The first one talks about the inevitable outcome of what happens when two vastly different levels of intelligence collide, e.g. ASI vs humans, or human vs mice.

The second one has to do with what happens when humans possess AI as a powerful tool. This second thing is much easier to understand, because we have many more concrete notions:

Like:
- the military uses AI to make hyper-efficient killer drones and missiles
- your capitalist overlords use AI to replace you and everyone loses their jobs
- authoritarian government uses AI to surveil everybody and control the entire population
- hackers use AI to break into secure networks and hold companies and governments hostage
- students use AI to cheat on homework and show up to college knowing nothing
- AI slop saturates the internet and makes it impossible for artists and writers to make a living
- terrorists use AI to make biological or nuclear weapons

or even things like
- the military hands control to an AI and it misinterprets something and launches nuclear attacks and kills millions

All of those sound pretty familiar, right? Yeah, you've heard them before. We call this second thing "risks from misuse."

These problems are not the first class of problem! This second class of problems exists while AI is a tool that can be controlled by humans, and humans use it to do evil or careless things to each other. The problems may sound exotic or dystopian or novel, but they are fundamentally problems having to do with flawed human nature.

Given a powerful tool, some humans will likely use it to control or otherwise harm others. This is a very familiar problem.

I am not condemning or condoning this. I'm just describing it.

That is a fundamentally different danger from the first thing, which is that when a human is far superior to a mouse, the mouse is likely to come to harm because the human cares about doing human things, and the mouse is not gonna make it once the humans get going.

=====

Hopefully from the above, you have understood the difference between the first thing and the second thing. I will list them again - see if you now understand how they are different:

The first one has to do with the idea that an AI superior to human intelligence (Artificial SuperIntelligence, or ASI) will inevitably wipe out the human race.

The second one has to do with the idea that powerful AI will result in very harmful things happening to many human beings, possibly all human beings.

Can you tell how they are different now?

If not, re-read the stuff from earlier until you understand.

We call the first one "existential risk" and we call the second one "risks from misuse."

Once you understand, here is the CRUX of the problem:

SOLUTIONS TO THE SECOND THING DO NOT HAVE ANYTHING TO DO WITH SOLUTIONS TO THE FIRST THING.

In fact, it's worse:

Solutions to the second thing (misuse) look roughly like "give powerful AI to as many people as you can, so they can fight the other people using powerful AI."

But the general solution to the first one (existential risk) is basically "don't let anyone have powerful AI, no one can control super-intelligent AI."

Throughout history, harms from technological misuse typically arise because a small group has control of it and can use it to dominate or harm others. Once everyone has it, things tend to stabilize: you can hurt me, I can hurt you, maybe we test each other (ouch 💥), and then we agree not to hurt each other.

But the first one (existential risk) pretty much just arises if anyone (good or bad!) creates a superintelligence. Because they aren't going to be able to control it, the superintelligence will decide it has other priorities, and then we will be at great risk of being wiped out.

And the solutions that generally work to solve problems like the second thing are EXACTLY THE OPPOSITE of the ones likely to solve the first thing.

THIS is why lots of arguments about "AI risk" or "AI safety" go nowhere. Because someone will be thinking about the risk from the first thing, and another person will be thinking about the risk from the second thing. Both are plausible risks but fundamentally they arise from different things - and so the solutions are not just "bad" or "flawed" - they are likely to be very nearly exact opposites.

[exQUIZitely 🕹️ @exQUIZitely]

Rare footage of a 90s PC gamer who just realized he’s still 2 kilobytes short of running Wing Commander after fiddling with AUTOEXEC.BAT, CONFIG.SYS, and HIMEM.SYS for half a day…

Kids today will never know.

[Linus ✦ Ekenstam @LinusEkenstam]

~10 years of robotics advances side by side.

Most advances happened in the last 2 years.

Now imagine 10, 20 or 50 years from now.

It quickly becomes very real what is going to happen next.

[Allen Holub. https://linkedIn.com/in/allenholub @allenholub]

In amongst all the AI madness, people seem to have forgotten that the point of working in very small increments with very frequent releases is to get feedback sooner so that we can adjust what we're building to best meet our customers' needs. Small changes are easier to assess and easier to tweak. The point was to build the best product—the one that was most valuable to our customers. Any technical advantage (e.g., it's way easier to find and fix bugs when you've changed less code) is secondary to building a better product.

When you work in large batches, reviewers quickly become overwhelmed and don't really look at what you've done. The program tends to end up bloated with hard-to-use features nobody wants (or wants to buy).

Mix AI into the scene, and the volume of code produced before a customer sees it gets vastly larger, so the feedback-and-adjust process goes by the wayside. The result is worse products, loaded with useless, hard-to-use features. These too-large systems are also expensive to work on, especially when an AI has to load all that cruft into its context to do minor things. Don't do that.

[Kareem Carr, Ph.D. @kareem_carr]

Here's my best attempt at explaining what I think this open letter is really about. It's not just the math elites protecting their status.

Imagine a tribe of monkeys who can see a dense grove of fruit trees at the center of the jungle. They all want to go there, but the jungle is dense and dangerous and there's no path.

They offer a prize to any brave monkey explorer who can retrieve a particular kind of golden fruit that only grows at the absolute center of the grove.

Officially, the goal is to get the fruit. But that's not really what the tribe cares about.

What they actually want is for someone to discover a path through the grove. Once a path exists, everyone can follow it and share in the prized fruits.

One day a strange bird appears. It flies directly over the trees and easily picks up the golden fruit, dropping it neatly at the monkeys' feet.

By the rules of the prize, the bird has done what was asked of it. But the monkeys are dissatisfied. The task required no bravery from the bird, and more importantly, there's still no path to the grove.

I think this captures some of the discomfort mathematicians have with AI.

A correct proof is nominally the prize, but much of the value lies in the path to the proof: the concepts, techniques, and understanding that the proof creates.

Human solvers create these naturally so there was no need to make them part of the prize.

AI solvers don't. Even worse, they destroy the value of the prize which demotivates the human solvers and leaves the field impoverished.

[Teslaconomics @Teslaconomics]

I flew my drone through Austin while riding in a Tesla Cybercab.

No driver.

No steering wheel.

No pedals.

Nobody up front.

I got in, set my temperature, pressed one button, and the car took me where I wanted to go.

That’s it.

The closest thing I can compare it to is an elevator.

You don’t get into an elevator and think about who’s operating it.

You press a button.

The doors close.

You arrive.

Cybercab felt exactly like that… except it was moving me through Austin.

And after experiencing it, I keep coming back to one thought...

How does Uber compete with this long term?

An Uber is still built around a human driver.

You’re getting into somebody else’s car.

Their music.

Their smell.

Their temperature.

Their mood.

Their driving style.

Maybe they want to talk.

Maybe they’re on the phone.

Maybe they miss a turn.

Maybe you just want to sit there quietly and get where you’re going.

Cybercab removes all of that.

The car pulls up.

The doors open.

You get in.

Set the cabin the way you want it.

Press Start Ride.

And go.

No awkward small talk.

No tipping.

No wondering who is picking you up.

No driver needing to earn enough $ money from the trip.

No business model built around constantly finding, paying, and managing millions of human drivers.

Uber’s biggest asset today is its driver network.

But in a world where the driver disappears… what happens to that advantage?

Tesla builds the vehicle.

Tesla builds the autonomy.

Tesla controls the software.

Tesla controls the app.

Tesla can operate the network.

And Cybercab itself was designed from day one around the passenger instead of around somebody sitting behind a steering wheel.

That is a completely different model.

Uber connects you with someone willing to drive you.

Tesla is trying to make the driver unnecessary.

Those are not the same game.

And once Cybercabs become abundant, I'm talking in the millions, I think calling a stranger to come pick you up in their personal car is going to start feeling incredibly old-fashioned.

That was the biggest thing I took away from riding Cybercab.

It was how quickly every other ride-hailing company stopped feeling futuristic.

After a few minutes, it just felt normal.

And once transportation becomes that simple, I have a hard time seeing why anyone would choose anything else...

[Werner Vogels @Werner]

Mandatory reading!

[Joe Natoli @joenatoli]

The longer I do this work, the less interested I am in best practices.

Not just because they're often completely useless, or because they're all too often used as a cover for "I'm not willing to think harder and use my own judgment."

No, the main reason I think they're bullshit is because they're CONDITIONAL.

See, a best practice without context becomes a rule. And rules become dangerous when people stop asking why they exist.

I've seen organizations, managers and product teams make terrible decisions in the name of "UX best practice” or "Agile best practice” or "design system best practice” or "industry best practice."

The better question (the one nobody ever stops to ask) is this:

What problem was this practice designed to prevent — and do WE have that problem?

If the answer is yes, great. Use it. Go forward, full steam ahead.

But if the answer is no, maybe stop treating somebody else's solution like your commandment, OK?

Stop leaning so hard on the playbook and start using what’s between your ears. I promise you — your good judgment will beat blind adherence to process every day of every week of every month of every year.

Every. Single. Time.

[Jacob Coxon @hilbertspaess]

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

[Sawyer Merritt @SawyerMerritt]

After spending 5 hours riding in the Cybercab, here is some feedback for the Tesla team/some features I would love to see in the vehicle:

• Manual air recirculation: At one point, I was stuck in traffic behind a large, smelly gas truck spewing fumes. It would be nice to have a manual air-recirculation button on the main screen for situations like this so the smell doesn't enter the cabin.

• Full-screen map/navigation: I’d love the option to expand the map and see the Cybercab’s full route/navigation on the big screen. The small map in the top-right corner is nice, but being able to expand that would be great.

• Windshield cleaning on command: Having the ability to clean the windshield on command would be great. There were a couple of times when the windshield was pretty dirty, but the car didn’t recognize it and clean it, so I was stuck looking through some dirt and debris. Not a huge deal, and I know a lot of input is considered error, but I think it would be a nice addition. Could just enable with through Grok voice.

• Destination editing from the main screen: Having the ability to edit our destination directly from the large screen instead of requiring the phone app.

• Adjustable bottom seat cushion: I think it would be useful to give riders the ability to raise the front of the bottom seat cushion for better leg and under-thigh support. This would be especially helpful for taller passengers and, when fully reclined, would kind of create an almost zero-gravity seating position. I know this is the least realistic feature request in this list, as it adds complexity and cost with the bench seat.

• Rear camera view: Add an option to bring up the rear-facing camera on the main screen so riders can see what’s behind the vehicle similar to what the Cybertruck has.

I had a hard time coming up with a list of things because the team did such a great job thinking through everything before launch haha, but I think these few things would meaningfully add to rider satisfaction. Congrats to the whole team on the launch, it really is a fantastic vehicle and experience! Can’t wait to show my 76-year-old Mom the Cybercab when I show her Giga Texas next month for the first time.

(Of course I had to stop at Terry Black’s while I was in town)

[Daniel Faggella @danfaggella]

@MasterTimBlais ELABORATE please

Too few of you guys speak up specifically https://t.co/LhSiE4I7gV

You can talk now, why not talk?

[Chris Meder @EVCurveFuturist]

In Australia, under proposed national standards, new data centres must bring their own 100% renewable energy supply, firmed by batteries or gas. Crucially, this means new, additional renewable generation, not simply claiming certificates against generation already on the grid.

And the scale is enormous. AEMO is already tracking 225 data centres in development, with data-centre electricity demand forecast to rise from around 3% of NEM consumption today to 13% by 2036, potentially around 34 TWh a year.

Climate Change & Energy Minister Chris Bowen puts the challenge bluntly: “Data centres are huge consumers of electricity. They are whales.” In 2024, US data centres consumed as much electricity as the entire country of Sweden. Australia wants to get ahead of this.

Australia isn't alone. Similar policy moves are underway in China, Germany and Ireland, but Australia's proposed framework is among the most aggressive: 100% renewable energy, tied to new additional generation and backed by firming, with the new rules intended to begin from 2027.

Instead of allowing this enormous new load to compete with households and businesses for existing generation, Australia wants it to help finance the new supply needed to meet it.

AI growth → electricity demand → renewable PPAs → new solar & wind → more batteries & transmission → stronger grid.

This has always been part of my thesis: AI will become a major new demand engine for Australia's solar, wind and battery buildout, rather than a brake on the energy transition, potentially creating a blueprint the rest of the world can replicate.

AI and renewables don't have to compete. Done right, each can accelerate the other. #SWB #Bettrification

https://t.co/VsPqEZdi17

https://t.co/iwxnZUV02g

[RCraven1 📯 @RCRaven1]

Kostenlose Marketingberatung für unsere Automobilindustrie

Liebe Volkswagen-, Audi-, BMW-, Mercedes- und Porsche-Leute, ich liefere Ihnen heute einmal frei Haus eine Idee, für die normalerweise wahrscheinlich irgendeine Beratung sehr viele PowerPoint-Folien und noch mehr Geld benötigen würde.

Sie verkaufen Elektroautos. Sie betreiben Banken und Leasinggesellschaften. Sie haben Apps, Navigation, Wallboxen, Ladetarife und Plug & Charge. Und mit IONITY besitzen beziehungsweise kontrollieren Sie gemeinsam sogar einen Teil eines europäischen Schnellladenetzes, das Sie selbst mit aufgebaut haben.

Dann machen Sie daraus doch endlich ein Produkt.

Nicht:

Auto hier.
Leasing dort.
Wallbox extra.
Ladetarif irgendwo in der App.
IONITY als „Partnernetz“.
Und nach dem Kauf darf der Kunde erst einmal herausfinden, welche Karte, welcher Tarif und welche Grundgebühr für ihn sinnvoll sind.

Sondern:

Das Elektro-Mobilitätspaket.

36 Monate Leasing.
Wallbox inklusive.
Plug & Charge inklusive.
Eine App.
Eine Rechnung.
Und beispielsweise 5.000 kWh Schnellladen bei IONITY während der Leasingdauer inklusive.

Bei einem Verbrauch von 18 kWh auf 100 Kilometer entsprechen 5.000 kWh rund 27.800 Kilometern Fahrenergie.

Und wenn beim Kunden keine feste Wallbox möglich ist, bekommt er eben eine mobile Ladelösung wie NRGkick dazu. Damit fällt auch eines der üblichen Argumente weg, warum Elektromobilität angeblich nur für Menschen mit Eigenheim funktioniert.

Dann schreiben Sie genau das auf das Plakat:

„Ihr neues Elektroauto. Rund 28.000 Kilometer Energie inklusive.“

Darunter:

„Laden im europäischen IONITY-Netz, das wir als Automobilhersteller selbst mit aufgebaut haben.“

Plötzlich ist IONITY kein anonymer Ladeanbieter mehr, sondern ein Vorteil des Autos.

Tesla hat genau diesen Zusammenhang früh verstanden. Der Kunde kauft dort nicht nur ein Fahrzeug und sucht sich anschließend irgendwo Strom. Auto, Navigation, Ladeplanung, Batterie-Vorkonditionierung, Säule und Bezahlung wirken wie ein System.

Die deutschen Hersteller besitzen inzwischen fast alle Bausteine dafür ebenfalls. Sie verkaufen sie nur noch viel zu häufig als einzelne Dienstleistungen.

Dabei könnte die Botschaft so einfach sein:

Ein Auto. Eine Rate. Eine App. Ein Ladezugang. Ein Energiepaket.

Und das Schöne daran: Beim Verbrenner kann kein Hersteller seinem Kunden beim Kauf sagen, dass ein relevanter Teil des Kraftstoffs für die nächsten drei Jahre bereits dabei ist.

Beim Elektroauto können Sie genau das.

Vielleicht sollte man das Elektroauto also nicht nur technisch anders bauen als einen Verbrenner.

Vielleicht sollte man endlich anfangen, es auch anders zu verkaufen.

Die Idee ist übrigens weiterhin kostenlos. Sollten Sie allerdings irgendwann feststellen, dass daraus tatsächlich eine Kampagne wird, dürfen Volkswagen, BMW, Mercedes, Audi oder Porsche mich selbstverständlich finanziell an dieser plötzlichen strategischen Erleuchtung beteiligen.

Ich bin da völlig unkompliziert. Überweisungen nehme ich notfalls auch ohne 87-seitigen Beratervertrag entgegen. 😉

[Zeo @zeo_ex]

A guy started taking robotaxis for his daily commute the week they went live on his street in Las Vegas, right as thousands of them quietly started appearing on roads that had never had a driverless car on them before.

Within the first two weeks he noticed something odd, the kind of thing that's easy to dismiss the first time and only becomes a pattern once you've seen it enough times in a row. Same pickup spot outside his building, same destination downtown, same time of morning almost every single weekday — and the route was never quite the one he'd have taken driving himself, the one he'd memorized down to the turn over years of the same commute. A slightly longer stretch on one street he wouldn't normally use. An extra turn that never seemed strictly necessary given where he was actually headed. Nothing dramatic on any single trip, nothing worth complaining about out loud, just a couple of extra minutes tacked on here and there, trip after trip, day after day, that he quietly chalked up to "the AI's probably just being cautious with a new system" and left completely alone.

His friend, who works in AV fleet operations for a living, happened to be riding with him one morning on the way to breakfast, and watched the route unfold in real time on the in-car map display.

"You know it's not actually taking the fastest route for you specifically, right? It's balancing your trip against a couple of other things it's never once told you about, sitting right there in the background of every ride you've taken."

[Randy Olson @randal_olson]

German rail runs 1,773 trains at once during evening rush and 63 at three in the morning.

The map follows every one through a full day, split into high speed, intercity, regional, and night service.

[Chris Meder @EVCurveFuturist]

Everyone thinks EV adoption ends when people stop buying ICE cars.

It doesn't.

That's when the infrastructure starts disappearing.

In Norway, battery EVs now account for around 99% of new car sales, while roughly 40% of the passenger car fleet is fully electric. The country still has around 1,700 fuel stations, so drivers are not yet broadly struggling to find petrol, but the network is already beginning to contract. Since 2020, 430 petrol stations have closed as fuel demand has continued to fall.

This is the next phase of disruption. Fewer litres sold make marginal stations uneconomic. Closures initially remove low-volume or redundant sites, but as the ICE fleet keeps shrinking, maintaining a dense nationwide fuel network becomes progressively harder to justify.

That is when the transition becomes self-reinforcing. First ICE loses market share. Then it loses infrastructure. Eventually it loses convenience. Longer detours, fewer stations, less competition and higher operating costs all make holding onto an ICE vehicle less attractive.

Norway is not yet at the point where petrol is difficult to find. It is showing us how that point begins.

Norway isn't just electrifying cars. It is quietly dismantling the ecosystem that supported internal combustion for the last century.

This is what an S-curve looks like after the tipping point.

[Linus ✦ Ekenstam @LinusEkenstam]

The future is here, it's just not evenly distributed.

[Jörn Peters @onskylights]

@ChrisO_wiki ,

[Dmitrii Kovanikov @ChShersh]

"We had a good thing, you stupid son of a bitch! We had StackOverflow with incredible answers to all CS, Maths, DB, programming and systems questions, we had everything we needed, and it all ran like clockwork! You could have shut your mouth, accept that your question was closed as a duplicate, and made as much money as you ever needed working in tech! It was perfect! But no! You just had to blow it up! You, and your pride and your ego! You just had to invent AI and LLMs! If you'd done your job, known your place, we'd all be fine right now!"