Anders Fjeldstad (@Hihaj) profile picture
Stockholm, Sweden
Joined April 2009
Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)Photo by Anders Fjeldstad (@Hihaj)
Anders Fjeldstad retweeted
Thermonuclear take, better models make it harder to have good code. Opus5.5 is phenomenal, it’s the best model out there. But it still does odd things, still duplicates a lot, still makes mistakes. The biggest problem, imo, is that because let’s say 75% of what it does is *fine*, it’s a lot harder to stay mentally engaged when reviewing it. If there’s a problem every few lines or major obvious flaws, then it’s a lot easier to focus because you’re thinking about how to do it better. Now that the PRs are mostly fine, I’m constantly fighting the urge to just merge without major revision, especially for “less important” bits of the codebase. And it’s not quite good enough to do that safely.
13
1
38
8,874
Anders Fjeldstad retweeted
I wanted to prove FSD is safer than human drivers. Genuinely. I went to the data expecting to find the number everyone quotes: one accident per six million miles on FSD versus one per 670,000 for humans. Nine times safer. Case closed. The claim cannot be made. Not "the data is weak." The comparison itself does not exist. Start with the equipment. The "human driver" baseline is every car on American roads, and the average American car is over twelve years old. That baseline includes bald tires, worn brake pads, cracked mirrors, and drivers with no emergency braking, no blind spot warning, no lane keeping. A modern Tesla is a rolling sensor suite with the best crash-avoidance hardware ever sold to consumers. So I went looking for the honest comparison: FSD miles versus miles in comparable modern EVs with the same braking, comparable sensors, comparable warning systems. That dataset does not exist in any usable form. Nobody publishes it. Tesla certainly doesn't. Then the miles themselves. FSD engagement is self-selected. Drivers turn it on in good weather, in daylight, on roads they trust it with. The owner's manual itself warns the system "may not operate as intended" in rain, snow, fog, and low light, which happen to be the conditions where humans crash most. So the FSD number is computed over the easiest miles in the newest cars with the richest, most insured drivers, while the human number absorbs every drunk at 3am and every teenager in a beater. And it's supervised. Every mile in the FSD column had a human babysitter whose interventions, whose saves, get credited to the machine. The comparison isn't machine versus human. It's machine-plus-human versus human alone, with all credit assigned to the machine. Darrell Huff described this exact move in How to Lie with Statistics in 1954: if you can't prove what you want to prove, demonstrate something else and pretend they are the same thing. Rolf Dobelli in "Art of Thinking Clearly" calls the other half of it the swimmer's body illusion: swimmers have great bodies because bodies select swimmers. Teslas crash less because new luxury cars with safety suites driven by affluent middle-aged owners crash less. Selection, not software. When Forbes decomposed Tesla's own 2020 report by road type, the Autopilot advantage shrank to about four percent, inside the margin of error. Tesla has refused every request since to release the breakdown. So here is where I landed. I wanted the claim to be true and provable. It is neither provable nor disprovable with the data that exists. The honest sentence is "we don't know." Anyone who tells you FSD is statistically safer than humans is guessing. Anyone who tells you it's statistically more dangerous is also guessing. The difference is that only one of those guesses is attached to a stock price, a monthly subscription, and a man telling you the feature will be finished next year, every year, since 2016. When someone quotes you the nine-times-safer figure, ask one question: compared to whom, driving what, in what weather, measured by whom? The silence is the answer.
Name one accident caused by Tesla FSD in Europe. I’ll wait.
93
262
24
2,168
94,167
Anders Fjeldstad retweeted
past week ships: - hex, great way to talk to agent - browse, great way for agents to use your browser - rune, great way to build native apps - deepseek $100 for $10 - new opencode website unfortunately wrote zero essays about what a software factory is so they will be punished
122
65
16
3,137
111,183
Anders Fjeldstad retweeted
I totally disagree with this. If you just want to build some automation tools, some simple stuff, closed software for 10 users - yes, you can ignore that you don't know those languages and just do all with agents without reading the code. But if we are talking about serious software, when the money is on the table, when people's lives are on the table (for example, for medicine software), this is where you absolutely need to have a deep understanding of your main programming stack. You are responsible for the results, not your AI agents for making bugs and mistakes. Believe me, business doesn't care if your agent is responsible for the outage in the middle of the night, losing hundreds of dollars per the website being offline. You, as a software engineer, are responsible for the result. You're not paid to write the code, and you're not paid for using AI agents. You are paid to solve business problems, and writing code manually, or using AI agents is just your tools, so use them responsibly. That's why I recommend everyone to have deep programming knowledge in your main programing stack, in system design, architecture and so you can clearly understand when your AI agents do something wrong, and it will be more than enough to maintain production.
I used to be a language lawyer. In the late 80s I _studied_ the ARM (Annotated Reference Manual for C++). I knew the language cold, inside and out. I could make it do backflips. I shipped a lot of code using it. A LOT. I’ve attacked other languages with a similar intensity and depth. Until recently, that was simply the only way to become adept at building software systems. I now consider the deep understanding of a language to be a supreme waste of time and effort. I don’t want to learn Rust, or Golang, or Typescript, or any of the thousands of other languages out there. I’ll leave that to my agents. What I need now is to understand the high level gestalt of the system. I want to see the systems and subsystems. I don’t need to inspect the cogs. If I truly need to understand a detail at the cog level, I’ll interrogate my agents. I’m not going to pretend to be a language expert anymore.
45
29
5
395
18,770
Anders Fjeldstad retweeted
Samtidigt som hans kandidater gick mot opinionens blodbad i fästen som Texas och Kansas höll han i måndags, dagen innan mellanårsvalet, en timslång Kurt Olsson-lik presskonferens om algblomning.
1
1
10
763
Anders Fjeldstad retweeted
Sam Altman thinks RAM sitting in personal computers is a waste when it could be put to better use in data centers. Weird. I was just thinking that all those billions sitting in billionaires’ bank accounts could be put to much better use in our communities.
442
13,287
218
91,040
1,025,371
Anders Fjeldstad retweeted
I get so annoyed by posts like this. People get sick. People get pregnant. Life happens. Unexpected things happen. And what exactly is the alternative? That people should just end up on the street because life didn’t go according to plan? It’s genuinely wild to me that this is even a debate. You can say whatever you want about Europe, and people certainly love to talk about our taxes. But you know what? Our schools are free. Healthcare is free or affordable. Universities are free or heavily subsidized. Childcare is affordable. And if you lose your job or get sick, there’s a system in place to help you get back on your feet. That’s the whole point. We have a social safety net because we understand that life doesn’t always go as planned. And I genuinely believe that having that protection should be a basic human right, not a privilege. Now, if you want to talk about how Europe can make it easier for startups to grow, attract investment, or build successful companies, I’m all for that conversation. There’s plenty we can improve. But please, let’s stop acting like our social welfare system is the problem. It’s one of the things we should be most proud of.
European hiring rules are genuinely insane. For example - if you own a company in the Netherlands and one of your employees gets sick, you are required to pay at least 70% of their wages for up to 2 years. If they can't go back to work after 2 years and you haven't done enough for their reintegration, you pay their wages for 1 more year. And you can't fire them until after that extra year!
504
128
42
2,003
204,724
Anders Fjeldstad retweeted
This enormous tweet requires context. Palmer Luckey is the CEO of Anduril, which makes drone tech for the military, and presumably, the sorts of drones that would hunt down enemy soldiers hiding in a forest, such as how FPV drones are used in Ukraine right now.
Replying to @Alacritic_Super
Super cool, this would be very useful for exercise
9
19
3
205
48,838
Anders Fjeldstad retweeted
I'm definitely not anti-achievement and never criticized real achievements, including by Elon. I even own two Teslas. I do expose marketing stunts that, while cool on the surface, are not the achievements they are claimed to be. Compressed air thrusters might give you acceleration boosts in any direction. They might even make your car hover for a couple of seconds if you accept replacing the passenger seat by a 400 liter high-pressure air tank. But they are not a path to actual flying cars. The same way Burt Rutan's SpaceshipOne was never going to be a path to orbital flights (just bunny hops out of the atmosphere). I am anti-Musk when it comes to politics. His political actions, the massive dissemination of disinformation through X and other channels, and his support for authoritarian regimes and racist ideas have the effect of weakening democracy and weakening human decency and social progress around the world.
289
837
70
7,971
650,521
Anders Fjeldstad retweeted
This gets difficult. 🇺🇸 threatens action against Europe if it restricts Starlink, but at the same time Starlink restricts use for the defence of a European country 🇺🇦. Europe needs sovereignty - not submission.
175
927
26
5,114
87,351
Anders Fjeldstad retweeted
Guy who said we would have coast to coast self driving cars in 2019 and 1 million robotaxis in 2020 says advertisers aren't allowed to overstate claims anymore. 😂😂😂😂😂🤣🤣 LOL 😆 😂 🤣 😅 😄
Going forward, all advertising on 𝕏 will be subject to a reasonable esthetic standard. Advertisers who post ads that are ugly, shrill, strident or overstate claims about their product will be barred from advertising.
225
420
15
6,367
198,157
Anders Fjeldstad retweeted
A self-driving Tesla crashed into a California highway patrol vehicle, seriously injuring two officers inside. According to the officers, the driver said he was asleep at the time and using the car’s driver-assistance technology.
2,813
3,325
1,211
8,106
1,881,357
Anders Fjeldstad retweeted
I can't emphasize this enough. Sure, the driver is responsible, and I hope he faces the consequences, but Tesla shares some of the blame. FSD gets more dangerous as it improves because of complacency. Type 'Tesla Full Self-Driving' on X or Insta and you'll see hundreds, if not thousands, of Tesla owners acting as if this car drives itself. I wonder why? Maybe the name? Maybe Elon Musk falsely claiming that you can text and drive now? Sure, FSD is truly impressive and works roughly 98-99% of the time, but as anyone who understands autonomy, 99% is nowhere near good enough. It needs to be 99.99999%. You need to be attentive and ready to take control for that last 1%. That gets exponentially harder as the percentage drops. When it worked great 80% of the time, you were highly motivated to stay attentive for that remaining 20%. As it gets better and requires fewer disengagements, you get more complacent, and the dumber you are- and oh boy, there are dumb people out there- the more likely you are to start believing that you don't need to pay attention because it feels like you don't have to 99% of the time. Tesla is doing a terrible job of addressing this issue.
38
27
129
6,879
Köp inte en Tesla. Använd inte Starlink. Prenumerera inte på X. Använd inte Grok. Använd inte Cursor.
And so the slaughter continues … Time for Zelensky to go
43
Anders Fjeldstad retweeted
Musk is pulling increasingly ridiculous numbers out of his ass, and no one ever questions him. $SPCX at a $100 trillion market cap is 85% of global GDP of $118 trillion in 2025. 🤡
45
47
8
292
12,371
Anders Fjeldstad retweeted
What about the guy who started the slaughter?
And so the slaughter continues … Time for Zelensky to go
6,294
3,536
250
46,402
2,108,729
Anders Fjeldstad retweeted
Elon Musk destroyed the career of FEMA CFO Mary Comans. In Feb 2025, he posted her photo to his followers and called her work "a criminal action." DHS fired her the same day. A federal court has now formally cleared her name. The judge found Musk's post was "entirely false," that he knew it was false, and that it amounted to "state sanctioned intimidation." And the payments he accused her of? They were approved by DOGE, Musk's own subordinates. Here's the evidence. Let's see if Elon is man enough to apologize.
227
2,024
112
5,530
139,609