What a terrible headline. What do you mean usually? Implying there are instances where matter and antimatter come in contact with each other and don’t annihilate? That makes no sense at all.

Also it doesn’t vanish at all, nothing inside our universe can ever be created or removed, it’s a zero sum game. It can only go from one state to another. In case of matter and antimatter, it gets converted into energy.

I always thought Roko’s Basilisk was a way of debunking religion. Like you can very easily invent a story that would steer if not totally govern your actions and motivations today, for something that might exist somewhere in the future. Like in some religions where you are declared to have an immortal soul and this soul will suffer for eternity if you don’t follow rules written by creepy old dudes.

I don’t think this is a very good argument. Because AI is all encompassing in scope, where the things you mention are not.

Sure most people would just use a library to handle TLS and not worry about it. But the library is there, it was written by humans that understand it, people created it for you to use. And if you feel like you can go in and explore it yourself and many people do so. For people brushing up on their assembly skills: Yes many people do that and I would recommend it a lot. For work it’s a more rare thing, unless you are in highly specialized fields, but for hobby programming assembly is used a lot. People love making romhacks or even whole new games for old systems. Z80 assembly is still very much alive and kicking due to the GameBoy. Just like doing recreational mathematics, people do recreational assembly programming.

But where your argument falls down the most is because of transferable skills. Whenever you do any kind of programming you are building up your skills. Your expertise might be NodeJS for example, but if you were to try some C programming, it would take some getting used to but you could easily do it. Because a lot of the skills are transferable.

With AI you get the issue OP described, actual programming skills atrophy. It’s like muscle that needs to be worked. Specializing in a field and simply using a library for another field still works that muscle. With AI you stop working the muscle at all and lose the skills. So presenting AI as just another level of abstraction or automation, like a library or just another tool like a compiler is a false comparison.

On the other hand, I do feel like I connect to Mr. Bucket on a spiritual level.

I was thinking we throw an arcade cabinet at him?

Oneweb isn’t comparable to Starlink. You need a huge dish with a very powerful radio. A typical installation runs around $10000+ and can’t be very mobile. It works well enough for larger boats, but not for anything that moves fast. Installations are large and power hungry, so powering them off a small battery like they do with Starlink wouldn’t be possible with Oneweb. The setup for those installations is also a hassle, not something a regular person could do in a few minutes without specialized technical knowledge. The price for usage is also much higher as compared to Starlink, around 10 times more expensive.

There’s also the issue Oneweb isn’t exactly EU. Eutelsat is French, so that’s fine, but the UK has a large share and India does as well. I’m not sure dealing with 2 extra governments in an already complicated project would be a good thing.

However Eutelsat is being used in Ukraine for a secondary internet source for fixed installations funded by Germany, so it isn’t like Eutelsat isn’t being used, but it is no where near a real alternative to Starlink.

I got excited about this when it was announced some time ago, so I dove into the project. It seems very much not useful and mostly a PR stunt at this point.

These mobile operators are competitors and are often actively working against each other, needing regulation to step in and make them behave. It might be different on this project, but their track record in the past has been not very good.

Then there’s the project itself, they do not intend to build their own sats, let alone launch them. And realistically there is no European launch provider that would be able to deliver even a smallish constellation of 300 sats for a low enough price. Their plan is to buy either a license or a part of a sat from some other provider. It isn’t clear as of yet which, but it looks to be AST Space Mobile, an American company. Chinese providers would be an option, but that would be a step back from US providers, so that’s unlikely. Russian providers are a whole thing I’m not going to touch with a ten feet pole.

AST Space Mobile offers a bent pipe construction where you pay to use a part of the sat, where the antenna and baseband is exclusive for your own use. Groundstations would be fully under EU control (at least within the EU) and in theory the satellite as well. But in reality the provider could still do some stuff to mess with the service, including stopping or degrading it. The exact issue we have with Starlink.

Then there’s the service itself. Starlink is so far beyond other networks, it isn’t a fair comparison. Literally, as Starlink has sats that operate at a extremely low orbit (within the upper parts of the atmosphere) they can offer low latency. With their large number of sats they offer a lot of coverage and stable service. With a large amount of sats and not that many users, the bandwidth is pretty good as well. This of course is because of the huge amounts of VC investments into Starlink and with the big downside of having to launch new sats all the time, as they de-orbit within a few years. So it makes a lot of sense for other providers to have a fewer number of sats in a higher orbit, because they can continue operating for 20-30 years. This means lower upfront costs and more time to recoup that money, but much worse service. High latency, less coverage, less stable and less bandwidth. This has meant any kind of satellite data was very expensive with relatively poor performance. Starlink is the first to offer good performance at a good price. Not sure if they can get it economically sustainable, and launching and de-orbiting all those sats might be destroying the planet. So it’s not a given, but from the customer point of view it’s much better than any other sat internet service.

There’s also the issue of capacity, I don’t think there is a constellation out there that could easily handle the extra capacity needed if we want to kick out Starlink. And that means enough capacity offering at least a somewhat viable replacement in a frequency band we can license throughout the EU. And preferably something that doesn’t require extremely expensive equipment for the customer. Some kind of huge dish with a very powerful radio is a pain in the neck.

So that would mean paying someone to either build out an existing constellation or get a whole new one going. This is going to be super expensive, on the level of 10-50x more expensive per month per customer compared to Starlink. Obviously for military purposes and Ukraine they wouldn’t need to pay, all of the EU tax payers would gladly foot that bill. But it does mean there probably won’t be as many private customers signing on (at those prices probably zero consumers and only few businesses). If we want this new service to actually replace Starlink and provide internet for people who don’t have access to fiber or reliable 5G, the price would need to be in line. Especially if the service sucks as well.

Unfortunately I found out all this and got a bit sad. It would be awesome if we could kick out all of the American services and do our own thing. But for the immediate future it doesn’t look likely. And projects like this often take 10+ years to get anything going. I pray to his noodlyness the war doesn’t go on for that long and there won’t be more wars in Europe by then. But that means Ukraine is stuck with Starlink and all of the Musk Trump bullshit.

I don’t think it’s much of a surprise when we look at America that they don’t value human life all that much.

The whole productivity boost is such a non-argument. A lot of that code was written by people who don’t know what they are doing, so without the tool they wouldn’t have written any code at all. For the people that do know what they are doing, it helps them shit out code faster. But is that actually a gain in productivity? It would only be the case if the quality remains the same, which in most cases it doesn’t. And if the sloperator took the time to read and understand the code, even review it perhaps, the time gains would evaporate. But we all know good sloperators don’t read or think, they ship! And then there’s the additional workload for the people that have to actually review all of that code the AI pooped out. Especially with complex code this can be very hard and with the increased amount of code submitted, their workload has grown. If we take that extra (harder) work into account, AI is actually making us as a whole less productive.

If the gains were as significant as the slop glazers claim, the results would speak for themselves. In reality we aren’t sure, people THINK they are faster, they might ship shit out faster, but at expense of quality or extra work for others. Software as a whole would be transformed and we would all be in awe, in reality we see a bunch of shitty slop projects on Github, shit out and forgotten within a month. We see more issues than ever before, users are complaining about how bad everything is. Best case scenario you get something that kinda sorta works, good enough, while looking and feeling like icky slop.

It’s been years now, where are the results?

As someone with a website, the AI scraper bot traffic is off the fucking scale. It’s like being under constant DDOS attacks. It isn’t like normal search crawlers that rate limit, do careful indexing and respect stuff like rel=“nofollow” and robots.txt. These AI scrapers just spam the shit out of anything and everything, hundreds of requests per bot. You block one IP, it seamlessly moves over to another IP. You do user agent detection, they just pretend to be a regular browser, totally fucking up all your stats while they are at it.

Having some form of bot protection like Cloudflare or Anubis is almost required these days. It’s like AI was designed specifically to fuck shit up, there is no reason they can’t just crawl the internet like a normal crawler. But they have to be assholes about it. Bad enough I won’t get anyone actually visiting my site, since the AI overview thing blocks that, but do they have to go the extra step to kill it altogether?

And don’t forget, an animal is much less likely to eat a whole bottle with the cap attached than they are eating a bottle cap by mistake.

It’s just a win all around and the people who get angry about them are just selfish assholes.

Sure in principle you could check all the output the AI shits out to see if it’s right and then spend time correcting it either by hand or by prompting some more. And in doing so, one might be able to do more in a given amount of time, or less. Recent studies have shown it might have felt like programmers were able to do more, but in reality did less in the same time. But that might vary from person to person and the given tasks.

However that skips over other issues:

Reading and understanding code is so much harder than writing code. This has been known for decades. So instead of doing the easier task of writing the code yourself, you are left with reviewing and understanding the generated code. This add a lot of extra cognitive load, which can lead to fatigue quickly. This fatigue feeds into another issue, where it’s so much easier to just hit approve on everything and skip out on the really in depth stuff. People tend to slack and just glance at the code, which is a huge trap, because LLMs are very good at writing code that looks right, regardless of if it is right.

Then there’s the issue where writing the code is a small part of the task usually. There’s also understanding the task and figuring out edge cases. It’s hard to do that all up-front and at least for me, while writing the code I think of edge cases or get a deeper understanding of the task. By flying through the writing part, the deeper understanding gets skipped as well.

And the issue of bias, where it’s often a lot better to think about an open question instead of a multiple choice question. With vibe coding it’s very easy to be lead down a path by the LLM, where it suggests a path which seems logical, but might actually be wrong. Then breaking out of that path can be hard. This might be an actual wrong path, or it might just stifle creativity. A common complaint is all AI generated stuff “feels” very samey, which makes sense because LLMs are designed to see common patterns in a lot of the data. The more common a pattern is, the more likely the LLM is to suggest it.

Another very common issue with LLMs is they tend to touch stuff they shouldn’t. This is because of how they work is by processing everything as a stream of text. So let’s say I want to expand a template with a new part, it has no choice but to read the entire template and then output the entire template as well. It’s supposed to just add the new part, but often it makes other changes as well. And this happens on a multiple-file base as well.

One big issue I’ve noticed is the regression of skills for people who use AI for their work (coding or other). It’s so tempting to just press the big green EASY button on just about any task. But then when the AI isn’t able to do the task right, people are stumped and are no longer able to do the things they used to be able to do for decades before. Especially with coding, it isn’t like riding a bicycle where once you know how, you are able to do it forever. It’s a lot more like a muscle you need to keep excising and if you don’t it’s going to lose power and capabilities.

And we’re not even touching other issues with LLMs in general, like concentration of power, vendor lock-in, obfuscated pricing, resource usage, (un)fair use of source data, hallucinations, manipulation, destruction of the internet etc.

So for me the people who say: “It’s just a new tool, you wouldn’t go back to hand punching paper tape right!??!!?” are being very disingenuous about the actual issues. Sure it’s just a new tool, but one with real consequences to using it. As with every tool we need to weigh the pros and cons and just because it’s available, doesn’t mean we have to automatically use it. We still see people enjoying a command line or text-based interface, even though most people use graphical interfaces. And I personally code without any form of autocomplete, because I find it annoying and distracting. Just because tools exist, doesn’t mean we should all use them and it certainly doesn’t mean we should be blind to the downsides attached to certain tools.

So the lawyers are using AI to shit out nonsense instead of doing their jobs. They aren’t lowering prices, they want those profits AI has promised and they now have an extra AI bill to pay each month. However since the results are shit, people need to use more tools to either catch or fix the mistakes the previous tool made. So in the end we all pay more, because of the extra tool costs and the quality still suffers.

Yay AI!

For the downvoters: This kind of AI is nothing like generative AI / LLMs. This is the kind of AI that’s actually been used for decades and has proven very good at handling large amounts of data. Machine learning excels at ingesting huge amounts of data and doing stuff like pattern analysis and predictions. The advancements in hardware have meant larger models, that can handle more data with more precision and weather prediction is one of the places where these advancements have shown good results.

So don’t just downvote because AI bad. I fucking hate genAI and what the handful of AI companies are trying to push onto people. But machine learning in data analysis is actual science and a good thing.

Also: Weather systems are chaotic by nature and by that I mean chaotic like in Chaos theory (https://en.wikipedia.org/wiki/Chaos_theory) often misunderstood and misquoted. But it means very small variations can have huge effects. That’s why the prediction van be 90% confident and still be absolutely wrong. That’s inherent in how these systems work and cannot be fixed. So better weather prediction does not mean perfect weather prediction.

Because he told the idiots he would make everything better. But in Trump language that means better for himself and his cronies, not the idiots that voted for him.

People also seem to be under the impression that a steam engine is a simple machine, because it is old tech. But I like to watch people on YouTube who restore and repair old steam engines and those things are complicated AF. So much engineering goes into the design, materials and function. Sure the basic principle is you put in water and heat and use the steam to do work. But that’s so far from reality it’s like saying an internal combustion engine is adding ignition to fuel and using that to do work.

I do think the American story of the Wright Brothers single handily pushing humanity into the era of flight while the whole world called them idiots and didn’t believe is kind of a stretch. There were at that time many people and teams working on the same general concept.

The idea of a flying wing was well understood, as people had been jumping from high places with them for a while. Some dude in Germany had a private hill constructed, just so he could play with his flying wings. The concept of propellers was also known. So people generally understood if one could build something light enough with enough power, powered flight would in principle be possible. The general public might have been skeptic, but all over the world there was money being invested and people working on it.

Doesn’t mean what the Wright Brothers did wasn’t brave and impressive, because it was. Building a rig with enough lift, while being light enough and sturdy enough not to fall apart. Building their own engine to shave down the weight. Figuring out some sort of control scheme and then trying it out without getting killed. One gust of wind and one of the brothers might have been dead, the project canceled. Some say Percy Pilcher could have flown before the Wright Brothers, but he died in a glider accident before his plane was ready to fly.

But in the end if those two could do it, others would be able as well. If they weren’t the first I would think others would have gotten there within the year. The technology was on the cusp of it being possible and as people had been trying for hundreds of years, it was bound to happen.

Thank you, fixed.

No not at all. That’s just a fantasy he likes to perpetuate. He was actually fired and kicked out of PayPal because the other dude hated his guts, when he was away on his honeymoon, but he still had shares so he still made a lot of money from it. That other dude was our other famous Nazi, none other than Peter Thiel.

Ugh another one of these slop projects that will be gone and forgotten in 2 weeks time.

The issue is if you don’t put any effort into it, why would we put any effort into trying it out? And if you don’t put any time into it, what sort of chances are there you won’t put any time in going forward?

It’s really fun to shit something out using AI, give it a fancy name, a fancy logo and a website. But then the fun wears off fast as you need to actually support, develop and improve the app. You need to actually fix the bugs and as soon as you do, you’ll quickly see the limits to what the AI can do. The whole thing quickly becomes a mess that nobody can maintain. And that’s where it’s not as fun anymore and you move on to something else.

This isn’t a critique of you personally, but a pattern we’ve seen thousands of times in the past few years since AI coding went mainstream. So many slop projects that get released and shared around on social media, only to die within weeks. We’ve learnt by now it’s a waste of time and aren’t interested.