it’s going to make it possible for people to identify more problems,
He may be right.
For every job taken away by AI, 2 jobs are created to fix the problems caused by AI.
If there was no AI, we wouldn’t be able to identify new problems that were caused by AI!
Yes of course you are right, my bad.
The 1 trillion is an extrapolation.
AFAIK it’s less than 10%. But it’s a start. 🤣🤣🤣
It’s not only a joke, it’s an analogy to show how stupid the claims of politicians are, that they want to have a backdoor for law enforcement.
Of course the analogy isn’t perfect for the reasons you describe, and those reasons makes it actually worse.
Is he going senile since he wants to play in the streets again?
We have an easy peasy solution to that.
We will just make it illegal for burglars to look under the mat.
And if they do look under the mat, we will also make it illegal fro them to take the key.
Finally we will also make it illegal for burglars to use the key.
See there an absolutely bullet proof solution, so why does the tech industry continue to claim this is a bad idea?
As a politician I simply can’t understand why they are so contrary to this idea that will increase safety for everybody!!
/s (just in case)
Exactly the same as when Microsoft ditched Outlook express.
you have to have some awareness of what AI even is or you are at risk of being harmed.
Yes that’s how it is, but that should not be the case, AI should legally be considered like asking expert advice, like asking a lawyer or a doctor, those are not considered risks, because they have legal responsibility for their advice. The same must be the case for AI, AI must have similar legal responsibility covered by the company offering the AI service.
If AI responses can’t be trusted and are false information, it’s not a service but a disservice. It can never be the case that normal users should have particular skills to use an AI service. That’s legally a slippery slope we should absolutely refuse to allow.
I never read any “fake” Asimov stories not actually being written by Asimov himself.
I have however read everything Asimov wrote, I think he was an amazing author, and I love how well thought out his stories are.
Is there any particular reason to read these Roger MacBride Allen stories?
I must admit my interest in reading has diminished, because I find mostly everything I read has become banal and without any new thoughts.
Thanks for the good constructive response. 👍
a company that provides unreviewed AI output must ensure users are aware that output can be harmful or incorrect.
I agree, and then we should require such warnings for every AI response, kind of like we have on cigarettes.
Meaning the responsibility to warn about harmful effects is up to the company offering the service, not for the user to assume.
I think responsibility comes once you present AI output to something who doesn’t consent or understand AI.
Which means 99.9% of the users currently using AI, or unknowingly exposed to it through services that use AI without it being clear.
I wonder if Elon Musk is laughing his ass off about BYD making robots, like he did about their EV cars?
https://www.youtube.com/watch?v=_9ftbRWqkj0
In China BYD is now selling 5 times more EV cars than Tesla!!
https://autovista24.autovistagroup.com/news/what-was-the-best-selling-ev-brand-in-china-in-2025/
Now Musk claims his stupid AI and Robots that aren’t even competitive today, will be worth billions or rather trillions in the future.
My guess is the competition will walk all over him in every field, because Tesla is lagging behind now even on EV where they had a solid head start. Elon Musk does not have a head start on robotics and AI, and I doubt Neither SpaceX (XAI) or Shitter (Grok) or Tesla (Optimus) will ever make a profit on any of their AI products.
Maybe BYD won’t either, but I give them a way better chance than anything Musk related.
The basic idea that there are abiously laws for humans too is fine.
But these are misunderstood, and are designed to put blame on users instead of the AI and the companies behind.
It is well known that Asimov’s laws were flawed.
What? No they aren’t!! Those laws are 70 years old, and are surprisingly well designed, and the concept still stands, which is why they are still so famous! The claim that they are flawed is a very bad start, and very noticeably the author does not describe in what way they are flawed, and also no, Asimov’s stories do NOT reveal flaws in these laws, on the contrary they demonstrate the necessity of them. Asimov’s stories demonstrate that responsibility of AI is a requirement, otherwise innocent people or even humanity will get hurt.
The inverse laws the article suggest:
Humans must not anthropomorphise AI systems.
This is mostly irrelevant. Judgement must be based on the facts, just like with any other type of professional advice.
I feel like 2+2 should be five, is essentially nonsensical, but still it is impossible to not have human users reflect on their own emotions when being advised, whether it’s a human, a book or an AI.
Humans must not blindly trust the output of AI systems.
This should seem obvious. But it completely removes the responsibility from the AI, and fails to account for the human factor that if the AI was right 9 times in a row, it is very human to think it is probably also right the 10th time, exactly as with a human advisor. You can’t make it a requirement of humans to automatically be skeptical of advice put forward with sophisticated language and argumentation that seems identical to a qualified authority and expert. And then require that we make the research manually afterwards. What would the point of the AI be then? We might as well simply skip the AI step altogether by that logic. Which might actually not be such a bad idea. 😋
Humans must remain fully responsible and accountable for consequences arising from the use of AI systems.
This sounds like blame is fully on the user, not the company responsible for the AI.
When services are offered to non specialists and ordinary people, the company behind the “product” must have responsibility of the quality of the product for the services they offer, exactly like with other products. Guidance to a teenager to commit suicide cannot be blamed on the user, and guidance to commit acts of terror makes the AI an accomplice, and the company behind such an AI must be exactly as responsible for the act as if a human had been an accomplice.
This looks like a whitewash of AI responsibility, and this is exactly the last thing we would want to become a legal norm, a world of irresponsible AI where the users carry all the blame.
Guidelines maybe, laws no.
Edit:
I’m surprised this is downvoted, since when did Lemmy become pro irresponsible AI? Laying the burden of responsibility on ordinary users instead of the companies behind!
Absolutely “don’t be stupid” is what we should all strive for, but you simply can’t make it a “law” that people must stop being stupid. An AI giving advice that obviously harm people should absolutely be illegal, exactly as Asimov suggested in his 3 laws of Robotics.
Expected errors are the worst, they don’t even pop up an error message, but just keep fucking things up in the background.
Of course you need a CPU capable of multi threading, which today means any CPU, but then there is no doubt that the multithreaded init process is way faster.
This was thoroughly tested when systemd demonstrated it.
Single threaded init processes have bottlenecks, and a single issue will stall the whole process. Of course systemd only influence boot speed of user space, but the Linux kernel itself is also multithreaded in it’s boot processes today, because it is without a doubt faster.
Yes upstart was relatively short lived 8 years:
https://en.wikipedia.org/wiki/Upstart_(software)
But Ubuntu did put some effort into improving the init process with Upstart, and it was the fastest init system until systemd beat it by being way way better at multi threading.
Ubuntu also massively improved how well Linux worked on laptops, and upstart was part of that effort too.
What is best can be subjective, being the most popular signifies that most people found it to be best.
I have made the measurements, and at 500 Mbit/s I actually got a bit more than 5x what I had at 100 Mbit/s. Actually my 500 Mbit connection ran as 550, because the rated speed here is the guaranteed speed of the connection. So the only limitation is the server at the other end.
It is true however that 1 Gbit/s didn’t quite double the 500 Mbit/s speed, Actual measured facts beat speculation.
But your examples of steeply diminishing returns are not true.
No that’s not true, there is actually competition here and a very transparent market.
30 years ago when 2 Mbit/s was relatively new here, ADSL on existing phone lines had a price of 69,- €. (cheapest provider at the time)
Even without accounting for inflation, the price now is cheaper for 1 Gigabit, despite the old ADSL was based on existing cables! And 1 Gbit obviously is on fibre optic cables made specifically for internet connection.
The cost of establishing fiber networks was expensive, and it is only recently that some of the companies are turning decent profits, and I think most of the profit is on selling TV packs and extra services like cloud storage and virus protection. My internet bill has about 5 points of extra services that all have a nice round zero on them. 😋
And on 500MBit it’s two minutes, so doubling the bandwidth only saves one minute.
This is simply not true, of course it isn’t entirely linear, but for big downloads you actually get pretty close to the full benefit of the speed, when the servers can handle it.
When the speed goes up, latency also goes down, making response times faster too.
Sounds a lot like your Fedora update is single threaded, which is a huge limitation. I start updates manually and monitor the whole process, and the whole process is finished in a couple of minutes for a big update. A single package can be literally less than 5 seconds for download, integrity check and installation. Firefox is among the most frequent single package updates, and that generally takes 5-6 seconds.
Maybe injecting the infections made it look like they were maintained? 😋
First create, then find. 😋