![](/static/66c60d9f/assets/icons/icon-96x96.png)
![](https://fry.gs/pictrs/image/c6832070-8625-4688-b9e5-5d519541e092.png)
People seem to hold computers to a higher standard than other people when performing the same task.
People seem to hold computers to a higher standard than other people when performing the same task.
Their problem:
So apparently NetHack has a mechanic that slightly changes how the game plays every time it’s full moon according to your system clock
The model wasn’t trained on a full moon. They had a system to set up the environment for replicable results but it didn’t include modifying the system time.
It reminds me of another bug with the system time, which a friend of mine encountered. He was working on hardware and he was getting a lot of units that worked fine at the factory, immediately failed at the client’s location, and then worked again when they were returned to the factory. It turned out that when these machines were turned on, their embedded OS automatically queried some server to update the current time. The client’s internet connection had such high latency that the server’s response only came back after the machine was already in use. This generated a huge delta-t value that triggered the sanity checks and shut the machine down. The factory had a much lower-latency connection and so the race condition could never be replicated there.
As for the weirdest bug I ever encountered myself: a compiler generating bad machine code. I have often said that the worst part of programming is that the computer always does exactly what you tell it to, but that was the one and only time in twenty years that the computer actually didn’t.
Before anyone gets too excited: some of their electrodes are no longer able to record a signal from the patient’s brain. They’re reprogramming their software to work with fewer electrodes. No one is being turned into a borg drone.
I suspect it isn’t even illegal, but I’m not an expert.
You can state what you don’t want, but no one will be paying attention. Except maybe the LLM reading your posts…
There’s not going to be a moment when the world suddenly goes from having oil to having no oil. Some oil reserves are relatively cheap and easy to extract. Other, very large reserves are currently so difficult and expensive to extract that doing so isn’t profitable. As the easy oil gradually runs out, the supply drops, the price rises, and sources of oil that were not profitable at the old price become profitable. This maintains the supply of oil and stabilizes the price.
Eventually oil will become so expensive that alternative technologies will be cheaper than it. This will happen with plenty of hard-to-reach oil left. So it’s true that the amount of oil is in principle finite, but that limitation isn’t really relevant.
One bad quarter and they’re doing this? I don’t even.
My experience with the healthcare system, and especially hospitals, is that the people working there are generally knowledgeable and want to help patients, but they are also very busy and often sleep-deprived. A human may be better at medicine than an AI, but an AI that can devote attention to you is better than a human that can’t.
(The fact that the healthcare system we have is somehow simultaneously very expensive, bad for medical professionals, and bad for patients is a separate issue…)
Note that you can turn the ads off quickly and easily. I agree that there’s someone off-putting about an operating system with built-in ads, but a tech-savvy person will see them once and then never again. (A person who isn’t tech-savvy probably won’t care.)
It’s a use case, but I would argue that it’s not a significant use case.
Emulators are still legal in theory, but I doubt that it is in practice possible to make an emulator for a modern video game system without violating some other part of the law.
Does it matter? I suspect that if that’s what you did, you were one of very few people doing so, and the law doesn’t require the absence of any possible legitimate use. In this case, something is illegal if it
is primarily designed or produced for the purpose of circumventing a technological measure that effectively controls access to a work protected under this title;
has only limited commercially significant purpose or use
How would they fight it if they had the money? Did they have a significant use case other than piracy?
Honestly, what’s the big deal? Your face is not secret and anyone who feels like it can photograph you while you’re out in public. Vending machines already know who you are if you use a credit card.
However, this is a good reminder to programmers: customers might sometimes see your error messages even if you didn’t intend them to. Don’t write anything Marketing wouldn’t like.
The alternative to military AI is not peace, it’s war the old-fashioned way. Humans are bad at distinguishing civilians from enemy fighters; artillery shells can’t do it at all. I anticipate that AI will make mistakes, but fewer mistakes than would have been made otherwise.
I liked the name Bard… Gemini is just random and unrelated to the concept.
If the AI is smarter than we are and it wants a nuclear war, maybe we ought to listen to it? We shouldn’t let our pride get in the way.
This time is different. If AI were to remain what it is today, the article would be correct, but AI won’t. It’s a fundamentally new kind of technology, unlike anything else that has ever been created by humans. It only seems like more of the same to some people because it’s so very new and primitive compared to what it will be soon. This won’t be humans losing their jobs, this will be humanity losing its job. There will be plenty of new industries created but they will be run by AI for AI.
With that said, it won’t necessarily be bad. It’s the end of the world as we know it, and I feel fine.
No, but I’m not paid hourly - if I sit around idle, I’m generally just wasting my own time.
From the article:
Both the “idle time”, which indicates a period of scanner downtime of ten minutes or more, and the “latency under ten minutes”, which tracks scanner interruptions between one and ten minutes are deemed illegal by the CNIL when it comes to data processing. The CNIL is using the GDPR as the legal basis of the case.
Amazon has also implemented a “stow machine gun” indicator to prevent mistakes. It signals an error if you scan an item less than 1.25 seconds after scanning the previous item. It sounds like a way to prevent double-scanning mistakes. But that’s a GDPR issue too, according to the CNIL.
I think these all seem like entirely reasonable things for Amazon to track.
When I bought my Windows 11 laptop a month ago, I was able to set up a local account after turning on airplane mode. (I had entered my wifi password in an earlier step since I thought it was just for installing updates.)