Because you're training a detector on something that is designed to emulate regular languages closest possible, and human speech has so much incredible variability that it's almost impossible to identify if someone or something has been written by an AI.
You can detect maybe your typical generic chat GPT type outputs, but you can characterize a conversation with chat GPT or any of the other much better local models (privacy and control are aspects which make them better) and after doing that you can get radically human seeming outputs that are totally different from anything chat GPT will output.
In short, given a static block of text it's going to be nearly impossible to detect if it's coming from an AI. It's just too difficult to problem, and if you're going to solve it it's going to be immediately obsolete the next time someone fine tunes their own model
. Trump had a history of corruption going back decades
The press shit on trump like no tomorrow. It didn't stick because they'd spent years and years eroding their own legitimacy, not because they didn't air bad things about Trump.
Build a proper router with PF sense. I can do everything that pie hole can do, and more, and you get to have fully custom eternally upgradable router software in general that you can use to do things like network scanning and all sorts of other fun stuff.
I briefly used the raspberry pi for that purpose, but honestly I'm much happier having built a full router.
Proud to be one of them. I tried to disable the job that runs windows update, they said I don't have permission, so I switched to Ubuntu on every single computer except the one that runs VR games.
As a bonus, as an enthusiast for artificial intelligence stuff, more programs run on Linux than they do on Windows
People at the internet archive literally gave away all the books they had in the library for free to as many people who wanted them, basically pretending they had a right to copy the books as many times as they desired as long as it was under the guise of being a library.
Not only did they deserve to lose this case, they displayed such arrogant weaponized stupidity in making that decision that I'm surprised they weren't trying to screw themselves over.
The internet archive is awesome, their decision in 2020 was fucking stupid
Well sure, the entire global warming crisis is a matter of energy. Almost every problem we have today is a matter of energy.
The problem is, at any given point in time a more productive use of energy then carbon sequestration is going to exist, because it is incredibly difficult to pull carbon out of the atmosphere and it would require a project of herculan scales to make a difference in the global climate.
Imagine it's 10 times as hard to carbon out of the atmosphere as it is to put in.
It has taken the entire world economy decades to get to the point that global warming is moving back a couple of degrees.
To offset that with sequestration you're going to need something the size of the entire global economy, and you're going to need to create that while the only possible input is through government programs and sequestration creates next to zero benefit in terms of profit for the people doing it.
It's going to be hilariously difficult, nearly impossible, and you can't wave that away with "it's an energy problem".
It's only ever going to make sense inside of coal smokestacks.
but if we don’t have transportation to get to these places I may as well live in London
Yes. These decisions are exactly what a carbon tax is designed to create. The people who really want to live a recreational non-urban lifestyle that is quite expensive in terms of carbon emissions are going to have to pay more to do it because that reflects the true cost of that lifestyle.
Ideally That extra expense encourages innovation in that space so that the area can become less carbon intense. Perhaps local authorities are encouraged to build trains or electric cars or some other system that lets you live your life without emitting carbon in those areas.
It's hilariously expensive and it's expensive because physics. We measure carbon in the atmosphere in parts per million. The entire surface area of the planet is already littered with Caron absorbers and they don't make a dent.
Since we have no genuinely reliable public transportation infrastructure, we all still drive. And we pay the carbon tax
A carbon tax is designed to be placed on the sources of oil and then ramped up over time. If you haven't felt the carbon tax it's because the tax isn't high enough yet.
The only problem with the carbon tax is that it's going to be difficult to get people to support it. Otherwise it's the single most effective way to actually produce change across the whole economy.
Except carbon sequestration is not ever going to work and it's always going to be more expensive than having just burned that fuel in the first place.
Maybe you'll get an advantage if it's nearly free to do and you use exclusively solar power in areas with excesses of it.
But on average? Sequestration is not an answer. The carbon dioxide in our atmosphere is just too rare to effectively pull out, and it's never going to be capable of even reaching fractions of what we're emitting right now.
We have one answer to this problem and one answer only.
Stop. Using. Fossil. Fuels.
Tax carbon.
Start getting ready to do geoengineering, because we are going to need it.
People like to bitch and say that we shouldn't be changing the environment, but guess what, we're changing the environment if we like it or not, it's only a question of it it's in our interests or if it's an uncontrolled self-destructive form.
Because you're training a detector on something that is designed to emulate regular languages closest possible, and human speech has so much incredible variability that it's almost impossible to identify if someone or something has been written by an AI.
You can detect maybe your typical generic chat GPT type outputs, but you can characterize a conversation with chat GPT or any of the other much better local models (privacy and control are aspects which make them better) and after doing that you can get radically human seeming outputs that are totally different from anything chat GPT will output.
In short, given a static block of text it's going to be nearly impossible to detect if it's coming from an AI. It's just too difficult to problem, and if you're going to solve it it's going to be immediately obsolete the next time someone fine tunes their own model