They didn’t include the most prominent problem with open models: they can be fine-tuned. Especially, they can be fine-tuned to remove the guardrails.
Don’t get me wrong, I don’t see this argument as a real problem since:
a) I don’t especially trust US companies (or Chinese companies, for that matter) to be the final arbiters of morality
b) There isn’t any model on the market that hasn’t been jailbroken
so I see that point as moot. But it’s at least a point that is understandable.
The first country to mass produce AI controlled soldiers will have a massive advantage on the battlefield. One side looses money and lives and the other only money. Supply suddenly doesn’t include food, medicine or water anymore. No issues with moral, fatigue or diseases. Fallen ‘soldiers’ can be recycled and repaired. Scary.
To be fair, they all could have gotten the sword after siring an heir.
What better time is there to go on birth control than after you have the right amount of kids?
How many programmers does it take to screw in a light bulb?
a. None, it’s a hardware problem.
b. None, programmers can’t fit in light bulbs.
c. 100. One to hold the bulb and 99 to debug the house.
d. None so far, but they’ll get to it Real Soon Now.
All answers are correct.
Why do I have the feeling that he will introduce himself as ‘Linux’?
Kenobi could just be a common name, like looking for a John Smith. Also, with multiple planets full of people, there could be thousands of ‘Kenobi,’ even if the name is not common.
Sounds like Inception.
Oh yes, Inception would be a great Muppet film: at every level they go deeper, it switches between real humans and Muppets.
That poor child in the stock photo getting shown for an article on how children are getting dumber…
That being said, the reason why children are getting ‘dumber’ is probably because a) education is getting less money every year b) social media is destroying their attention span c) intelligence isn’t valued enough by society
At no point is getting a notebook part of the problem. Young people need to learn how to use technology.
Then why haven’t we heard about all the criminals (pedophiles, terrorists,…) being caught by Microsoft surveillance? Because they don’t care about the crimes you do. Their tracking is for money making only.
If I were a criminal I would not bank on it - they might use it for blackmail, too - but for the most part they are not in danger.
Let me try: Three capitalists enter a bar. The first one pays for his drink and doesn’t tip, leaving the bartender with less than minimum wage for the transaction.
The second capitalist leaves without paying since his wealth is in stocks and he can’t be asked to pay for anything.
The third capitalist says: “I just bought the bar, fired half the staff, and sold off most of the furniture. Your bar is now very profitable, and I just sold it for double. Try not to declare bankruptcy before I leave - it bums me out.”
Isn’t ChatGPT the sycophantic one? Especially the older models were horrible.
That sounds serious, but I wonder what it is that OpenAI wants from Apple? Currently they are mostly on the software side and Apple’s secrets more on the hardware side.
Is it for the rumored physical AI thing they wanted to build?
No, they are probably just building data centers to replace any non-physical labor with AI. And hope to be able to do that in the future for physical labor with robots: To create a permanent underclass of unemployed people and the 1% living like gods.
Not really better but that’s what it looks like to me.
You would love my backyard. I’m lazy af.
I would be less concerned (still not buying it, but…) if pedophiles we already know about were prosecuted. How about putting them in jail before looking for more ways to protect children? What about the rest of Epsteins pedo ring? Are we supposed to think that was just the work of two people?
I’m all for protecting children, but we could do so much more even without sacrificing privacy.
Have you heard about “Tokenmaxxing”?
Since many AI companies didn’t have a reasonable limit on the number of tokens for the amount of money you paid, companies started telling their employees to use as many tokens as possible. LLMs improve with more tokens (although there are diminishing returns).
So users tried to exploit the ‘lure offer’, and AI companies had to change the billing. It’s still below the real cost, but no longer this insanely expensive option.
Well, if we had risen up against the rich, instead of letting them rule the world at least the last image would have been avoided.
Yeah, sorry. I’m tired and English isn’t my mother tongue.
Oh, the huge manatee!