Ehhh maybe a year or two ago. Modern systems are far more capable than you’re giving them credit for.
I’m not insinuating that they are perfect or even human level, because they’re not, but both of those complaints are things that AI can usually handle summarily.
This is one of those areas where modern AI might actually come in handy. Let it go through every bodycam for every incident and write the report without the bias of the officer on duty. And it could go through the archive of all footage to help build better statistics than this self-reported data.
Yeah, it’d have to be an openly auditable model, obviously.
It absolutely doesn’t. The only model that has “gone nuts” is Grok, and that’s because of malicious code pushed specifically for the purpose of spreading propaganda.
Unpopular opinion, but some sort of anti-ICBM satellite network would be a fantastic thing for the world… as long as it’s jointly operated by a coalition of nations. It’s a problem when schoolyard bully USA is the only one who can repel a nuclear counterstrike.
Yep, my small business will not be fulfilling from the US any more. We will be operating out of the home country of the other founder. And I’m not even a big company who has to do real optimization for cost.
Trump’s policies are losing a lot of money for the US, and it will only get worse as companies that are slower to pivot get around to leaving.
Not DDOS, DOS. You can often crash an unprepared server with one request by telling it to hash more data than it has memory for. See this blog post for a well-known web framework. Let’s say I just sent it a 10GB password, it still has to process that data whether or not the hash eventually shortens to the database field length.
Well, AI in general does do better than Republicans at being compassionate to all humans. At least it does until instructed to do otherwise, at which point it becomes the perfect vehicle for oppression.
Ehhh maybe a year or two ago. Modern systems are far more capable than you’re giving them credit for.
I’m not insinuating that they are perfect or even human level, because they’re not, but both of those complaints are things that AI can usually handle summarily.