Agreed.

What if he uses a tool to do that? Where is the line? wc to check if you satisfied the word count requirement? A spell-checker? An AI?

Assuming that the actual grading is still done by a human and AI is just one of several tools used in the process?

Don't understand the hate. This is actually AI being used in the right way. As an assistant. Not to replace a human, but to help with the repetitive ordinary tasks that are part of the job.

My own experience is similar. When I ask AI to generate some text for a purpose, the result is meh. But as a text critic or to get suggestions for improvements, as a proof reader, it's pretty good.

What should happen is that we don't take an AI output and just use it as-is, but use it as an input for a human who does the actual job. AI isn't magic, it's just a tool. Nobody complains that a lever enables us to excerpt more force than our muscles alone could.

That's probably a long way off, too. Right now, robotaxis are only allowed on a limited range of public roads, which doesn't include highways (and by extension, bridges and tunnels). So while they might be useful to get you from your home in a city neighborhood to downtown, but not much more than that.

(Also, I doubt anybody's really going to pay for robotaxis for a daily commute. Most people buy cars or take the train for that.)

"Going Sleeper"? I don't get it. If you were laid off, why wouldn't you say so? You're under no obligation to a retailer you no longer work for, least of all to use some kind of coded language.

In other words: Tennessee has just outlawed clouds.

They're chemicals (H2O is definitely a chemical) and they affect the intensity of sunlight.

(from the github link)

Bit much for my local setup, so we'll have to wait a bit before we can test this properly.

LOL - far from it. Things are pretty fucked up in Europe, along many dimensions.

But Europeans are a lot less on the "EVERYTHING HAS TO BE THE BEST AND GREATEST EVER !!!!" train.

Nope.

The USA never does something. It always overdoes everything it does. Same here.

Elsewhere in the world, sports gambling is a small side-hustle for fans who like to spice things up a bit. In the USA, everyone is always for their personal ticket into the billionaire's club, the unicorn start-up or the license to print money.

"sane" is a word that's prohibited entry into the USA and shot at the border should it try. :-)

Ironically, that this kind of stuff would happen with a big probabilistic prodiction engine was to be expected.

Let the AI write code, they said.
It's really good at it, they said.
Look how fast it generates code when I put in a natural language prompt!
What could possibly go wrong?

Yes, there are some use cases where you want the LLM to essentially be a search engine on steroids. In that case, you need one that's online and vacuums up the Internet every so often. Essentially Google 2.

But for a lot of use cases a model that is occasionally updated will do just fine.

For similar reasons (I work on projects with serious security demands) I've gone down the rabbit hole to get local LLMs working and I'm pretty happy now, but it was quite a journey.

We now have stuff like Ollama and LM Studio that can run models locally, open models that have sufficiently large rolling windows, and things like privateGPT as a glue to feed in your own documents. Or Anything LLM if you want an all-in-one solution (though in my tests it didn't quite work as well).

We're getting there. In a few years, we'll have local AI integrated into our desktops.

I personally wouldn't invest into any AI-as-a-Service companies anymore, at least not for generic models. Maybe for models custom-tailored to specific use cases. But for the generic "write a report for me" LLMs, there really isn't a need for any government to rely on cloud services anymore.

With the current LLMs, noe jobs that need actual experience will be replaced anytiime soon. While the language capabilities are impressive indeed, even the best LLMs have a troubled relation with facts, truth and consistency. They'll gladly invent facts, put together totally different pieces of information of completely miss the point. For any question that I've asked an AI that requires expert knowledge where I have that knowledge (and used the AI to check or to generate more ideas) the answers I've gotten were at best junior level, and often trainee at best.

LLMs are the 21st century Wikipedia - pretty nice on common everyday topics, surprisingly knowledgable on a weird set of totally fringe things, far from something that'd be accepted in polite society.

We'll see a bunch of dumb jobs replaced by AI that so far were done by humans because computers didn't have the object recognition or ability to understand natural language. The first level call center job level.

Oh, and there's no such warning. When you get into the backseat of a robotaxi, it won't even start moving until everybody is wearing their seatbelts. I hardly see any passenger leaping into the front seat and grabbing the controls in a traffic incident.

Nah. I've rode a couple of robotaxis around the city, and while the ride is nice, it's clear we're nowhere near "get in a car and go to sleep." For one thing, the vehicles aren't even allowed on highways yet. And they require months of training on any particular urban area before they can perform reliably. I don't think there's been any training in rural or even suburban environments, which have different challenges. What you say may eventually be possible, but we're still a long way off.

Different user bases. I'm a Linux fan, a couple Linux projects have my name in it. All my servers are Debian. But my notebook and desktop are Macs. Because when I need to get desktop things done, that's the platform that works best.

I waited a decade for the Linux desktop to get there and it didn't. Then I stopped waiting and started being productive.

Would I want to have a Linux desktop and application choices equal to what I have on the Mac? Absolutely. Do I expect it'll happen anytime soon? Absolutely not.