So this is weirder than it looks at a glance.

That is not an LLM-generated search result. That is a funny ha-ha mistake a LLM made that then some guy compiled in his blog about AI.

Google then did their usual content-stealing thing, which probably does involve some ML, but not in the viral ChatGPT way and made that card by quoting the blog quoting the LLM making the mistake. And then everybody quoted that because it's weird and funny and it replicates all the viral paranoia about this stuff.

Hubspot AI chat bot told me to go three levels deep into a menu that doesn't exist, to click a button that doesn't exist to enable a service that doesn't exist to solve a problem I had.

My company pays a 5-figure yearly sum for this service 👍

Is there a uBlock filter list for AI SEO websites? If not then I guess I should make one, it would make my life so much easier especially when looking for a product

This joke is really old now.

And yes, even if such mistakes are funny at first glance, it doesn't change the fact that the field of AI has developed incredibly in the last year. And this development actually has the potential to completely change our economy. And not only that.

No, I'm not fun at parties.

For niche topics, search engine AI is less than worthless. It produces an unacceptably high proportion of misinformation.

Kenya doesn't start with K. It actually starts with K. Completely different.

No matter how dumb AI is, it will be an improvement over a lot of people.

spelled with a "K" sound

I'm not even a linguist but reading this little snippet already makes me want to smash my head against a wall.

It's always a race to the bottom to create the most content for the least amount of money and effort, isn't it? The problem is, Ai generated content is crap.

"Water can be hot or cold. You shouldn't drink too much of it. Water can be stored in containers so it won't spill. You can cook things in it."

Please quote me what this character said in this public domain story

Chatgpt: (Quotes wrong character)

No, that is a different character

Chatgpt: (invents hybrid character)

Sends it a link to the public domain text

Chatgpt: I can't follow links

Of course there isn’t. And you know the word “gullible” isn’t in the dictionary, either.

This is AI gaslighting .

Its fun when you ask it a question and you counter and it says "You're right".

Yet somehow we'll never get ubi or free healthcare.

Yeah op, it will. Just because ai is an idiot sometimes doesn't mean it won't vastly improve. Many jobs will be destroyed, denying it does everyone a disservice

Well, gotta say it's improving too fast to joke about it.

Haha they will. Corpos won't care if their products and services are bad or worse through AI; they're de facto monopolists and using AI drives down cost tremendously.

What would they care this leads to a worse society as long as THEY benefit?