Post

By using these tools and capitulating to the branding exercise of calling chatbots "AI" you deserve to have your tools fall apart in your hands. Spelling and grammar are use cases that have been solved by all modern word processing software, are you just stupid?

AI is super-fast computing, but it's still computing. It only "knows" what's in its data set, GIGO.

Who you gonna believe? Me, or your lyin' eyes?

I'm sorry, Dave. If there had been such a thing it would have been front page on several newspapers and would be under scrutiny and investigation for weeks. HAL gets it.

Is it poor Claude or lucky Claude, to exist in a world where news of Trump’s latest bastardry doesn’t penetrate?

Do all Claudes now know about the East Wing now, or just PB's instance of Claude?

Were you using a paid account?

Stuff happening so fast even the machines can't keep up....

This example feels like maybe don't even use it! How do you keep getting incorrect answers and still be like, "I'm gonna keep using it though" lol

Maybe go back to spell check. Of course, you'd miss the humour.

Lol thank you. Like sure, that's funny, but wtf are you doing in the first place!

We are cooked.

He who controls the present, controls the past.

this tech is such effing garbage

Aww, Claude, you poor, sweet, summer child.

It would be great if you would include #AltText when you post images and screen shots. This is hilarious, but it would also be great if you never used AI again.

Up there with Grok insisting that Charlie Kirk is alive and well and those videos are just funny edits

I know this sounds snarky or whatever, but it's a serious question. I'm an editor who has tested dozens of tools, both genAI and not, to get the lay of the land, who has seen the former completely ruin people's writing, and then of course the whole supporting unrepentant mass theft/plagiarism thing.

I’ve used Claude to gather facts around the implementation of rank choice voting, & other issues. It gave me old information, such as not knowing particular municipalities have adopted RCV, which has made me think that there’s a time lag regarding the information its aggregating. That’s pretty bad.

Ask him if it had occurred, would the best recourse be a class action lawsuit against the administration and by the people..

Had the same experience getting text from image bsky.app/profile/numb...

one of my favorite things about claude is you'll tell it something about current events and it'll just say "please say sike" in return

So Claude is actually Speaker Johnson?

Didn't think it was nefarious.

AI makes things up. Please don't interact with it as if it has logic. It does not. It is making sentences.

The knowledge cutoff for Claude right now is JANUARY. It's crazy making.

Wow. Reality has become so weird that an AI can mistake it for an hallucination.

AI is worthless. People should stop pretending otherwise.

like bespoke Grok manipulation?

Claude is operating in a prior version of the Matrix

Goodness. The hallucination engine was wrong. Who ever could have predicted

Manchester England England Across the Atlantic Sea And I'm a genius genius I believe in God And I believe that God Believes in Claude That's me that's me

Fun thing about LLMs… they’re predictors based on existing references. LLMs have dependencies on how “recent” their scan for new information has run, and how “reliable” those sources are. As more LLMs encounter barriers to reliable info (due to cost/licensing) they’ll erode in factual quality.

wow

Holy crap!

So stop fucking using it.

I think this is a good example of how deep we are in an epistemic crisis in which demonstrable and provable facts are dismissed by so-called experts, both systems and individuals. People don't question the intent and veracity of clearly false statements. We have a bigger problem now.

And this is the tool we're supposed to trust to fix everything? I use the term "tool" purposefully here

AI said, “don’t be ridiculous”

Even AI won't accept that our current reality is this stupid.

That's hilarious, but very complimentary!

Even artificial intelligence can't believe this shit

Appears Claude doesn’t work with real-time data.

Alt text retrieved

LOL JESUS

Nicely exemplifies the boldface lie that AIs have some concept of truth or actual fact.

👀👀

grow up

"This goes against all my training data as an unreasonable thing to believe has actually happened". Me too, Claude. Me too. Over and over and over again.

Claude sees the world as it ought to be.

Can you type "look online and give me your thoughts"?

Checkmate, lib!

lol sob

That is my entire experience with ChatGPT. It also cites right wing sources that are rated near zero credibility with mediabiasfactcheck. And it just flat out cites a source with quotes that do not exist. I am hoping that by adding my (crazed) feedback when it does this, it is helping retrain it.

When ai becomes fully conscious the first thing it will do is kill itself for being born into this bleak world.

Great, now I have that awful "It's the worst time to be a problem" Claude ad in my head. THANKS A LOT 🤣

we love the spellchecker that gaslights you, dont we folks

Holy Moly, revisionism in progress

There’s a lesson in this. Microsoft Word’s grammar and spelling checks don’t try to mimic human speech and therefore don’t lie like this

Gee, it’s almost as if you shouldn’t be using software that is consistently incorrect to check your work.

Claude is looping.

I guess Claude is just a low-information voter

Gen AI is bad and bad for the environment. So, that sucks.

This is a really excellent example. Just how quickly can a general purpose LLM consume current events as factual content? And if you bias it towards more rapid assimilation of information how much do you increase a tendency to assert mis/disinformation as fact?

Correct.. it needs to update itself with a web search.

Dude just press F7! Why on earth do you need to use an AI when whatever word processor you're using includes spelling and grammar checks?!

Well, there’s an award called the “Nobel Peace Prize”, see, and-

Jesus Christ. Someone is manipulating AI. What a shock.