wetdry.world is one of the many independent Mastodon servers you can use to participate in the fediverse.
We are a community focused on gaming, tech, entertainment, and more.

Administered by:

Server stats:

731
active users

microsoft paid $13 billion for this technology

this really blew up, thanks everyone. these are silly examples but i hope you took away that the "instant answers" that google and bing provide shouldn't be trusted. there's always the possibility of them misunderstanding your question or grabbing irrelevant sources to mislead you. "AI-powered" research is inferior to human research.

don't rely on paraphrasing; scroll right past and find a high quality source on the safety of drinking battery acid instead.

@keat [wetdry.world] @thepoliticalcat said
>For REASONS, I don't foresee this turning out well AT ALL. For ANYone.

It'll turn out fine for those of us who don't use Microsoft products. ;-)

@keat I love the last one. I mean sure technically you CAN drink milk and battery acid while you’re diabetic. If you’re still alive afterwards is a totally different question!

Though peanuts are apparently not possible to eat at all 😅

@keat lol, chatGPT 3.5 is free, why the hell do they pay 13 billion for this... 🤣

@keat @cjmoorehead It is astounding to me that huge corporations thought this was ready for public use.

@keat @maique That’s what we call edge cases, in the industry. 🙈

@keat this is an existential threat to humanity *accidentally bumps the plug to the server rack and takes the entire revolting thing offline*

@keat A fraction of what they paid for Activision.

@keat This is way to good to fact check, but I’m a killjoy - @pivic is this true?

@Dimduning @keat @pivic Let me Bing that for you, because nothing is too good to fact-check.

- bing.com/search?q=can%20I%20pu
this one gave the result in the screenshot when I visited the URL, but didn't show up in the archive, oddly: archive.is/Xq9ag

- bing.com/search?q=can+i+eat+pe (interestingly, you get different and accurate results if you capitalize the "I")
archive: archive.ph/OOTts

- bing.com/search?q=is+it+safe+t
archive: archive.is/Xaby0

- bing.com/search?q=can+i+drink+
archive: archive.is/r0O1P

I tried archiving the pages in the Wayback Machine first, but none of them included the panes with the large-print answer, and 3 of the archive.today ones do. So it must depend on the user agent.

Bingcan I put chihuahua meat in a microwave - BingIntelligent search from Bing makes it easier to quickly find what you’re looking for and rewards you.

@nev @Dimduning @pivic thanks for making these archives, a few people have already accused me of photoshopping 😅

for the record, i took these screenshots on firefox for android, using a VPN server based in toronto, ontario. i'm sure there are a ton of weird factors that affect the results you get, but those are the only two i can think of that you can modify yourself.

@keat Of course you can drink battery acid. This doesn't mean you should 🤣

@gunchleoc @keat this would seem to imply that you cannot and should not eat peanuts. EVER.

least deranged microsoft opinion tbh (and i LIKE peanuts)

@keat @waitworry I love how you can lead it on by doing stuff like adding "allergy" to an inquiry about peanuts. It's all "well shit I don't know but I know peanut allergies are serious"

it's like asking the drunk uncle at family events domain specific questions and he has no problem just bullshitting you by guessing based on intonation and context

@aismallard @waitworry @keat Haha, I use that analogy a lot, myself. People definitely answer questions in the same haphazard manner, so we have reached *that* level of parity.

@keat To be fair, there is a difference between "You can" and "You should"

@keat And "serious" people are treating it like it's mind-blowing tech

@liquor_american @keat have you had ChatGPT write code for you? its amazing!..until you ask it to send you the same code again and watch it rewrite it from scratch

@nutsling @liquor_american @keat I'm active on WordPress Stack Exchange. I've run across several questions where someone posts nonsensical code and asks how to fix it; when I ask for clarification, they tell me "Oh, ChatGPT wrote it for me".

@liquor_american
It literally is mind-blowing technology, in some cases... /s

@keat

@liquor_american @keat It being flawed doesn’t mean it’s not impressive. The two aren’t mutually exclusive. It’s both impressive and flawed. The problem is when people pretend the flaws don’t exist.

@nmn @keat I'm impressed by the amount of pretending going on !

@liquor_american @keat There’s a financial motive to pretend for many. For others, it’s a kind of naive hope that the flaws will go away soon.

For the second group, I can see where they’re coming from if you look at the pace of improvements.

@nmn @keat I think we have to be very careful when we decide what we think improvement looks like, and I think the people you mention with the financial motives are spending a lot of money and effort to convince us that being better and more convincing at bullshitting is the same thing as improvement. And a lazy news media is generally very happy to go along.

I think the average person has enough trouble detecting *obvious* bullshit in 2023, don't you?

@liquor_american @keat LLMs have their use-cases but we’re still discovering them as a society.

Also, while they’re getting better at being more convincing with lies, they’re getting more accurate too.

Based on the benchmarks for accuracy that have been created so far, they’ve gone from around 40% accurate to as high as 70%.

Wrong 30% of the time is still WAY too high to use reliably for now, but people are obviously extrapolating the most optimistic future.

@liquor_american @keat Sure, they're "serious" people, but they probably have a Tesla and an iPhone. They're impressed by the shiny, rather than the useful.

@keat I mean, yeah, you can microwave the meat of any animal.

Technically correct is best correct. Nothing to see here, move along.

EDIT: My comment applies to the first panel only, didn't spot the rest at the time.

@keat People will look back at philosophy books about A.I. and wonder why they never talked about the dumb A.I. problem

@keat I remember my parents having to pay hundreds because of the old AI "Smart Answers" which miscited a website. This is only going to get more common.

@keat add in the paid ads and bumped results from the battery acid selling to drinkers companies on top of ai's liking it and....

@keat i was surprised to see a fellow wet dry worldian in a random tech meme channel on telegram

@gravitos i'm glad they credited me unlike the repost i saw on twitter... not like i did anything that warrants credit though lmao

thanks for letting me know, never had anything go viral like this before and feels good knowing it made a bunch of people smile

@keat and this is what Satya calls responsible AI, try to imagine what irresponsible AI would look like 🤣