LaurelEHOct 2, 2023
Neverfadingwood 🇺🇦
Just seen a new favourite response to AI text. "Why should I bother to read something nobody could be bothered to write?"
Oct 1, 2023 at 5:55pmWeb
Show threadNicole ParsonsOct 1, 2023

@Neverfadingwood

ChatAI etc is how tech, & its investors on Wall Street & OPEC+, plan to craft plausible deniability for the onslaught of malign influence campaigns underway for 2024

They'll hype ChatAI until some OPEC billionaire buys its IPO on Wall Street, sells its services to Putin & Xi.
Then they can claim that they aren't responsible for moderating harmful disinformation on the 3 C's, covid, climate, & coups

Founder cashes in & founds another Bluesky to erode democracy, rinse & repeat

Show threadCGdoppelpunktOct 2, 2023

@Npars01 @Neverfadingwood

This is a plausible scenario.

Show threadRaysOfRedOct 1, 2023
@Neverfadingwood it’s only worth reading AI writing when the prompt you give it is such a fucking curveball that it trips over itself trying to give you an answer. For example, pretty much any AI SpongeBob or AI Family Guy scene
Show threadNeverfadingwood 🇺🇦Oct 1, 2023
@RaysofRed Ah, thank you very much.👍
Show threadDec [{()}]Oct 2, 2023
@RaysofRed @Neverfadingwood
No software based AI can beat Family Guy's writing team of Manatees in a tank filled with idea balls. #SouthPark
Show thread@ NovaNaturalist HAS ALSO MOVEDOct 1, 2023

@Neverfadingwood

I think we are all going to need an electronic Monk Plus for this...

#douglasadams #ElectricMonk

Show threadNeverfadingwood 🇺🇦Oct 1, 2023
@NovaNaturalist Must reread Dirk - it's been 30 years at least.
Show threadDutchOct 1, 2023

@Neverfadingwood @NovaNaturalist

Same. That quote sounded vaguely familiar and upon scrolling over the picture, the AltText told me why.

Show threadKeith V.Oct 1, 2023
@Neverfadingwood If AI is just conveying information, even though AI is suspect, I can sort of see value in that. But to have AI write books, which Amazon is already doing, or films and TV shows, I see no value in it. Perhaps one day it will get to the point where I can't tell, but it's not there now.
Show threadTom WalkerOct 1, 2023
@Neverfadingwood I never hit the boost button so fast
Show threadtechnetiumOct 1, 2023
Interesting @Neverfadingwood, last week saw an email from the unemployed agency, suggesting to use chatgtp to get a job.
This confirms what is happening to cv's and cover letters.
Show threadLisPiOct 1, 2023
@technetium @Neverfadingwood I'll welcome another nail in the coffin of cover letters.
Show threadLazarou Monkey Terror 🚀💙🌈Oct 1, 2023
@Neverfadingwood I've been saying this
Show threadOrbitalMartianOct 1, 2023
@Neverfadingwood now, that’s a good one.
Show threadJustin ⏚Oct 1, 2023

@Neverfadingwood This really sums it up. There will be only a very brief period of time where LLMs are effective at saving time. Very soon, if you want to demonstrate that something is *worth* reading, you will need to demonstrate that effort went into its construction.

It's like junk mail vs spam email. Junk mail at least occasionally got read.

Show threadjordanOct 1, 2023
@Neverfadingwood me with computer logs
Show threadDarwin WoodkaOct 1, 2023
@Neverfadingwood I'll have my AI read it for me!
Show threadAndrew MaybinOct 1, 2023
@Neverfadingwood Entirely correct. I’ve never understood the “GPT can generate copy” thing. If you’re employing a tool to write words like that, those words didn’t need to be written at all.
Show threadaethervisionOct 1, 2023
@Neverfadingwood I was under the impression that every word was actually written, just not necessarily at the same time as the other words in the sentence. More of a Frankenstein’s Monster than a killer cyborg, if you will.
Show threadIoanOct 1, 2023

@Neverfadingwood well...
English is not my first language and in some of my business replies, I apparently come across a bit "strong".

So I use ai to put my point across while not being so obviously pissed off.

Show threadRui ReisOct 1, 2023
@Neverfadingwood brilliant!
Show threadAdam R. WoodOct 1, 2023
@Neverfadingwood That's how I feel about videogames that don't include basic control instructions. Guess they didn't want me to play them...
Show threadLight Owl ☸️Oct 1, 2023

@Neverfadingwood Even worse yet: "why should I bother reading anything which was authored after the advent of #AI authorship, in case it was written by an AI - but is of course being misleadingly passed off as authored by a human - thereby highly likely wasting my time?"

#reading #books

Show threadactive landslideOct 1, 2023
@Neverfadingwood this is the exact argument that I have been making in my own community about peer review of papers that disclose having used LLMs. Just posing the question!
Show threadNeonard LimoyOct 1, 2023
@Neverfadingwood I'll send an AI reader to read it
Show thread'dan Oct 1, 2023
@Neverfadingwood awesome. And it's adaptable for any type of AI output...
Show threadSeth 🎙️Oct 2, 2023
@Neverfadingwood Good point.
Show threadKevin JardineOct 2, 2023
@Neverfadingwood Sadly people have been reading hack journalism for years. AI promises to be a step up from that.
Show threadEther DiverDec 14, 2023
@kevinjardine @Neverfadingwood no it doesn't
Show threadAnatisadalia 🐞🌺Oct 2, 2023
@Neverfadingwood
Show threadSean ReifschneiderOct 2, 2023
@Neverfadingwood In a podcast, they said "AI text is smooth, human text is jaggy.": Humans say unpredictable things sometimes, where the generative AI just says the predictable thing. I found that interesting.
Show threadChristopher WalkerOct 2, 2023
@Neverfadingwood it reminds me of a comic I saw where in one panel a person used an AI to save time writing by turning a bullet list into a page of flowery prose, and in the next panel the recipient used an AI to distill the page down to bullet points to save time reading.
Show threadDoc Edward Morbius ⭕​Oct 2, 2023

@Neverfadingwood Many moons ago I ran into someone after I'd given a presentation who handed me a high compliment:

"I've written programs to read the messages your programs generate."

Which isn't to say I don't get your point. But automated generation can, at times and in context, be useful.

When the presumption is human interaction, however, I prefer human generation.

Show threadMatt HellerOct 2, 2023
@dredmorbius @Neverfadingwood If the ouput of a program needs an entire other program to interpret it for human consumption, then... are you sure it was a compliment?
Show threadDoc Edward Morbius ⭕​Oct 2, 2023

@mh Let's just say that these were advisory reports of poor behaviour which required expedited response.

The programs expedited that response, and as they implicitly trusted my reports (or at least weighed them into the response factoring), I'd say without hesitation or caveat that this was in fact a compliment.

Both of us laughed when they'd said it.

@Neverfadingwood

Show threadMichael ZOct 2, 2023
@Neverfadingwood @lucy_who So AI was designed to write corporate mission statements!
Show threadpgcdOct 2, 2023
@Neverfadingwood which is why they're forcing AI down our throats.
I'm not exactly looking forward to the inevitable rise of AI Readers, but we all know it's coming.
Show threadAbu HodjaOct 2, 2023
@Neverfadingwood Someone writes what they want to say as bullet points, runs it through GPT, posts the result to the internet, where others grab it and ask GPT for a summary. It's a really weird sort of anti-compression scheme.
Show threadCGdoppelpunktOct 2, 2023

@Neverfadingwood

Give it to your AI to read it!

Show threadZeugsOct 2, 2023
@Neverfadingwood because it's your job?
Show threadsketchyTechOct 2, 2023
@Neverfadingwood the future is that nobody ever reads anything, one AI writes another AI reads, and then they decide together the fate of humanity.
Show threadJustin DerrickOct 2, 2023
@Neverfadingwood I feel like I can detect AI-written text through the lack of insight in writing. It says stuff, but can’t convey an opinion or present anything interesting. When I had it write code, it was broken in ways I couldn’t begin to fix because it was broken in so many ways.
Show threadZZ BottomOct 2, 2023
@Neverfadingwood 
Show threadMarcus MüllerOct 2, 2023

@Neverfadingwood real problem :(
It gets worse: People are asked to bother correct incorrect understanding that nobody could be bothered to form themselves.

For example, learners are using ChatGPT/competitors to get "easy" explanations to complicated topics. Then they go on the internet and ask questions based on texts or understanding based on these wrong statements and waste expert time, who get to debunk nonsense before they get to explain the truth.
for example:
https://electronics.meta.stackexchange.com/questions/9793/should-questions-about-chatgpt-responses-be-closed

Should questions about ChatGPT responses be closed?

There seems to be a consensus that ChatGPT answers should be deleted. What about questions that are essentially "ChatGPT told me X. Is X correct?" This question is a long exposition abou...

Electrical Engineering Meta Stack Exchange
Show threadbit101Oct 2, 2023
@Neverfadingwood mostly you're not supposed to read it. AI bots generate tons of trash content. Other AI bots visit and click on it, driving up fake traffic. Other AI bots see the traffic and buy ads on those heavily trafficked sites. Yet other AI bots click on those ads. Some human walks home with the profits.
Show threadUrzlOct 2, 2023
@Neverfadingwood I consistently refuse to use the "Important" flag on email at work on the principle that if I don't consider an email important enough to read, I'll let you know by not writing it.
Show threadinfocyteOct 2, 2023
@Neverfadingwood What a very strange perspective. Shouldn’t ideas, in general, be judged based on the merits of the ideas themselves, rather than the source of the ideas? The alternative, which this quote expresses, is a kind of cultural chauvinism that probably feels good, but is really unhealthy in my opinion.
Show threadSome Kind Of AnhedoniaOct 2, 2023
@Neverfadingwood Great quote, but I guess the person who said it assumes people and places will always be forthcoming about whether something is AI-written?
Show threadJez 🍞🌹Oct 2, 2023

@Neverfadingwood While I agree with the intent behind this, that phrase "nobody could be bothered to write" assumes that people use a technology because they're lazy and consider the task beneath them. Disabled folk who use assistive technology to communicate, for example, would strongly disagree.

The problem, of course, is that enabling these genuinely life-changing use cases is of no interest to those peddling generated text right now. More money to be made from advertising "content".

Show threadNeverfadingwood 🇺🇦Oct 2, 2023
@petrichor That's a fair point, absolutely. I would just say that "assistive" is the right word. It still needs human input or it's going to produce bland unengaging texts. In my experience so far, at least.
Show threadChris JonesOct 2, 2023
@Neverfadingwood Is there an artificial reader for artificial content?
Show threadWarmasterPalakOct 2, 2023

@Neverfadingwood

Absurdist technological conjuction:

AI written story being read by a screen reader that's been muted.

Show threadÓscar Morales VivóOct 2, 2023
@Neverfadingwood paraphrasing the old Soviet joke, we pretend to read it and they pretend to write it.
Show threadAviva GaryOct 2, 2023
@Neverfadingwood this is the way...
Show threadrequiem 🦫Oct 2, 2023
@Neverfadingwood I literally said this to a co-worker who told me he has LLM’s write his docs and proposals.
Show threadNeverfadingwood 🇺🇦Oct 4, 2023
@requiem Oh good grief. Does he think nobody will notice? Or that they don't care?
Show threadrequiem 🦫Oct 4, 2023
@Neverfadingwood they don't care. I'm sure that he will get promoted for this choice.
Show thread6t8kOct 3, 2023
@Neverfadingwood Bingo ❤️