Mastodawn

I'm so tired. People, I am not interested in reading *blogs* generated with claude, gemini or chatgpt. No, not even ones proofread by AI. Do it yourself. I do not wish to even accidentally click on "one-shotted" dummy projects. Don't show me your AI-generated analysis of anything. Above all, I do not wish to see images/artwork someone prompt created.

I value real content, with all its good and bad quirks, from actual human beings.

Somebody please normalize auto-tagging of AI generated content across all formats - text, images and code so I can blanket ban them from my personal computer.

This is a plead for help I never thought I'll have to make.

#noai #enshittification #writing #bookstodon #reading #books #quotes #quote #cybersec #mentalhealth #adhd #art #cooking #blog #indieweb

Jason Morse Feb 18

@pheonix I’m seeing certain outlets proclaim themselves AI free, which is great, and I think we’ll be seeing more of this practice.

Windy city Feb 18

@cloudhands A differentiator which will be a USP probably from next year onwards 🤞

Grow Fediverse Feb 18

@pheonix Right? Like at this rate how about we watch AI sports, where instead of paying human competitors we just see generated video of fictional humanoids in made up Olympics 🤪

Grow Fediverse Feb 18

@pheonix Seriously though. We should actively start talking about replacing athletes with GenAI. Watch how big mad the sportsball people will get 🤣 But observe the merits:

- Watching AI athletes is safer for humans. No risks of injury to people.

- It's cheaper for humans. We can cut all the public spending on school sports programs and publicly funded arenas

- It's better quality, the AI athletes will never trip, miss their jumps, fumble a ball, or slip on surfaces.

- It's cheaper and faster to produce. Training human athletes takes years and specialized equipment, and only happens at the rate humans can have babies. AI can crank out millions of atheletes and teams in seconds.

- AI refs will never make mistakes either. In fact they and coaches can just be abstracted away and eliminated entirely because the rules and strategy of the game will be part of the AI agent players.

Or maybe putting AI in everything is an annoying idea #NoAI #enshittifaction #AISlop #GenAI

@growfediverse @pheonix

In terms of turning the general populace against Gen AI, that proposal might work. Sports have lots of time and cost investments, and this proposal would have all those people, children and their supportive parents and peers, rally against your awful proposal, as their sunk costs (no fallacy here, move along in an orderly fashion) would have the marching down the streets pretty quick. What if we did to sports what we are doing to workers?

@pheonix It is *so* annoying.

Duckduckgo has a no ai search, which works as well as you’d expect. This morning, I heard of an artist and wanted to see more examples of his work. Even with “AI: hide,” about one in five images came from domains that seem dedicated to generating slop “in the style of” an artist.
And yeah, I could see the difference between flat art nouveau and shaded slop, but I shouldn’t need to waste my very limited energy on that.

I tagged those to ddg, hope they’ll shit list those domains.

https://noai.duckduckgo.com/

DuckDuckGo - Protection. Privacy. Peace of mind.

The Internet privacy company that empowers you to seamlessly take control of your personal information online, without any tradeoffs.

DuckDuckGo

Windy city Feb 18

@toni I didn't know about the no ai duckduckgo variant so thanks for bringing it up! 😍

@pheonix Sure, happy to have helped. :)
I’ve added the URL to my original post, to make it easier for others who might stumble over it.

@toni @pheonix This is because DuckDuckGo only uses a block-list to filter out sites that post primarily AI images. They can't actually tell what images are AI in a reasonable amount of time/processing. So some will get through.

@JessieImproved @pheonix I suspected as much. That’s why I tagged images from craiyon, midjourneyart etc.

𝙳𝚊𝚒𝚕𝚢 𝙵𝚒𝚜𝚑𝚠𝚛𝚊𝚙 Feb 18

Sounds like DDG needs to implement an AI agent to filter out those images 😶

@Sheep_Overboard

blobfoxsweating

@toni @pheonix I tried to search for a comic I’d seen on the topic of AI. All I got in return was garbage ai comics.

blobfoxnotamused

@pheonix 💯 agree

Nu Modular Feb 18

@pheonix #BoycottAI

Howitzer105mm Feb 18

@pheonix I participated in a virtual meeting late last year and the subject of AI/LLM came up. I expressed my multiple objections to the dreck that it is. One of the participants mentioned a way to tag things 'human created'. I didn't follow up at the time. It appears they may have been talking about notbyai.com?

Filter number one for (auto-)tagging: Does the text contain spelling errors or editing mistakes?
Number two: Count the fingers.
Number three: Exuberant writing or telling style.

Windy city Feb 19

@jerrej IIRC, Wikipedia has a section on some popular AI writing styles. That could be a great start!

Themikina Feb 18

@pheonix We need something crowd-sourced like SponsorBlock, but for AI content.

It has its issues, especially with false positives being super damaging to actuall artists you want to support, but at this point, I'd love to have something like that.

Jordan Maris 🇪🇺 🇺🇦 #NAFO Feb 18

@Themikina @pheonix ublacklist + the AI blacklist: https://ublacklist.github.io/docs

https://github.com/laylavish/uBlockOrigin-HUGE-AI-Blocklist

Introduction | uBlacklist

uBlacklist is a browser extension that filters Google Search results, available for Chrome, Firefox, and Safari.

ΞVΞ🦋Feb 18

Agreed. I like to read personal blogs, and what I hate is taking the time to read something that sounds interesting, and then realizing about two paragraphs in that it’s AI. Like it’s trying so hard to pretend it’s real writing, but losing the human composition cues to keep it up for long-form writing. It starts to sound robotic, repeating things, talking too much about a topic, too wordy, saying weird phrases not native to human conversation, etc.

It reminds me of when the scourge of SEO hit the internet. Bloggers wanted to increase visitors to their sites by keeping to the top of Google search, so they abandoned writing organically and only wrote in a way to appease the SEO algorithm. This led to the deterioration of writing quality of blogs and good, creative personal blogs got buried to the back of Google search under all the SEO slop.

Benjamin Schmidt Feb 18

@eve unfortunately this kind of human behavior will always repeat. It's never about who's best, just who's loudest (or has the most money to be the loudest).

Windy city Feb 19

@eve and now those same SEO spamsters have these chatbots writing unlimited slop 24x7. What a colossal failure of the system in that regard.

This was my wish as well from the first moment I was suspectof the images and "news" reports done by AI.

I can gyuess why the lawmakers haven't just passed a law that states everything AI must be labled as "created with AI" or "written with AI assits?"

martin lentink 🇪🇺 🇺🇦📎Feb 18

@pheonix
Hear! Hear!

Artstories Feb 18

@pheonix But it is also so, so very difficult to cut through the noise. If you don't use AI your ability to produce content is limited due to your time. AI spits it out in seconds and google loves new content.

Then there are summaries on top of the search engine so that people don't see the need to visit your blog. Next is a google conversation so that it can give a complete summary of the contents.

It is exhausting. You can't keep up with this.

MintyFries Feb 18

@pheonix I wish I could find the words to properly express to the people who keep sending me AI slop, how incredibly dull I find such content. No I don't care about your conversation with Chat-GPT yesterday, a clanker's opinion is meaningless.

@mintyfries @pheonix I don't understand the joy people get from sharing AI responses. My favorite is the screenshots of it instead of copy/paste because even that is too much work.

Windy city Feb 19

@cereal_cable @mintyfries IKR. so lazy

derptron Feb 18

@pheonix "Somebody please normalize auto-tagging of AI generated content across all formats - text, images and code so I can blanket ban them from my personal computer."

I guess I could maybe come up with an AI for that. Don't have a lot of experience doing that, but it seems like it doesn't take much. All I have to do is say it does the thing, right?

The Turtle Feb 18

@pheonix ...but it's FUN to misuse AI to make trump look ridiculous, far more so than my actual talents could achieve.

Making trump look ridiculous is a worthwhile cause. It was my absurd brain that came up with the idea; AI just moved some electrons around.

DeeAnn Little Feb 18

@the_turtle I think though- this is hard to explain but I think there might be something fundamentally different in the way some of our brains are processing generated information. "Uncanny Valley" is generally used as shorthand, but doesn't convey the actual, physical fatigue effect, even when used for humor. The synthetic graphics have humor, but are still tiring to look at. The same as reading synthetically composed text. I don't really have the vocabulary to articulate it though. @pheonix

DeeAnn Little Feb 18

@the_turtle I don't know exactly what it is, and it's different than something like fiction written by a person, or a funny photoshop.

Edit to add- it might go along with the "ultra-processed food" idea that @christianrickert mentioned upthread.

The Turtle Feb 18

@chillicampari @pheonix @christianrickert i see parts of it as rather freeing for some people who just don't have the skills to do it by hand or in Photoshop, or the money or patience to hire an artist to explain "I need a bridesmaid in distress being chased by hundreds of giant hot dogs in the Niagara River" or "I need an image of Anne Hathaway confidently leading an army of blue giraffes on the Moon."

Orb 2069 Feb 18

@chillicampari @the_turtle @pheonix @christianrickert

Normalizing fake photos only hurts the side that cares about truth.

The Turtle Feb 18

@chillicampari @pheonix

31 years ago, the lady who photographed my first wedding was very skeptical of digital photography. A year and a half later, i got a really shitty, crude Kodak digital and showed it to her. You could see her brain kick in when she understood the possibilities, and within two years she was shooting mostly digital.

In 1972, my music teacher scoffed at synthesizers.

Seven years later, I got an F on a math assignment because the teacher was appalled "you used a computer!"

DeeAnn Little Feb 18

@the_turtle don't get me wrong, as my own anecdotes I had the first Agfa consumer digital camera and a very early adopter of new technology (I also maintained the first edition of the plasma display FAQ and worked in emerging/experimental hardware at SUSE). I also did early video streaming as an Akimbo partner. So I know there are bumps. But whether it is general lack of technical maturity at this release stage or something else- there is still a fatigue effect. @pheonix

@pheonix I am a law school TA and it actually warms my heart to see a typo because there’s humanity behind it. Infinitely more forgivable than hallucinated case law

Windy city Feb 19

@esnupi That's so much worse than a simple email. I can't even imagine how meticulous the courts have to be, since any paralegal or unprofessional legal person can run their documents through an LLM and it may hallucinate precedents and whatnot.

@pheonix I'm so tired of seeing AI slop images on LinkedIn and the like, especially professionals I used to look up to using this constantly... it's annoying and shows lack of creativity and care.

Windy city Feb 19

@simplyxtata I've noticed LinkedIn is a particularly notable offender in this regard. Somehow, they ended up sorta normalized writing and posting images made with these chatbots.

Not that it was any better before, but the people at least put in some effort to write corporate drivel to gain popularity on the platform. Plus, this chatgpt writing style is so boring, predictable and cookie-cutter, it feels like an insult to the reader.

@pheonix I have a weird take but I find the worst offenders of using AI for writing, making charts, coming up with ideas etc. essentially have low confidence in their skills and view the chatbot as a free executive assistant that in reality is not trained to do any of this. The most insidious part is that these generative AI companies are basically selling a LLM (a glorified version of auto complete) as a wonder tool when in reality there are so many errors, issues and problems with using these tools— and that is before we get into stuff like the unethical nature of these companies and things like IP theft.

Something that grinds my gears is how people will say— oh well I use Claude or Perplexity vs. Grok AI or Chat GPT when in reality, all these things steal and copy off of each other and have the same ethical issues (with Grok having the most in terms of producing rampant exploitative material). A lot of my classmates love to state they like using Claude for homework help and essay writing. To me, it gives off desperation, lack of concrete skills in writing and researching as well as the early 2010s vibes of a hipster saying "oh, I use Apple vs Windows because it's more protected" when in reality it really wasn't that much more secure if you worked in actual pen testing/hacking.

Overall I find it lazy and I can go in and break these things pretty easily and I have NO CS or computer background.

Windy city Feb 19

@simplyxtata Very valid points and I couldn't have said it any better!

MissConstrue Feb 18

@pheonix Right? I am so ready for a human only internet.

Windy city Feb 19

@MissConstrue exactly!

Bruce Mirken Feb 18

@pheonix Amen!!!

@pheonix I absolutely get not wanting to waste time on AI slop.

However, there's something I don't understand.
What's the problem with posts proofread by AI, though, assuming we're talking actual proofreading here (spotting mistakes, as opposed to stylistic advice)? Why would it matter whether a typo was found by a friend of yours or some large probabilistic model?

@pheonix Right? Give me something real so i can appreciate the effort.

Care > Content

Adam Faircloth Feb 18

@pheonix I feel the same way but also think this change feels inexorable.

What I don't understand is why humans are so eager to build a world not suitable for humans.

@pheonix What is the main reason you value the 'real content' in this case?

Personally, I want to see quality content, and inherently I don't care what tools were used to make it.

@pheonix : welcome to the #smallWeb :

https://ploum.net/2023-08-01-splitting-the-web.html

#gemini_protocol , #gopher #smolnet

Yes, you will need to learn a bit. This is the price to pay ;-)

Splitting the Web

Splitting the Web par Ploum - Lionel Dricot.

Thomas Arildsen Feb 19

@ploum @pheonix stupid question: what is Gemini that you refer to? I assume it is not Google Gemini?

https://en.wikipedia.org/wiki/Gemini_(protocol)

https://geminiprotocol.net/

addresses start with gemini:// like gemini://ploum.net

#geminiprotocol

Gemini (protocol) - Wikipedia

The Febrile Muse Feb 18

Yeah, no grok screenshots either.

Jack Yan (甄爵恩)Feb 18

@pheonix I wonder if I would be OK with “AI”-proofread blogs if the writer was not writing in their first or most familiar language. But I have received “AI”-translated emails and they were just horrible. It took out all their personality and I actually had to ask one friend to write to me normally as there was little wrong with the way he usually wrote. Sure there were mistakes but for a private email, who cares as long as I understood him?

@pheonix Well, my blogs are in Galician, but IA free. You can try one of them: https://blogs.amarinha.gal/ribadeando/

Ribadeando – Ribadeo e as miñas cousas (chámalle Blog / Weblog / Bitácora / Caderno … )

Ω 🌍 Gus Posey Feb 18

@pheonix 🏆🍺

@pheonix Are... people actually doing that? What the fuck?

@iveyline Feb 19

@pheonix I'm with you. AI is not intelligent. It's being imposed on us without our consent. Like you, I only want to see human generated content created with real human intelligence. It would be fantastic if all AI generated content was tagged as such. The we can make our own choices instead of having them imposed on us.