#will chatgpt kill google

[juviin]

if u need to use gen ai for simple language hw im afraid you are terminally stupid. like brain dead terminally stupid.

[cuckoo-on-a-string]

When a new client comes crawling into your inbox begging for help because they tried using ChatGPT for their business site and it done fucked 'em.

[the-real-google]

Look at all the blood on your hands, Bing.

So Bing's AI is not helpful with real, day to day problems it seems.

[rellbot]

is google just shit or is the shit im studying some fucking government secret man wtf

[tumblasha]

every once in a while i play devil’s advocate and go to a “learn how to use gen ai” zoom training for work. today the zoom host starts by asking chatgpt about the weather in a major city… maybe next time!

[generation-of-vipers]

the decline of human creativity in the form of the uprise in AI generated writing is baffling to me. In my opinion, writing is one of the easiest art forms. You just have to learn a couple of very basic things (that are mostly taught in school, i.e; sentence structure and grammar amongst other things like comprehension and reading) and then expand upon that already-foundational knowledge. You can look up words for free— there’s resources upon resources for those who may not be able to afford books, whether physical or non physical. AI has never made art more accessible, it has only ever made production of art (whether it be sonically, visual, written—) cheap and nasty, and it’s taken away the most important thing about art (arguably)— the self expression and the humanity of art. Ai will never replace real artists, musicians, writers because the main point of music and drawing and poetry is to evoke human emotion. How is a robot meant to simulate that? It can’t. Robots don’t experience human emotions. They experience nothing. They’re only destroying our planet— the average 100-word chat-gpt response consumes 519 millilitres of water— that’s 2.1625 United States customary cups. Which, no, on the scale of one person, doesn’t seem like a lot. But according to OpenAI's chief operating officer , chatgpt has 400 million weekly users and plans on hitting 1 billion users by the end of this year (2025). If everyone of those 400 million people received a 100 word response from chat gpt, that would mean 800 MILLION (if not more) cups of water have gone to waste to cool down the delicate circulatory system that powers these ai machines. With the current climate crisis? That’s only going to impact our already fragile and heating-up earth.

but no one wants to talk about those stats. But it is the reality of the situation— A 1 MW data centre can and most likely uses up to 25.5 MILLION litres of water for annual cooling maintenance. Google alone with its data centre used over 21 BILLION litres of fresh water in 2022. God only knows how that number has likely risen. There are projections of predictions saying that 6.6 BILLION cubic meters could be being used up by AI centres by 2027.

not only is AI making you stupid, but it’s killing you and your planet. Be AI conscious and avoid it as much as humanly possible.

[tunisian]

just overheard someone say “im going to google this…no wait. ill ask chatgpt” and i never wanted to tell a person to kill themselves so severely

[thedialoguedilemma]

The Dilemma Bulletin: Monday January 27th, 2025

Keeping you informed about the daily events of the Trump Administration

China releases an AI source called DeepSeek set to rival ChatGPT and OpenAI. China claims it created this AI at a fraction of the cost causing US markets to trend downhill.

China releases an internet communications satellite set to rival Starlink.

NVIDIA has fallen out of being the world’s most valuable company after losing $600 billion dollars in market capitalization today under the Trump administration

Trump continues to threaten tariffs on nations that don’t agree with his policies which will skyrocket prices for everyone.

A January 6th terrorist that was pardoned and released into the public by President Trump was shot and killed by police during a traffic stop in Indiana following an altercation.

The Trump administration has instructed public health officials across the United States to stop working with the World Health Organization effective immediately as bird flu cases rise in the US

After Trump threatened Colombia with tariffs this past Sunday, coffee prices have risen in expectation of possible trade war with Colombia.

President Donald Trump announces a tariff on foreign semiconductor chips which will cause electronics such as refrigerators, game consoles, tvs, computers to all skyrocket in price.

Trump claims the United States military entered California and just “turned the water on” What does that even mean????? The pumps were temporarily off for a 3 day maintenance and are now up and running.

Trump says Microsoft is in talks to acquire TikTok

Google Maps updates Gulf of Mexico to Gulf of America despite every other world country still referring to the body of water as the Gulf of Mexico

Trump expected to sign an Executive Order banning transgender people from the military. Will be challenged in court.

Legoland Florida announces massive upcoming layoffs as a result of Governor DeSantis and President Trump leadership

Trump announces an idea of wanting to fly American criminals to other countries who will incarcerate them for "a small fee". Again— will be immediately challenged in court.

[cursedhaglette]

listen to me very carefully. come close, huddle up. take my hand.

unless it's set to private, the fic you have posted on ao3, the fic that is indexed and searchable through google, has been scraped and has already been used to train AI.

i'm going to say that again.

your fic, whether you want to admit it or not, has already been fed to the AI monster at large. this is the reality of the internet we use and the age we live in. AI knows what's in every fic existing on the internet right now, and Gemini likely knows what is in most, if not all, of your WIPs.

that's reality. it sucks.

if you are so concerned about this reality that it is giving you horrible, mind breaking, earth shattering anxiety, causing you to lash out to the people on the internet around you - it's time to put your future works on private. it's time to stop clicking on chatgpt and saying "hi i'd like one fanfic about blorbo penis please" and pressing enter, then posting, because that is what is actually making use of that scraped content.

but if you're someone doing that, it's not actually about the anxiety or the AI or the wah wah wah of it all. it's not about concern for how AI is going to kill the rainforests and melt the icecaps, nor is it concern for the death of creativity at the hands of the techy elite. it's about lying and attention and fandom clout. it's about making a scene. it's about playing the wounded dog over and over and over again until people can't help but come and listening to your bleating and whining and say "aw puppy what happened" and ignore the sense put in front of them.

but just know, there is absolutely logic to this, whether some people want to see it or not. AI is going to be here for who knows how long, and it is not making content that will make your heart sing, regardless of how the beaten dogs cry. it will give you four snowball fights in ten chapters because it doesn't know better. it will give you the same metaphor repeated three times in three paragraphs because why shouldn't it? it's a computer. you're the person.

you're supposed to know better.

[prontaentrega]

group project classmates have done fuck all so far like genuinely only one of them even opened the google doc and it was to put her name on it and add nothing more and its fine. i can live with that. if any one of them pulls up at the last minute with 30 pages of chatgpt slop im killing a hostage though

[independentanon]

"EVERYTHING IS NORMAL" "THEY'RE █████████ THE INTERNET!"

Sometimes you'll notice changes being quietly made to your favorite internet services. Be it a video platform, your search engine of choice, your favorite social network, or wherever you go to buy the things you need or want. Sometimes you'll also see changes in advertisements that were, suspiciously, only relevant to your own unique situation.

You know it in your gut that something definitely changed, but it was only worth mentioning in conversation. "This changed for me, did it happen to you too?" Some of these changes are experienced by everyone all at once, but others are limited to specific groups, and sometimes are rolled out in staggered waves, meaning only some people are affected at different points in time. By the time the change is fully implemented - when every person targeted for this change is affected - it doesn't even matter anymore. The companies making these changes could report them publicly if they wanted to, and all people could do in response is be annoyed by it but eventually accept it and move on. The idea of "boiling the frog" comes to mind.

Our services have been getting worse in some ways, better in others, but there's undoubtedly some changes that are bad for everyone but the companies supplying these internet services (and sometimes, secretly, the governments of various countries around the world).

For me, personally, I've noticed changes to Meta (Facebook), to Google (and its services, Google Maps and YouTube), to ChatGPT, to Twitter - oh sorry, to "X", and many more. These changes are relatively small and are mostly unnoticeable... but I noticed them, just like all the other little changes they've quietly rolled out over the years. However, these changes feel a bit more insidious.

With Meta (Facebook for me), it was that they started suppressing accounts that frequently posted political content. This became most obvious during and after the 2024 election.

With Google, it was how it seems to bury certain content that's relevant to your given search, such as proof - one way or another - that something was happening with our politicians that's valuable knowledge to the public, but apparently isn't relevant enough to be on the very first page (or is simply hidden away entirely). This isn't even mentioning that Google modified its maps service so The Gulf of Mexico now reads The Gulf of America...

With YouTube, it's how it prioritizes click-bait, rage-bait, heavily-one-sided discussions of political topics, rather than pushing the very proof (or at least the very best evidence) that paints the clearest picture these overblown discussions are about. It's clear they're prioritizing watch time and engagement instead of truth.

With ChatGPT, I knew they had to control their generative text AI behind-the-scenes for certain situations (naturally you don't want your service to be generating stuff like "kill yourself," hate speech, lies, etc...), but recently it seemed to change its sources when looking up news online, to the point that it now paints a favorable image of Trump and his people.

And Twitter... sorry, with X... well, I shouldn't even need to explain this one, but I will try. The richest man in the world bought Twitter, changed how some of the back-end works, dramatically changed which voices were suppressed and which ones were heard, allowing hate speech and misinformation to spread freely on the platform, even promoting misinformation directly by retweeting it... there's a lot to it, but just know that Twitter used to be less shitty than it is now. Now it's really bad.

The point I'm making is that a lot of these changes happened around or soon after the 2024 election, and the people controlling these companies showed up to Trumps inauguration. On top of their million dollar donations to Trump, they're also doing work on his behalf to mask what awful things him and his people are doing while simultaneously promoting the things that make them look good. In short, information is becoming less accessible.

All of this, of course, is ignoring what Trump and his people have done to our government-provided websites and services, like removing the constitution and more from whitehouse.gov, how they're scrubbing decades of data from the CDC, etc...

The worst part about all this is I don't know if I could even prove anything anymore. These changes have made it difficult to know what services can be trusted going forward.

These are terrifying times. If the censorship was bad before, it's so much worse now.

Although I'd usually go out and protest with these signs, I've decided not to do it with these ones. I'd practically be an actor or an NPC, repeating the same visual joke over and over. These are my first signs I won't protest with. At least, for now.

Nonetheless, don't forget to fly your flags upside-down, boys and girls and non-binary types. Stay safe, and fuck Trump & Co!

[mugasofer]

So, NGL, I initially found this pretty persuasive. It doesn't negate due process, or more clearly innocent ppl, but still.

Marijuana (marihuana), Smile (sonrisa), a skinny cross that looks like a 1, and skull = Calavera or Cráneo = C = 3rd letter of the alphabet). That last one might seem like a stretch, but number codes are a pretty common thing with gangs.

Then I saw someone on bsky suggest:

It got me thinking: what if he'd been accused of belonging to a different gang?

Cannabis, RIP, |, Poison = CRIP (Crips)

Bud, Laughing, On God, Dead = BLOD (Bloods)

Skunk, Kill, |, No More = SKIN (Skinheads)

Alfalfa, Laughing, King (of Kings), Noggin = ALKN (Almighty Latin King Nation)

Bud, Eyes, Saviour, Terminated = BEST (Barrio Eighteenth St.)

Of course, the accusation is that the meanings are in Spanish, which makes sense. I don't speak Spanish. But ChatGPT does!

Santa María, último, religión, esqueleto = SURE (Sureños)

Narcóticos, ojitos, religión, tóxico  = NORT (Norteños)

THC, risueno, evangelio, no-muerto  = TREN (Tren de Aragua)

Brote, emoticono, santo, toxico = BEST (Barrio 18)

Lechuga, alegre, 1, huesos (h=8) = LA18 (Barrio 18)

These aren't cherrypicked; these are the first gangs that came to mind or came up when googling. Notably, Barrio 18 are the gang he's allegedly in the US to escape after they forcibly recruited him; the reason he was specifically barred from deportation to El Salvador!

h/t @slatestarscratchpad for first giving me a sense of just how easy it was to create these kinds of "spooky coincidences", with his blog post on this sort of thing The Pyramid and the Garden and unsongbook.com.

[drmsss]

desi aur pahadi mein kya fark hota hai batao

[btw I felt your vibe is scary cuz your scary posts came on my feed but I'm still gonna go without anon and follow you cuz scary people are not scary scary they just scare scary people and scary here doesnt exactly mean bura and darawna wala scary just bade tauji ka ladka scary pls don't kill me]

Go check ChatGPT/Google vrooo… I am scary right? They’ll explain it more clearly than I can…. :)

@cee-caw

[just-antithings]

https://www.tumblr.com/just-antithings/770058264120573952/where-do-i-even-begin-on-this?source=share

I kinda want to talk about this, not just because of this post but I’ve seen a 21 year old rely on someone else’s google screenshot to say “yes fiction affects reality, fiction is propaganda!” (And they were arguing about the “moralities” of a Genshin ship)

Propaganda is basically the dissemination of (biased) facts and information, rumors or lies to influence public opinion, it can be used to convince people to believe an idea, buy a product or support a cause. Very often propaganda exaggerates a stereotype or the truth, it contains information that has been manipulated to influence public opinion

what antis don’t understand is that fiction is less likely going to be used for this kind of thing because fiction is often too fantastical and rarely contains any sort of information. I mean, how are you going to be influenced by a fanfic of a Genshin ship? What kind of information does it contain?

now compare this to “the turner diaries” novel, which is often referred to as the “bible of the racist right” because it dissiminates white suprematism and it was credited with influencing the criminal activities of hate groups (like the KKK) and individuals.

One of these individuals was Timothy McVeigh, who bombed a federal building in Oklahoma and which killed 168 people

Propaganda influences people’s already existing prejudices and biases. Emphasising on “already existing” because it can’t give people new opinions or biases, but it can influence and shape those

which is also why it won’t work on people who inform themselves, who think critically and stay educated. A person who will look up information rather than blindly trusting information given to them (by others) is less likely going to be influenced by propaganda.

Of course no one is immune to propaganda, but it’s important to stay informed and to think critically, and that’s something antis can’t do. They’ll blindly believe (mis)information, rumors, screenshots from google or heck, maybe even ChatGPT

.

[kalikalahansa]

On generative AI

I've had 2 asks about this lately so I feel like it's time to post a clarifying statement.

I will not provide assistance related to using "generative artificial intelligence" ("genAI") [1] applications such as ChatGPT. This is because, for ethical and functional reasons, I am opposed to genAI.

I am opposed to genAI because its operators steal the work of people who create, including me. This complaint is usually associated with text-to-image (T2I) models, like Midjourney or Stable Diffusion, which generate "AI art". However, large language models (LLMs) do the same thing, just with text. ChatGPT was trained on a large research dataset known as the Common Crawl (Brown et al, 2020). For an unknown period ending at latest 29 August 2023, Tumblr did not discourage Common Crawl crawlers from scraping the website (Changes on Tumblr, 2023). Since I started writing on this blog circa 2014–2015 and have continued fairly consistently in the interim, that means the Common Crawl incorporates a significant quantity of my work. If it were being used for academic research, I wouldn't mind. If it were being used by another actual human being, I wouldn't mind, and if they cited me, I definitely wouldn't mind. But it's being ground into mush and extruded without credit by large corporations run by people like Sam Altman (see Hoskins, 2025) and Elon Musk (see Ingram, 2024) and the guy who ruined Google (Zitron, 2024), so I mind a great deal.

I am also opposed to genAI because of its excessive energy consumption and the lengths to which its operators go to ensure that energy is supplied. Individual cases include the off-grid power station which is currently poisoning Black people in Memphis, Tennessee (Kerr, 2025), so that Twitter's genAI application Grok can rant incoherently about "white genocide" (Steedman, 2025). More generally, as someone who would prefer to avoid getting killed for my food and water in a few decades' time, I am unpleasantly reminded of the study that found that bitcoin mining emissions alone could make runaway climate change impossible to prevent (Mora et al, 2018). GenAI is rapidly scaling up to produce similar amounts of emissions, with the same consequences, for the same reasons (Luccioni, 2024). [2]

It is theoretically possible to create genAI which doesn't steal and which doesn't destroy the planet. Nobody's going to do it, and if they do do it, no significant share of the userbase will migrate to it in the foreseeable future — same story as, again, bitcoin — but it's theoretically possible. However, I also advise against genAI for any application which requires facts, because it can't intentionally tell the truth. It can't intentionally do anything; it is a system for using a sophisticated algorithm to assemble words in plausibly coherent ways. Try asking it about the lore of a media property you're really into and see how long it takes to start spouting absolute crap. It also can't take correction; it literally cannot, it is unable — the way the neural network is trained means that simply inserting a factual correction, even with administrator access, is impossible even in principle.

GenAI can never "ascend" to intelligence; it's not a petri dish in which an artificial mind can grow; it doesn't contain any more of the stuff of consciousness than a spreadsheet. The fact that it seems like it really must know what it's saying means nothing. To its contemporaries, ELIZA seemed like that too (Weizenbaum, 1966).

The stuff which is my focus on this blog — untraining and more broadly AB/DL in general — is not inherently dangerous or sensitive, but it overlaps with stuff which, despite being possible to access and use in a safe manner, has the potential for great danger. This is heightened quite a bit given genAI's weaknesses around the truth. If you ask ChatGPT whether it's safe to down a whole bottle of castor oil, as long as you use the right words, even unintentionally, it will happily tell you to go ahead. If I endorse or recommend genAI applications for this kind of stuff, or assist with their use, I am encouraging my readers toward something I know to be unsafe. I will not be doing that. Future asks on the topic will go unanswered.

Notes

I use quote marks here because as far as I am concerned, both "generative artificial intelligence" and "genAI" are misleading labels adopted for branding purposes; in short, lies. GenAI programs aren't artificial intelligences because they don't think, and because they don't emulate thinking or incorporate human thinking; they're just a program for associating words in a mathematically sophisticated but deterministic way. "GenAI" is also a lie because it's intended to associate generative AI applications with artificial general intelligence (AGI), i.e., artificial beings that actually think, or pretend to as well as a human does. However, there is no alternative term at the moment, and I understand I look weird if I use quote marks throughout the piece, so I dispense with them after this point.

As a mid-to-low-income PC user I am also pissed off that GPUs are going to get worse and more expensive again, but that kind of pales in comparison to everything else.

References

Brown, T.B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., ... & Amodei, D. (2020, July 22). Language models are few-shot learners [version 4]. arXiv. doi: 10.48660/arXiv.2005.14165. Retrieved 25 May 2025.

Changes on Tumblr (2023, August 29). Tuesday, August 29th, 2023 [Text post]. Tumblr. Retrieved 25 May 2025.

Hoskins, P. (2025, January 8). ChatGPT creator denies sister's childhood rape claim. BBC News. Retrieved 25 May 2025.

Ingram, D. (2024, June 13). Elon Musk and SpaceX sued by former employees alleging sexual harassment and retaliation. NBC News. Retrieved 25 May 2025.

Kerr, D. (2025, April 25). Elon Musk's xAI accused of pollution over Memphis supercomputer. The Guardian. Retrieved 25 May 2025.

Luccioni, S. (2024, December 18). Generative AI and climate change are on a collision course. Wired. Retrieved 25 May 2025.

Mora, C., Rollins, R.L., Taladay, K., Kantar, M.B., Chock, M.K., ... & Franklin, E.C. (2018, October 29). Bitcoin emissions alone could push global warming above 2°C. Nature Climate Change, 8, 931–933. doi: 10.1038/s41558-018-0321-8. Retrieved 25 May 2025.

Steedman, E. (2025, May 25). For hours, chatbot Grok wanted to talk about a 'white genocide'. It gave a window into the pitfalls of AI. ABC News (Australian Broadcasting Corporation). Retrieved 25 May 2025.

Weizenbaum, J. (1966, January). ELIZA—a computer program for the study of natural language communication between man and machine. Communications of the ACM, 9(1), 36–45. doi: 10.1145/365153.365168. Retrieved 25 May 2025.

Zitron, E. (2024, April 23). The man who killed Google Search. Where's Your Ed At. Retrieved 25 May 2025.