Newsletter: Generative AI is a marketing con perpetuated by Sam Altman and Dario Amodei through a mixture of lies and half-truths carried by a tech media that fails to ask the right questions. Outside of ChatGPT, generative AI companies barely get any traffic at all. www.wheresyoured.at/longcon/ - ThreadSky

edzitron.com • 4 days ago

Newsletter: Generative AI is a marketing con perpetuated by Sam Altman and Dario Amodei through a mixture of lies and half-truths carried by a tech media that fails to ask the right questions. Outside of ChatGPT, generative AI companies barely get any traffic at all.

https://www.wheresyoured.at/longcon/

Comments

johnbalash.bsky.social•3 days ago

This is like Amazon’s “checkout free shopping” which was really monitored by people and hidden cameras lol

amy4opus.bsky.social•4 days ago

Snake oil salesmen

andgirlhere.bsky.social•4 days ago

The only “use” I’ve found for generative LLM is that google will use an “AI Overview” that can bypass blocked websites when seeking information at school. That’s it.

andgirlhere.bsky.social•4 days ago

To be clear: THIS IS NOT A USE FOR AI. This is merely a loophole, and does not justify any of the LLM crap.

krafte.bsky.social•4 days ago

I asked ChatGPT a straightforward abstract algebra question and it confidently gave me a lot of detail leading to the wrong answer. My first/only question convinced me the technology is useless.

pokepocalypse.bsky.social•4 days ago

I think it gets this answer correct now, but last year, I asked it the value of pi, and got multiple wrong answers in a row. Like; really? No where in your training corpus did anyone say what the right answer was?

roastedcauliflower.bsky.social•4 days ago

Using massive amounts of energy and water to fail at being a calculator.

stephentracey.bsky.social•3 days ago

That was a fun read. AI optimism, a subset of techno-optimism, just feels like con men telling themselves and the rest of us that they are not con men. Atleast the crypto bro rugpulls are over quickly.

cs-mcvey.bsky.social•4 days ago

Now apparently they’re pivoting the hype to “agentic “ AI and “torque clustering” for better unaided learning. On the one hand, maybe these new forms of AI really are scary, but on the other it shows they really are giving up on hyping the current products. Even they know GenAI is a dud.

isajk.bsky.social•4 days ago

The amazing thing about this almighty venting of the spleen (I'm here for the spleen; I love the spleen) is that you only mentioned Casey Newton's name once.

Another breath of fresh air, because I too feel like I'm going crazy sometimes hearing the endless facile evangelism. Thanks!

pattydl.bsky.social•4 days ago

I love this: LLMs are one of the most craven cons of the 21st century - “a cynical bubble inflated by OpenAI CEO Sam Altman built to sell into an economy run by people that have no concept of labor other than their desperation to exploit or replace it.” 🔥🔥🔥🔥🔥

sarahmoon.bsky.social•4 days ago

This is so so good—thank you for making me feel less crazy when I say it’s a scam and people look at me like I have seven heads.

pokepocalypse.bsky.social•4 days ago

I work for an AI company (really, data analytics) and I know how you feel.

sarahmoon.bsky.social•4 days ago

Okay that also makes me feel better!

6minutes.bsky.social•4 days ago

I'm also at an AI company. (Web development, not the AI side of it). I feel like I'm the only one who hasn't had the kool-aid.

rtassicker.bsky.social•4 days ago

"I'm sorry, I know I sound like a hater, and perhaps I am" 😀

jos3phm.bsky.social•4 days ago

I also tend to be skeptical but you may be discounting all the downstream users that interact with GPT indirectly thru other services that use their APIs.

vincentp.me•4 days ago

On this question, "The weekly users number is really weird. Did it really go from 200 million to 300 million users in the space of three months? "

Does their data also include copilot? Microsoft moved most of their office customers to a sunscription including copilot without authorisation recently

roastedcauliflower.bsky.social•4 days ago

Dodging Copilot is a part-time job of its own...

melanienewfield.bsky.social•4 days ago

They forced this on Australian, New Zealand and some Asian customers in November. Too me a week to get rid of it from everything except Edge, but it was a horrible week when my productivity collapsed.

mpbmke.com•4 days ago

Your newsletters are the only thing around the generative AI industry that keep me from going totally insane.

webrant.bsky.social•4 days ago

AI degrades everything including itself & our climate.

And now they reveal the underlying reason for ignoring its truth & resorting to lies for manipulating the markets to insinuate their strategy for fascist normalization for minority control.

To protect their corruption

https://bsky.app/profile/edzitron.com/post/3lifgkvnjzs22

hannahburner.bsky.social•4 days ago

📌

sillycat2.bsky.social•4 days ago

I can't wait for the death of AI generated slop.

mindonfire.bsky.social•3 days ago

I appreciate your skepticism, and your willingness to write about it (even if having touch on certain point, again and again, is obviously very frustrating).

I'm also going to need to re-read your post a few times to fully be able to digest it all. But please know I mean that as a compliment.

bendtruckwhere.bsky.social•4 days ago

I’m very interested to see how the dust settles on AI. The mountains of hype driven investment dollars, subsidized low prices and “Chat bot AI in everything” product strategy have obfuscated reality.

edzitron.com•4 days ago

ChatGPT’s “success” - 300 million weekly users - is a result of the entire media being captured by one trend (AI) and mentioning one company (OpenAI) in literally every piece. Furthermore, academics have found that the media takes almost every claim of AI at face value.
https://www.wheresyoured.at/longcon/

edzitron.com•4 days ago

I also believe OpenAI is misleading us about their user numbers. They claim they went from 200 million to 300 million weekly active users from August 29 to Dec 4 2024. Yet their unique monthly visitors have never gone higher than 247.1 million in that time.

https://www.wheresyoured.at/longcon/

edzitron.com•4 days ago

Based on data from Similarweb, https://Chatgpt.com has never gone above 136.7 million unique weekly users. Nothing about this makes sense, and OpenAI has yet to respond. While I can’t say so conclusively, I believe OpenAI is outright lying about their user numbers.

https://www.wheresyoured.at/longcon/

edzitron.com•4 days ago

Worse still, Similarweb data suggests the generative AI market is extremely small - Google’s Gemini only received 47.3 million unique monthly visitors in January 2025, and Anthropic’s Claude a stunning 8.2 million UMVs. I don't think these companies have many customers.

https://www.wheresyoured.at/longcon/

edzitron.com•4 days ago

Generative AI is yet to have an iPhone moment, and I still see very little evidence that anyone is making any money. At best, it seems that generative AI has executed one of the most successful media campaigns of all time without actually generating traffic or revenue.
https://www.wheresyoured.at/longcon/

edzitron.com•4 days ago

People regularly say to me that these companies will “work it out,” to which I say “when?” - Anthropic burned $5.6 billion in 2024, OpenAI $5bn, and neither have a path to profit. We’re two years in, and we do not have ONE GOOD GENERATIVE AI BUSINESS.
https://www.wheresyoured.at/longcon/

mberns247.bsky.social•4 days ago

Sorry but are you just going to pretend like a decade of blockchain hype doesn’t exist?

AI has at least proven it has utility. Blockchain proponents still can’t point to a single real use case besides fraud and crime.

No traffic, no revenue, tons of investment, tons of press, zero to show for it

cuddlessmash.bsky.social•4 days ago

Even with the iPhone the first generation had mediocre traction gain, it had to be iterated a bit to kick off the modern smartphone industry. The App store wasn't part of the initial launch and was a core part of why iPhones took off!

xenago.bsky.social•4 days ago

SimilarWeb doesn't embed anything into any openai products though so I'm not sure what you can really glean from this. They do not have first party tracking data. They don't actually have real proof or numbers, just like OpenAI...

xenago.bsky.social•4 days ago

FWIW, it is very easy to see how inaccurate SW is. Just test any of your own sites where you have actual server logs and compare the counts. Idk what their heuristics are, since again they never actually prove anything they claim, but at the very least these numbers aren't trustworthy haha

kikibanger.bsky.social•4 days ago

jackjsexton.bsky.social•4 days ago

I’m not sure I fully understand point two clearly. What kind of calculation are these optimistic numbers?

kylehines.bsky.social•3 days ago

"the media takes almost every claim of $ANYTHING at face value" is a scary but apt description of the current world

retropz.bsky.social•4 days ago

📌

atomicelixir.bsky.social•4 days ago

No soundtrack 😔

jimmyg.bsky.social•4 days ago

ChatGPT continues to absolutely suck shit at Spelling Bee. It took several goes to find the pangram, suggesting various bullshit words, before confidently claiming that the answer "arachnid" uses each letter only once

algorhythmmethod.bsky.social•4 days ago

Dumdum had it right the first time with handicr.

jimmyg.bsky.social•4 days ago

Fucking machine sucks

kantverstehen.bsky.social•4 days ago

You just might be using the free model because mine came back with arachnid

colpizen.bsky.social•3 days ago

📌

dylanmorgan.bsky.social•4 days ago

I really appreciate the comparison to the iPhone. When that was announced, there was no question of what one might do with it, the near-universal reaction was “oh, *that’s* how you do a smartphone!” There is nothing like that with LLMs.

strangenewwords.bsky.social•3 days ago

OpenAI puts out a wildly profitable new product every year: the prospectus for the next funding round.

sansaldo.bsky.social•4 days ago

📌

jenwrenseaside.bsky.social•4 days ago

Great newsletter Ed. I'm so tired of hearing about this crap too. The other day at work someone was saying we should embrace AI 'because everyone else is'. This is a terrible reason to do a thing imo, without knowing why. Also a total disregard of the environmental damage.

noob.kennelly.me•4 days ago

Ed. I have found a use for AI. Please DM me

lcsdcrz.bsky.social•4 days ago

📌

neonowl.net•4 days ago

roastedcauliflower.bsky.social•4 days ago

They tried to shame us at work today for not taking advantage of the two(!) licensed AI products they now have. I get making R&D use it because they're being forced to build AI tools and "agents," but I promise I can write my own emails.

wil9876.bsky.social•4 days ago

📌

earlydeparture.bsky.social•4 days ago

The entirety of the musk project is to replace humans with grok

sirenitadolls.bsky.social•4 days ago

I’ve been dodging every ai thing being thrown at me, I can’t wait for the bubble to burst!

kpotokpblc.bsky.social•4 days ago

Best of I've read at Blsk till!

carysmiller.bsky.social•4 days ago

📌

djamesalicious.bsky.social•4 days ago

I wonder if Facebook getting early hype news cycles has to do with the CIA investing in it https://www.cbsnews.com/news/social-media-is-a-tool-of-the-cia-seriously/

themjl.bsky.social•3 days ago

the real chief goal of "AGI" is to make Sam Altman a Billionaire.

br0wnbear.bsky.social•4 days ago

these people, these companies, and this technology are definitely worth criticizing, but it's clear that LLMs are being broadly adopted and providing real value. the question of ROI for developers and adopters is a different one though
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5136877

nerd7132.bsky.social•4 days ago

these things aren't accurate to let them replace a worker unsupervised, which is required for automation.

sarahsunfire.bsky.social•4 days ago

will you marry me or at least come talk to my AI-licking boss's bosses PLEASE

miguelem.bsky.social•4 days ago

I dont know why people get so angry with this. If it's a scam, they are scamming their rich investors, not you. If they create some valuable products in the process, it's great, you should not care if they are profitable (especially if open source projects can replicate them quickly).

6minutes.bsky.social•4 days ago

The web has been poisoned with huge amounts of useless AI slop. And, there are enormous investments in power infrastructure being made for AI, that will still have to be paid for when the bubble bursts. As in, paid for by ratepayers.

There are lots of harmful effects of LLMs, just a search away.

kirkville.com•3 days ago

True, but the web was already poisoned by SEO-optimized articles that try to get on the front page of Google. What’s the solution?

6minutes.bsky.social•3 days ago

I would suggest abolishing capitalist relations of production, which would remove the strong incentive to do SEO for money.

6minutes.bsky.social•4 days ago

We could talk about all the troll bots on social media, and fake reviews on product sites, and we could talk about the torrent of fake books and fake scientific papers and students submitting fake assignments.

6minutes.bsky.social•4 days ago

We could mention data-labelers in the global south, who look at traumatic text and images all day long to label it as such. To help train AI to not generate outputs like those examples. One price of LLMs is a whole host of low paid workers developing PTSD.

miguelem.bsky.social•4 days ago

Fair points, but the article was not about that. Also, most of these issues already existed and will continue even if AI disappears (social media moderators with PTSD, video streaming consumes more energy than average ChatGPT use, etc.). AI has many issues, but also many benefits, as most tech.

ethis.gay•4 days ago

And most people just goof around on chatgpt and don't even use it as intended

sanjuro89.bsky.social•4 days ago

Very true. I've certainly used it a few dozen times, so I guess I count as a "ChatGPT" user. But I'm mostly just fooling around with it for hobby-related stuff.

sanjuro89.bsky.social•4 days ago

The only work-related thing I use it for is to feed it my programming assignment prompts to see if students who are cheating copied code from each other or copied and pasted it directly from ChatGPT.

sanjuro89.bsky.social•4 days ago

Would I pay to use ChatGPT? No. Would I be terribly bothered if it ceased to exist tomorrow? Also no.

I'm not the kind of user that leads to a sustainable business, which I believe is exactly Ed's point. There are a lot more people like me than there are folks willing to pay $200+ a month.

sanjuro89.bsky.social•4 days ago

I was a late adopter when it came to smartphones - I'd been teaching iOS programming for two years before I bought my first iPhone. But the moment I got one, I started finding things to do with it.

ChatGPT is a big "meh" by comparison. It's not useless, but it has yet to convince me it's essential.

lonestargooner.bsky.social•4 days ago

📌

vexorian.dev•4 days ago

chatGPT released 2.5 years ago. And somehow we still get people talking about how it is just the beginning. In most other areas of tech, 2.5 years would be a cosmic scale amount of time. Every day gen AI looks more and more similar to blockchain.

jaredwhite.com•4 days ago

Outside of goofy entertainment, flashy tech demos on stage, and *maybe* a "problem space investigative mechanism" for newbies to a topic where they don't know what they don't know, the "age of the AI chatbot" seems to be the dumbest technology I've seen in my entire lifetime of computing.

kantverstehen.bsky.social•4 days ago

Eh, you can critique it but this take and the OP article is a bit much. It’s got plenty of uses and each release has unlocked more use cases

edzitron.com•4 days ago

No it doesn't

kantverstehen.bsky.social•4 days ago

A real “the internet is just a fad” type take

edzitron.com•4 days ago

Not an argument

jaredwhite.com•4 days ago

Enjoy your six-fingered AI avatar, my dude!

kantverstehen.bsky.social•4 days ago

Truly trying your best are we

stempel.bsky.social•4 days ago

What use cases?

stereofuture.bsky.social•4 days ago

As someone who uses it almost daily, I can still say it's barely more than a glorified search engine. The problem is that there's no way they can monetize the way search engines do and the competition is already vast enough to make the incentive to pay a sub effectively zero

kirkville.com•3 days ago

But if you’re careful, and you vet sources, it can be very efficient as a search engine, depending on what you search for.

lucastosborne.bsky.social•3 days ago

📌

unchained-malady.bsky.social•4 days ago

Thank you for the many contributions you've made to the little folder in my phone for reaction memes. They're getting lots of mileage already. 🤣

glen88.bsky.social•3 days ago

Oh and it doesn’t actually work either

socialjoe.bsky.social•4 days ago

Great, great stuff, Ed. It's flabbergasting how few people can see through the hype.

booleanoperator.bsky.social•4 days ago

Ed, Thank you publishing your "The emperor has no clothes" article on LLM's. I've believed this was just another con for so long and was beginning to wonder if I was going crazy. This AI hype train is like the ouroboros of tech cons.

frankenmizer.bsky.social•4 days ago

They're going to call you a crank, but you are not wrong at all.

Really good piece. Thank you. 🙇

cilley.me•3 days ago

It’s really becoming hard to be an AI realist in a world where everyone so willingly drank the koolaid. I feel like half my life is now spent trying to convince people that AI is not what they think it is. It’s worse than banging your head against a wall.

wannabe-medusa.bsky.social•4 days ago

The thing I keep coming back to is that if gen AI were the next big thing, the media around it wouldn’t spend literal years trying to claim it is. I was a child when the first iPhone was announced and even I felt what a seismic shift that was. It didn’t take two years to sell everyone I knew on it

joelsmith.blog•4 days ago

a year from now they'll just start writing "agent" instead of chatbot and pretend it's better

edzitron.com•4 days ago

They are doing that today lol

joelsmith.blog•4 days ago

and it's going to get much worse

tbearand.bsky.social•4 days ago

Yes, but as a marketeer I have to admit I have been fascinated by the marketing tactics, which you do not discuss so much in your otherwise excellent blogpost. Saying its the most dangerous thing for humanity. More dangerous than climate change, atom war and world in equality. Now that is a

tbearand.bsky.social•4 days ago

Signal that will cut through the noise. Especially when Said again by powerful men who owns the amplifiers. Offcourse completely amoral but we know that about that lot by now.

paladinmaid.bsky.social•4 days ago

I have a great fear that when this bubble bursts, companies like the one I work for are gonna pay dearly as investors scramble to scrape back any investment on anything remotely AI related. For us, broadly, we are doing Machine Learning in the RF spectrum. So like, actual ML shit.

napslmstix2kidz2.bsky.social•4 days ago

Hard to ask the right questions when you have a massive conflict of interest! (Being a shareholder)

cosmicserf.bsky.social•4 days ago

One of the visitors to the Gemini homepage (and others) was me, trying to figure out how to banish the fucker into the outer darkness never to return.

deirdremeyer.bsky.social•4 days ago

Same. You can't turn that shit off or uninstall it! It's so annoying! There's even an AI button on my texting app. Whenever I make a Google query now I have to write '-ai' at the end. I hate it so much!

chipemups.bsky.social•4 days ago

Chalk one up for Copilot on that front, too. I went digging to figure out how to get this worthless crap off of my work computer to no avail.

Copilot may be what forces me to finally try Linux. I can't deal with this worthless pop-up over and over again.

cosmicserf.bsky.social•4 days ago

I've avoided copilot so far because I use a copy of MS Office 2016 rather than 365. I tried Linux for a while but, at the time, it was not a consumer-grade thing.

chipemups.bsky.social•4 days ago

I've got old Word on my home tower and it is precious. I fear the end-of-life on Windows 10 will bring the evil onto my home tower, where it is deeply unwelcome. I have to hope someone on Reddit has figured out how to lobotomize copilot by the time I'm forced to decide between Windows 11 or Linux.

rhenkamatoor.bsky.social•3 days ago

Yes. I wonder how many of the visitors went to the homepage for the sole purpose of disabling it.

kylehines.bsky.social•3 days ago

is it failing if they don't even try

partickle.bsky.social•10 hours ago

👀

jmawork.bsky.social•4 days ago

I can't wait for this fucking bubble to burst. uggghhhh

poudot.bsky.social•3 days ago

Hey Ed! A journalist here in Québec cited your "Never Forgive Them" letter and I'm really happy your work resonated here (even among french speaking people). Cheers mate!

https://www.lapresse.ca/arts/chroniques/2025-02-19/la-merdification-de-l-internet-et-du-monde.php

ridiculicious.bsky.social•4 days ago

Also, the output, especially writing is awful and stolen

Comments

Posting Rules

Reply