Newsletter: I am sick and god damn tired of everybody pretending that generative AI is the next big thing. The media is complicit in accepting fantastical nonsense - both in the numbers put out by OpenAI and the silly jobs created by Anthropic - and it has to stop.
https://www.wheresyoured.at/reality-check/
https://www.wheresyoured.at/reality-check/
Comments
You mean the same media that drove us for years into platforming/legitimizing MAGA?
We are in a doom loop built by American ignorance & hubris.
Now, we've ignored all ladder history and never read a book but we estimate $2.2T in market share!!
Tell me what to say when I defend you online against all critics with a weird amount of fervor...
GenAI is a bubble, capitalism's last hope in these dire times. Maybe its the last bubble ever, who knows 🤞
Maybe not right now, but when the AI bubble is over they'll just go and get excited about something else. I'm not even sure they have a choice.
In scamming.
Meaning: none of your competitors are *not* doing AI. There's no "what if we lose to non-AI company Q?" There's no such competitor. So nobody cares if it turns out to be nothing.
the VC shops (that actually had good returns) closed up just before cloud craze as the writing was on the wall already :^)
Roose then tugs on the fake beard he is wearing and it makes a sound like a train whistle.
https://www.wheresyoured.at/reality-check/
https://www.wheresyoured.at/reality-check/
https://www.wheresyoured.at/reality-check/
https://www.wheresyoured.at/reality-check/
https://www.wheresyoured.at/reality-check/
https://www.wheresyoured.at/reality-check/
It's just pay to play public relations and puffery now
And yeah it's going to stick around because it's actually valuable for specialised jobs and not just a statistical text generation engine that people push into all sorts of things that it's not remotely suitable for!
Sensors, object recognition and predictive path finding *could* solve the self-driving problem. They're just stupidly optimistic with when.
"Stringing words together based on likely occurrences" only solves "create plausible but often wrong text"
Man, that is not what an ER does. And have you been to a hospital or seen the state of healthcare in the US? AI ain't fixing all that.
Delusional from the jump.
https://www.darioamodei.com/post/on-deepseek-and-export-controls
thx again for your hard work, covering this shit must be a huge pain in the ass
I don't get it.
Deepseek wouldn't be profitable either. It would be cheaper, but even at the significantly reduced costs people just don't want to pay to use AI as it is right now. That's when their models all include a bunch of uninvented tech that will save them and assumes endless improvement.
But they had to say they invented god.
But Ed's main point is that AI isn't profitable, OpenAI loses money with every single query, even from paid customers. Deepseek's costs are a lot lower, so why isn't OpenAI copying their methods?
This is also Ed's view.
I mean, if the guy lies about his own voice then you really don't want to buy anything he sells because its all woo-woo snake oil bollocks.
Ideally, they would charge everyone, EVERYONE, $200/mo to use it 5 times per month, and even then I'm not 100% positive they'd be able to profit.
…
https://www.theregister.com/2025/04/29/generative_ai_no_effect_jobs_wages/
Ed is 100% correct about this.
What often gets overlooked: it’s not just bad internal decisions and ads — SEO and fake reviews have been poisoning Search for years.
LLMs will face the same fate, just faster.