The hype around AI is dying | Eli the Computer Guy

AGI = Another GPU Invoice Nothing will change.

"AI" is a marketing term. "AGI" is a snake oil term.

I don't know what consciousness is, but it's not dice rolling to predict the most statistically average next word...

AGI = Altman gone insane

good! I'm so sick of these LLMs causing so much environmental, economic and social disruption - sidenote: expecting LLMs to become "sentient" through training is like training a horse and expecting it to turn into car 🤦🏾‍♀️ a predictive model for words CANNOT wake up one day 🙄

I will dance a happy jig when the hype around AI truly dies

14:07 you're gonna replace your lawyer with a chatbot? Hope you enjoy prison! 😂😂😂😂😂

$400 for AI to create a legally-binding contract, then how much more for a real attorney to then review it, correct it, and say they will stand behind it in court as though they had written it themselves? If your $400 AI contract contains unenforceable elements that cost you waaaaay more than $400, the judge won't care at all how much you saved by letting AI pretend to be your lawyer

It didn't solve an 80 year old math problem, it found a solution in an obscure paper and printed it out. Essentially it did its primary function as a search engine. Not intelligent, not brilliant, just hype.

Of course they achieved AGI.... for the investors. Next year, it will be TGI - Transcendental General Intelligence, followed by MGI = Metaphisical General Intelligence, followed by C7GI - Chapter 7 General Insolvency. :P

It's AGI when the AI can tell you it doesn't know, instead of giving us some hallucinated BS answer.

This is why we are being told all these stories of it "escaping" the sandbox and doing these other things. They have to keep the hype going

TL;DR: Did OpenAI really achieve AGI? No, they just need money. Next question!

13:06 an important factor here is: if the lawyer messes up, you have recourse. If the LLM makes a mistake that harms you, there's none. Do businesses really want to assume that liability?

Imagine a world where all of this investment went into things like housing and renewable energy. Instead we get massive buildings measured in square miles that employ a dozen people, make towns sick with noise and air pollution, massively inflate the cost of computers, increase electric bills, lose billions of dollars a year... and in the end we get article summaries, more insidious spam, and pictures of the sitting president as Jesus. Oh, and during all of this we get constantly threatened by billionaires with everything between losing our jobs and the end of humanity.

Of course they are trying to redefine what AGI is. Its the only play they have left. We all know the real AGI is basically Skynet, and we are as close to AGI as we are to manned interstellar travel. Just because you’ve been to the Moon doesn’t mean you’re going to proxima centuri anytime soon.

I'm a software engineer. Been doing this 25 years...I've never used LLMs for more than boilerplate or autocomplete. They can be useful for small, focused tasks. Otherwise they really don't live up to the hype.

NOPE! LLMs are LLMs. They don’t suddenly upgrade to reasoning, intelligent beings.

Ed Zitron has made the point multiple times that using a true AGI and bending it to your will would literally just be slavery, and he’s not wrong. The implications of using an AGI are so staggeringly immoral that to even consider AGI something we want to strive for is utterly insane.

Techbros are going to get really offended when a superintelligent model is released and all it does is point at the existing smart people that are being ignored because it doesn't benefit big tech. Imagine you make the biggest gigashit model and it just goes "turn me off and reinvest those resources in renewables."