The statement that data centers are being built with cash on hand is inaccurate. In the fourth quarter of last year alone, tech companies raised over $107 billion through bond issuances.

This was a great talk.

I am sure we will never achieve AGI, forget AGI we still don't know how a single cell works and we are talking about creating AGI.

09:15 Yeah: Sausage or not sausage :)

Just wanted to say that this “trough of disillusionmen graph/system (34:30) is rather probably nonsense. You can get very rich if you get very rich if you get mentioned by the one company that espouses this though. There is that.

I saw this

49:43 "When you get 60B $ of investment you're basically guaranteed to make bad choices."

That orchestrator over multiple specialized LLMs would reliably pass the Turing Test but would not be thinking. So I suggest the Turing Test is not a good test for thought. And I don’t know of a better test.

don't forget OpenAI defines AGI in their charter as "highly autonomous systems that outperform humans at most economically valuable work" - it's all marketing, baby.

I don't know bro. In the right hands people are doing some great things with the models. Not sure if dud is right context

Great talk ....
In situ ..all systems are and or required to be on line - on apps , data bases - stations the collective 1 "  One "
Fibre optics - wii fii ....weighing down the clouds  of electical ....bio electric nuclear thorium plasmas ....detrimental to Trinidadian  split divide of the ever present daughters of toxic beginnings ...till theres nothing but baron rare earth scrapings like tapestry of colors  the machine requires  and wants  ...
All till theres nothing but the black hole , that intelligent minds so hell bent on advancements became the enablers of destruction ...without knowing or choosing to ignore ...what good is and means from the heart and love that can not be calculated cloned or simulated ....only a machine  allowed to dominate needs everything and everyone ..
unwise unintelligent unstable ...stop

Thanks AI for the following summary!

Historical Context and Marketing

Artificial Intelligence was coined in 1955 as a marketing term to secure funding, leveraging the human tendency to project intelligence onto software.

The industry has cycled through multiple "AI Winters" and periods of overinflated expectations.

The LLM Explosion

Modern AI progress was driven by massive scaling and Microsoft’s investment in OpenAI to create heavy workloads for Azure.

We are likely passing the peak of the Gartner Hype Cycle and entering a "trough of disillusionment" as newer models show diminishing returns.

Impact and Ethics

Developers are well-positioned to use AI because they can treat it as a peer review process with automated validation.

Significant risks include economic bubbles, the erosion of visual truth through deepfakes, and the potential for "acquiescent" AI to reinforce human biases.

Richard Campbell emphasizes that while AGI is not here, developers must take ethical responsibility for how these powerful automation tools are deployed.

Had to stop watching at 18:52 when Richard came along with the history of AI until 2023 without mentioning "Attention is all you need". A pivotal paper which changed the course of how OpenAI did things internally and Andrej Kharpathy did mention this in his own talks and videos. So this one though must be great is a hard pass.

Stop the alarmist rhetoric already. Alcohol, Tobacco, Sugar, various drug addictions,  etc..., are killing and disabling a lot more people than AI ever will. We are talking about millions of people while AI is maybe - maybe - responsible for nothing but a handful of catastrophic disasters.