This is the most financially dangerous thing we have talked about for a very long time because this is kryptonite. You know, the stuff that can take down Superman and superman in our little world is obviously Nvidia. This is a fuse being lit and no one has been paying attention, but now just now we are getting tidbits. I’ve been yakking about it for some time now, but it is finally starting to dawn on some people and they are starting to talk, just on the fringes but it is starting.
We go now into the latest bubble, the AI bubble. Tristan Greene writes for Cointelegraph and his piece is titled, EU says ChatGPT outputs too much false information to comply with rules. He writes, “While the assessment remains ongoing, OpenAI appears to have made little in the way of progress since 2023…The chief complaint appears to be that ChatGPT is prone to outputting inaccurate information. “As a matter of fact, writes the EDPB, “due to the probabilistic nature of the system, the current training approach leads to a model which may also produce biased or made up outputs. “ No kidding. The interesting thing to me is that they have been trying for a year and can’t make it better. It is a regurgitator not a contemplator. That is the fundamental limitation.
Greene writes another article this time about Elon Musk’s proclamations about the dangers of AI. Musk doubled down saying that AI would replace humans and do everything better and that humans would lack meaning in their life. My favorite part of this article is when Greene mentions that Musk is raising money for his own AI company. He also mentions “It bears mentioning that Musk’s AI-related predictions haven’t always fared so well. In 2019, he famously promised that Tesla would field 1 million fully autonomous robotaxis on the road by 2020.” The grift is strong in this one.
From hype to actual AI information, in the Dell earnings call, Toni Sacconahi, an analyst from Sanford C. Bernstein & CO, LLC, asked why AI server revenue went up $1.7 billion but operating profit was flat and margins on AI servers effectively zero. This kicked off numerous questions and Dell executives dancing around the question. Well, the market responded, and Dell’s stock dropped 17.9% that Friday despite large revenues for AI servers. The profit wasn’t there folks. I have been saying for a very long time, the profits are coming where to support these valuations? I know people are buying AI chips from Nvidia, but who is monetizing AI? Without monetization somewhere somehow, this bubble bursts and bursts spectacularly. If ChatGPT, the flagship that started this boom last year, can’t get its house in order and Elon can be trusted about as much as Don T. in the eye of a storm, then what chance does this industry have besides spending money for a little while longer? If you believe what they are telling you, it’s the future. If you choose to believe them fine, I’m sure they, like Dell, don’t have a vested interest in telling you that.
I personally think AI could serve a purpose, that of the greatest regurgitator of all time. Instead of Google we get the regurgitator. It actually has the ability to grab a lot of search items and recall them. It isn’t thinking and it probably needs help ordering them, but the recall power to sift through data sets would be akin to having a photographic memory, one that is really close to perfect. So close that when it is wrong no one expects it or probably even notices. This is why Google seems like the one major company that could be affected by this technology. I think code writing, language, and logical operations are obvious areas that raw computing power can provide benefit. Emily Bary had a very unique insight in MarketWatch when she asked Rosenblatt Securities analyst Barton Crockett why investors should take a pause on investing in Google and it was his last point I found most interesting, “Finally, he mentioned that Alphabet could wind up in a “higher-than-anticipated [capital-expenditure] spending cycle for AI” as companies feel competitive pressure to ensure they’re well positioned on the technology.” This is a high stakes game of keeping up with the Jones’ and all the big boys are playing. The chips are the way to keep your stock price high, the mention of AI gets your stock to jump, and some seem to think that it will never come back to earth. Like the Challenger flight, it will land sooner than you think.
Ken Griffin from Citadel also thinks so. If you don’t know who he is, he is one of the smartest traders on earth. He is amassing a fortune in financial trading and very smart. He in essence started a hedge fund in college at Harvard and did so so he could short the market in 1987 and profited handsomely. He is worth more than $30 billion
He is so smart he moved his whole operation from Chicago to Florida. However, he too is a man of shaky morality. In the great financial crisis he wouldn’t let people have their money of out his fund. He was leveraged and if he would have started liquidating the whole damn thing could have fallen down. Don’t be surprised if we don’t see this again, but maybe Ken learned his lesson.
However he is very connected and very astute and here is his latest quote on AI from Yun Li’s report on CNBC, “We are at what is widely viewed as a real inflection point in the evolution of technology, with the rise of large language models. Some are convinced that within three years almost everything we do as humans will be done in one form or another by LLMs and other AI tools,” Griffin said Friday during an event for Citadel’s new class of interns in New York. “For a number of reasons, I am not convinced that these models will achieve that type of breakthrough in the near future.” I think near future is exactly right. The technology is nowhere ready for prime time thinking and perhaps that is a good thing. He continues, “Machine learning models do not do well in a world where regimes shift. Self-driving cars don’t work very well in the North due to snow. When the terrain changes, they have no idea what to do,” Griffin said. “Machine learning models do much better when there’s consistency.” Garbage in and garbage out, they basically can’t adapt to anything that they haven’t been fed. How do you think they would have responded to to 9-11?
Jeffrey Funk and Gary Smith write for Marketwatch and they have another insight, it is systemic. “The fundamental problem with ChatGPT and other large language models (LLMs) is that they do not understand what words mean. They are very much like a young savant who can recite every word in all six volumes of the History of the Decline and Fall of the Roman Empire without comprehending any of the content. Without such comprehension, LLMs are not going to morph into artificial general intelligence (AGI) — the ability to perform any intellectual task that human beings can do.
Many AI enthusiasts — including Tesla’s TSLA Elon Musk, Jensen Huang of Nvidia NVDA and pioneering AI researcher Ben Goertzel — nonetheless claim that AGI is just a few years away. This cheerleading certainly helps raise funds (just ask ChatGPT’s Sam Altman) and sell computer chips (just ask Nvidia) but it is increasingly recognized that the breathless hype is another case of Silicon Valley’s “fake-it-til-you-make-it” mentality…It is easy for startups and major players to make seductive predictions and promises about the Next Great Thing.”
Bingo, these people are hope trafficking and selling it. They have seen a few people do this and now the room is full of imitators. For every Bezos that faked it until he maked it there are Elon Musks promising self-driving and miracle cars. They are so miraculous that it seems there are more brands of EV cars making a car every bit as good as he does than when I was a kid making internal combustion engines. When the public stops believing these sales hypes the stock of Tesla goes from 220 to 20. Volkswagens stop buying Rivians, we stop investing in wind turbines at sea, and for goodness sakes stop building AI data centers to suck up all the electricity. We stopped listening to Al Gore on the environment and its time we start thinking and reading for ourselves on all this other nonsense as well.
I keep warning about the malinvestment on this and here is Christopher Mims writing for the Wall Street Journal, “A mature technology is one where everyone knows how to build it. Absent profound breakthroughs—which become exceedingly rare—no one has an edge in performance. At the same time, companies look for efficiencies, and whoever is winning shifts from who is in the lead to who can cut costs to the bone. The last major technology this happened with was electric vehicles, and now it appears to be happening to AI.
The commoditization of AI is one reason that Anshu Sharma, chief executive of data and AI-privacy startup Skyflow, and a former vice president at business-software giant Salesforce, thinks that the future for AI startups—like OpenAI and Anthropic—could be dim.” That’s why EVs and self driving hasn’t advanced any more either. It is an application of the Pareto principle the 80% is easy but the remaing 20% is hard. Mims continues, “.An oft-cited figure in arguments that we’re in an AI bubble is a calculation by Silicon Valley venture-capital firm Sequoia that the industry spent $50 billion on chips from Nvidia to train AI in 2023, but brought in only $3 billion in revenue.
That difference is alarming, but what really matters to the long-term health of the industry is how much it costs to run AIs.”
Numbers are almost impossible to come by, and estimates vary widely, but the bottom line is that for a popular service that relies on generative AI, the costs of running it far exceed the already eye-watering cost of training it.” This is a unique blend of inefficient and costly. This is akin to hiring a narcoleptic to oversee security. It is a cash burning system that the entire world stock market is leaning on right now. Do you know how god damned stupid this is? No, let me help you out still more.
Here is an article that talks about Meta’s new AI gizmo called Cicero. This from ZeroHedge according to the futurism site, “Led by Massachusetts Institute of Technology postdoctoral researcher Peter Park, that paper found that Cicero not only excels at deception, but seems to have learned how to lie the more it gets used — a state of affairs “much closer to explicit manipulation” than, say, AI’s propensity for hallucination, in which models confidently assert the wrong answers accidentally. As Park explained in a press release, “We found that Meta’s AI had learned to be a master of deception.” This is where we are going to outsource our thinking, our busywork, our tasks for our jobs. What kind of world is this going to be. Its going to be an incorrect one with lots of errors, deception, and made up bullshit. This is why we can’t outsource or thinking, we can’t allow a device to tell you something and we believe it no questions asked, we can’t be that lazy. We have to think, actively think, critically think, and the more we think the more we realize that this doesn’t make any sense. Know this, bookworms like me will read the books written about this in 20 years. The authors will say after the fact that is was pretty obvious to everyone, but it simply isn’t right now. People believe so many things that aren’t real right now. If you pulled an earbud out of a Gen Zer there would be nothing but bullshit and bong resin oozing out of their ear for a week. I want to give them a hug like Robin William’s in Good Will Hunting and simply whisper its not your fault over and over again. It is true, it isn’t their fault that their parents were idiots and their parents’ parents were idiots. This is just the cycle of life and Tom Petty we are freefallin’ right now.
Sincerely Yours,
C Thomas Printer
On this date in history… 161 years ago to be exact, the Battle of Gettysburg ended. It also unofficially ended the South’s chances for winning the war and America’s chance at preserving itself as a Republic.
Also born on this date … the newly freed Julian Assange. Enjoy your freedom, Julian.
Thank you for listening today and you can find all of our articles and more on our website cthomasprinter.com.
You can’t spell failure without AI
Dell isn’t making money on AI?