Exactly right. All the exuberance in 98/99, all the investment, the tech crash, the washouts, the companies picking over the bones. But the chaos faded, restructuring happened, creative destruction worked, and we got past it to the Internet we have today. I was in the thick of it back then massively overbuilding optical networks. Then, literally in one month, orders for optical transmission equipment stopped, followed by 100,000 people were fired, and the company went out of business.
My advisor believes this is due for a washout. I do too. Just about all involved are building and planning as if they will contronl near 100% of the market space and we know that can’t happen. Somebody will be right. In addition to Microsoft and Uber I see that Frodo’s Farcebook has retreated some finding the cost of entry excessive and choosing to be like a fast follower.
One week, not long ago, I spent time doing some research on the subject of AI, not exhaustive but what I believe is adequate. I searched in vain for any business plan by the providers. Instead of modeling the business plan I used the best projection of cost and revenue I could find and built a simple economic model. The most glowing forecasts resulted in economic failure. One would be better off putting his money in just about any simple index fund.
I also found one podcast interviewing a Sillycone Valley capitalist. His conclusion is similar to my own.
Some of the other factors I see that are not in favor of the AI hype: Equipment being installed now will be obsolete in 5 years or less. Newer more power efficient equipment is needed. Not all the facilities are needed, if it continues more will be done with less. The fees necessary for profit are probably unaffordable at the consumer level.
Not all my AI experiences are bad. It is a good tool to cut to the chase on certain questions but is not always to be trusted. It gets things downright wrong and I know this because of my considerable background in my field of engineering, mechanics and agriculture.
AI does not think, it harvests and assimilates and compiles and reports what seems to be consensus. AI is a very poor customer service agent and problem solver. It only “knows” what is programmed in response to a very specific set of canned questionss. I spent more than three hours just yesterday trying to find the answer to odd messages from PayPal. AI just kept circling back to the same useless responses. I finally got a human on the line who tells me that PayPal’s internal system is generating some alarming messages to users saying a new account has been created in my name and to call a bevy of numbers to complete set-up. Alarming. I have frozen my account pending just canceling it. My AI experiences in lieu of human interaction, even difficult human interaction with someone in India, have been very unsatisfactory. I expect this will persist, nobody will admit a mistake and instead will continue to insist it is growing pains that will get better. I doubt it.
The money in AI is in the parts and service suppliers. I have seen the same thing in oilfield booms when the only parties that made real money are the parts and service suppliers and the producers who survived the last collapse being present to reap a windfall.
Strangely enough, back in the .com boom I bought some of the eventual winners. It was hard or impossible to tell who they were going to be when the shake-out started. I owned things like Broadcom, Micron and AMD and don’t now and have not for ages. Their time has come again for now and the gain is more than impressive, I’d call it improbable. Huge risk takers are sometimes rewarded I’m just not one of them any longer. These winners are hardly easily identified on anything like fundamentals and instead it more luck in the picking or following the heard using the bigger fool tactic.
Just my thoughts of course.