AI: Here come the AI Doubts. RTZ #376
...like clockwork, the media focuses on glass half empty
It happens every tech wave, like clockwork. The doubts emerge on the initial enthusiasm, just as the race starts and gets going. Is AI the wrong peg in the wrong hole? And is the glass half empty?
And it now happens to the AI Tech Wave. The patterns are the same, every time. Breath-taking enthusiasm for the new, new thing, with non-stop cover stories on the ‘founders of the moment’, their seeming ability to walk on water, and the epochal, ‘disruptive’ capabilities of the technologies being invented by their companies.
Followed inevitably, with questions of the technology’s efficacy and mainstream adoption. Especially at the first, predictable signs of a pace slower than loudly pronounced just a few headlines ago.
The frequency some times is fast or slow, but as autumn follows summer, so do media doubts follow media enthusiasm after the dawn of the ‘next, new thing’.
After barely a year and a half into the OpenAI’s ChatGPT catalyzed LLM AI boom, and the robust, global media enthusiasm for the same, the season may be shifting. Something I’ve observed already before, but it’s particularly pronounced as we enter the middle of 2024.
It’s already been reflected in the stock market this past week, when enterprise cloud companies took a pause of 10-20% after being up a 100% or more year to date. And the media is noticing, as one can see here. The WSJ’s Christopher Mims had dead-on point piece on Friday titled “The AI Revolution is Already Losing Steam?”. The punch line is as follows:
“The pace of innovation in AI is slowing, its usefulness is limited, and the cost of running it remains exorbitant”.
Going on to of course Exhibit no. 1, Nvidia, the current driver of the AI Train:
“Nvidia reported eye-popping revenue last week. Elon Musk just said human-level artificial intelligence is coming next year. Big tech can’t seem to buy enough AI-powering chips. It sure seems like the AI hype train is just leaving the station, and we should all hop aboard.”
“But significant disappointment may be on the horizon, both in terms of what AI can do, and the returns it will generate for investors.”
The long piece then goes into the inevitable and predictable points, outlined here below:
“The pace of improvement in AIs is slowing”.
“AI could become a commodity”.
“Today’s AI’s remain ruinously expensive to run.”
“Narrow use cases, slow adoption”.
And then the ending paragraph that adds some protection in case of a reversal of the perceived ‘reversal’ in AI enthusiasm above:
“None of this is to say that today’s AI won’t, in the long run, transform all sorts of jobs and industries. The problem is that the current level of investment—in startups and by big companies—seems to be predicated on the idea that AI is going to get so much better, so fast, and be adopted so quickly that its impact on our lives and the economy is hard to comprehend.”
“Mounting evidence suggests that won’t be the case.”
The whole piece should be read to digest points well raised. Just a tad early, in my view. The right doubts perhaps at the wrong time.
Regular readers here have seen my pieces addressing each of the above, with specific discussions here, here, here and here.
I won’t repeat or belabor the counter points in detail here. But just repeat the following high-level responses to the above key points raised:
The pace of AI improvements is NOT slowing, but ACCELERATING EXPONENTIALLY, due to ‘AI Scaling Laws’ that for now beat Moore’s Law. Just wait a couple of years or less.
Tech only becomes a commodity after market saturation. We’ve barely seen any mainstream adoption by businesses and billions of end users. We’re at least half a decade or more away from that point.
Yes, VERY expensive to build and deploy ‘Big AI’ today, but will quickly be offset by ‘Small AI’ on local devices in the next 2-3 years. Again, just wait.
Narrow use cases right now, but the peanut butter is barely out of the jar. Give it a chance to be truly spread around. We’ve not seeing anything yet in terms of mainstream adoption at scale.
I’ve long likened this AI Tech Wave to the PC tech wave of the eighties and the Internet tech wave of the nineties. Just looking at the Internet wave, when Netscape showed the world what the Internet could be with its browser to access growing websites in 1995, there were less than 25,000 websites at the end of that year, a 100,000 by January 1996, over 250,000 by the end of 1996, over a million by end of 1997, almost 2.5 million by end of 1998, and over 3 million by end of 1999.
I stop there for now because a little company called Google was founded in 1998. THREE whole years after the Internet gun went off with Netscape. And they came along with an invention called ‘Pagerank’ that allowed links in web pages on a web site to be measured and ranked against links in other web pages on websites. And of course evolved into one of the best businesses ever invented over the next 25 years.
Today we have over a billion websites in the way we think of websites, and over four billion users accessing their links over smartphones. Over ten billion searches a day. Even as we debate how AI will change Search this time, for Google, OpenAI, and others.
The AI Tech Wave, even with the historic build-out underway of AI Tech infrastructure in the hundreds of billions today, is barely at the beginning of the beginning of its ‘accelerated computing’ buildout and mainstream use.
So the questions in articles like above, are the right questions, at the wrong time. Save them for a few more years. At least a couple of years to start. Then let’s revisit them.
Let’s be patient, and pause on the doubts. Wait for AI products to actually be built, deployed, re-built, and then potentially adopted by mainstream users in the billions. It will all happen in time.
Let’s see if it’s the right pegs in the right holes. And if the glass is half full. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)