It’s Not Any Different This Time
Generative-AI Promises — Fantasy & Reality
Yes the AI Revolution will change the world and yes AI stocks will crash. Two or more things can be true at the same time, and just because certain stocks are going into the stratosphere doesn’t mean that the industry will deliver. Generation after generation, economic cycle after economic cycle, willfully ignorant investors who never pay attention to history become the useful suckers that contribute to the next tech bubble.
So yes, it’s not going to be any different this time.
Investors are in full swing once again. The S&P500 index has gone into rarefied air and there is a new so-called Magnificent Seven group of big tech stocks that are being pumped up under the narrative of future gigantic profiting from the artificial intelligence revolution. Amazon, Apple, Google, Meta, Tesla, Microsoft, and the new and mighty Nividia. Note, that these stocks are well established and have solid and broad business models and solid growth/revenues/profitability. Nevertheless, the hype surrounding AI is pushing GenAI-specific stocks higher with irrational exuberance. And the M7 ones too are naturally caught up in that because people assume they are a “safer” way to play AI. However, if AI doesn’t eventually deliver to the level of the hype, there will be a reckoning.
Unfortunately, I’m not nearly smart enough nor do I have a crystal ball to tell you what to buy, when to buy or when to sell. All I can offer is some applied intelligence (ai) commonsensical approach to thinking.
A quick comparison to the dot.com era —1990s to early 2000s bubble and the eventual bust. Reminds me of another magnificent group of telecommunication fibre-optic stocks: Worldcom, Nortel, Lucent, Alcatel, and even an up-and-comer named JDS Uniphase. Another tech revolution within a broader tech revolution (dot.com) — with an unstoppable upward trajectory.
However, within only a two-year period the dimensions of the collapse in the telecommunications industry were spectacular. Over half a million people had lost their jobs. The Dow Jones communication technology index dropped 86 percent; the wireless communications index, 89 percent. These are declines in value worthy of comparison to the Great Crash of 1929. A $7 trillion decline in the stock market since it peaked, and about $2 trillion disappeared in the capitalization of telecom companies. Twenty-three telecom companies just went bankrupt in a wave capped off by the collapse of WorldCom, the single largest bankruptcy in American history back then.
I have no idea what is going to happen next with AI but I do know that all markets and human nature are the same. Eventually, the hype game will end and those fragile AI companies will have nothing left in the tank to push it higher. Gravity will show up like the Grim Reaper. Panic and sell-offs will happen, and those excessive and bogus fantasy valuation levels will too collapse. The destruction and collateral damage will be dramatic and much greater than we’ve ever seen before.
However, let’s not spend precious time doing the impossible — making predictions. Let’s take a deeper dive into the “AI Revolution” itself and look for the objective truths about it. Specifically generative AI (GenAI).
Let me preface things by defining the stock market; it’s a casino, so if you can understand that you might do better at it — the game. Casinos sell dreams of riches and that’s how it lures you in…it needs its fair share of useful suckers — the house always wins.
When it comes to GenAI, what we’ve seen so far are promises promises promises, pushing up stock prices with nice stories. For example, investors have sunk over $100 billion into the driverless car industry but we are nowhere close to the driverless world the likes of Elon Musk and others have promised. In reality, with all the mishaps, it’s going in the opposite direction.
When it comes to LLMs (large language models). With all the hype and ever-increasing stock prices, ChatGPT and Gemini continue to hallucinate in epic proportion — overanthromorphic…making stuff up. So just because stock prices are going up doesn’t mean LLMs are delivering. Because it clearly isn’t.
As Gary Marcus puts it “Silicon Valley wants you to believe that the [hallucination] problem is about to end. Ten days ago, in the era before Sora, Gemini, the OpenAI meltdown, and the Gemini image scandal — Andrew Ng optimistically told the Wall Street Journal that hallucination was “not as bad an issue as people fear it to be right now”. In September, Reid Hoffman claimed to Time Magazine that the problem would be largely solved within months.” This of course has not happened and there is no evidence to believe it will happen anytime soon, or if at all.
Constant promises and heuristic pontification by the Big Tech aristocracy and their VC courtiers is the common narrative being used to keep the investor suckers on the line. These two massive models (ChatGPT and Gemini) both just make stuff up and even though they both have multi-billion dollar budgets, they continue to fail — i.e., just last week Gemini had to shut down because its hallucination went off the rails.
Stephanie Palazollo, with The Information, put it recently, interviewing DataBrick’s CEO Ali Ghodsi, who said “By themselves, large language models that power today’s chatbots leave a lot to be desired. To turn them into a truly big business, the models need ways to access real-time data in a reliable way, otherwise known as retrieval augmented generation”.
To turn them into “truly big business,” is the main point here, and the main challenge for GenAI companies. But again, there is no credible evidence they are anywhere near this.
So don’t buy into the hype narratives surrounding those stocks. However, I’m not dissuading anyone from making money in the casino. I’m just saying be aware.
To get a good understanding of how problematic things are for GenAI. I recommend reading the following article (No, RAG is probably not going to rescue the current situation) on Substack, by the formidable computer scientist and AI entrepreneur Gary Marcus. If you want to learn you must seek the objective truth.
GenAI/LLMs are just in their very early days and just because stocks like Nividia are blowing away their earnings with real sales/revenue/profits, GenAI is still not delivering. At some point, there will be a come-to-Jesus moment — a moment of sudden realization, comprehension, or recognition that often precipitates a major change. A period of extreme volatility, where the commonsensical and long-term applied intelligence (ai) investor mindset will win big.
“Two things are infinite: the universe and human stupidity: and I’m not sure about the universe.” — Albert Einstein
The same stupid mistakes made in the past will be made this time around too. So if you’re a more sophisticated and intelligent investor, you might develop a contrarian investment strategy that aligns with reality, and take advantage of the coming volatility and fragility of certain tech stocks.
Without a doubt, LLMs can do some great things and have created real value and even appeared amazing at times. But it’s nowhere close to what has been promised. Nor has what is being touted right now lived up to the hype. There is no revolutionary magic bullet here. Reality and fundamentals will always matter most. No matter the time.
To make LLMs work as the hype says they could, and be reliable or non-hallucinating enough would mean that LLMs would have to achieve a level of human General Intelligence. That’s not in the cards. So we must not allow ourselves to become detached from reality.
For the last 50 years or so we’ve gotten many promises about AI but the ball keeps being kicked down the road. This doesn’t mean we don’t continue to research, develop and innovate, and it also doesn’t mean AI has not changed our world. It has, relatively speaking. But the question is to what degree and whether or not it has delivered on the hype. It has not!
As investors, we can’t ever accept being played with nice stories, and not holding storytellers to account.
“But we are kidding ourselves if we think the solution is near at hand. Making it work is very likely going to require a whole new technology, one that is unlikely to arrive until the tech industry gets over its infatuation with LLMs and quick fixes.” — Gary Marcus.
Again, it’s impossible to predict the future and we simply can’t ever tell where things will be in even 3 to 10 years. However, we can apply some commonsensical applied intelligence to help us make good investment decisions. Remember, the only thing that changes is time and technology, human nature doesn’t. The only true thing is the objective truth. So it won’t be any different this time around.
