I recently wrote an article here titled 1 Year Since Co-Founding a SaaS Startup. This post had a lesson-learned titled "Lesson #4 - Do not get distracted by tulip mania." in which I described falling for the latest "tulip mania" in the form of generative AI, and how this distraction cost my startup valuable time and resources. This was not to insinuate that AI is not going to be a long term game-changer, but that the hype in the last year or two has not been proportional to where the tech actually is.
This is how bubbles form. A new fad hits the world and the Fear of Missing Out (FOMO) sets in quickly. This FOMO accelerates asset prices as people, often not informed on the "hit new thing," pile in to try to reserve their seat on the rocket ship. Before you know it, the bubble pops, and reality sets in.
One of the most famous cases of this happened in the 17th century: Tulip Mania.
In the 17th century Netherlands, during the Dutch Golden Age, tulip bulbs became incredibly popular and valuable. This "tulip mania" saw prices for some rare bulbs skyrocket to outrageous levels, exceeding the cost of houses and exceeding a year's salary for the average person. The most expensive tulip receipts that History.com found were for 5,000 guilders, the going rate for a nice house in 1637. By February, the tulip market collapsed, and many were left holding the bag. A cautionary tale, indeed.
History doesn't repeat, but it does rhyme. Let's take a look at some other, more recent, examples of tulip mania that we can learn from.
In the late 90s, the internet was all the rage. The world had very quickly transitioned from making fun of email as a concept on popular talk shows to throwing money at anything with a website (hence the name). The frenzy was wild. With FOMO in the air, investors were piling into startups that often didn't have a product or even a fully fleshed out idea.
As concerns rose about the sustainability of these inflated valuations, the bubble burst. Investors panicked and began selling their tech stocks en masse, leading to a sharp decline in prices. Many dot-com companies, unable to sustain themselves without the constant influx of investment capital to burn, went bankrupt. The crash had a significant impact on the tech sector, leading to widespread layoffs and a loss of investor confidence. While some established tech companies weathered the storm, the dot-com crash served as a stark reminder of the risks associated with speculative bubbles and the importance of sound financial practices.
Between 2019 and 2022, starting largely with the Bitcoin halving in 2020, the world went crazy for anything Web3. Major brands like the NBA were releasing Non-Fungible Tokens (NFTs), which are essentially a way of verifying ownership of a digital asset on a blockchain. Many NFTs were selling in the order of tens of thousands of dollars and more. Many of these are now worth pennies on the dollar compared to their prices just a few years ago.
The Web3 craze was not limited to NFTs, however. Remember how many celebrities were doing commercials for now-defunct exchange FTX?
New crypto currencies with no history were launching by the day. These new coins often had no real benefit above established players like Bitcoin and Ethereum. Many of these coins were used for pump-and-dump schemes, where a small group would manipulate the price and stoke FOMO. People, having seen the meteoric rise in the Bitcoin price, were hoping that they could buy these new coins at a cheaper price and eventually strike it rich. Most of these coins went to zero, and the people who invested in/bet on these assets lost their entire holding.
Transformer architecture was invented in the 2017-2018 time frame, primarily pioneered by Google. This technology has been a game changer for Artificial Intelligence and Machine Learning technology. In the time since then, companies like OpenAI, Google, Anthropic and Meta have been on the cutting edge of developing generative AI.
The way generative AI works is fairly simple, but also very resource intensive. You train a model on a large amount of domain knowledge. From there, it is able to use weighted pattern matching to effectively predict answers to asks. This can be in the form of text, as we're mostly familiar with, but also in the form of images, sounds / music, vertices in 3d models, and basically anything else that has a large domain of knowledge on which models can be trained.
When AI became main stream last year, the public was entranced. I know this because I was one of the entranced people. This ability to generate responses in plain English surely signaled the rise of the machines, right? That's what the market thought. In a time of tight monetary policy and tough fundraising conditions, anything with a .ai domain name was getting funding -- sound familiar?
It remains to be seen how this potential bubble will play out, but what is clear is that the technology currently amounts to a party trick for most use cases. I'm bullish long term and do think this technology will massively boost productivity, help cure many diseases, and may even help us travel the galaxy. Right now, however, it's just not reliable enough for most cases.
For each of these bubbles, the bubble does not signal that the technology has no use. In the case of dot-com, the internet has obviously been a game changer. The hype and loose investment policies were the problem, not the technology itself. Similarly, I think blockchain and AI are here to stay and similarly will be game changers, but the short term hype does more reputational harm than good to the industries and may cause many to lose out in a big way.
Other honorable mentions to look into, if you are interested in bubbles / crazes, would be the 2000s United States housing bubble, the roaring twenties and the great depression, and even Google Glass which seemed to be all of the sudden everywhere, and then all of the sudden gone.
--
Sheridan from Essembi -- https://essembi.com
If you'd like to follow me or essembi:
@sheridanbulger on X
Sheridan Bulger on LinkedIn