As I think about the hype around AI and Large Language Models, I’m painfully reminded of a quote from Douglas Adams.
― Douglas Adams, The Salmon of Doubt: Hitchhiking the Galaxy One Last Time
“I've come up with a set of rules that describe our reactions to technologies:
1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works.
2. Anything that's invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it.
3. Anything invented after you're thirty-five is against the natural order of things.”
I’ll confess, two technologies have moved to prominence after I turned 35: crypto and LLMs. I feel vindicated on the first, but feel like the second one is more complicated. There is invariably a lot of hype around AI right now, and I am sure we will see a correction at some point. However, there seems to be something promising, even if it can’t yet do your job.
If found myself searching for an analogy, some earlier comparison that may help navigate the path we’re on. After a bit of thought, I’ve settled on pet supplies.
In 1998, Pets.com launched as an online retailer specializing in the online sale of pet supplies. The company’s value peaked when it made a public offering in February 2000, with a valuation of approximately $400 million. The company failed spectacularly after spending millions on advertising and losing money on most orders. By January of 2001, the company was finalizing its liquidation.
Ten years after that, Chewy was founded with a focus on the exact same market. However, its fate has been very different. By 2014, it had over $200 million in sales, dwarfing the $6 million that Pets.com had reached at its peak, and it has continued to grow. As of this writing, Chewy is profitable with revenue of nearly $12 billion and net income of $400 million in 2024.
I’m not the first to make this comparison; others have made and dismissed the argument. Ultimately, a lot had changed. The number of online users grew massively, smartphones became widespread, online purchasing became more convenient, and shipping has also gotten much easier. I would argue that these underlying changes may have played a significantly larger role in differentiating the success and failure of these two companies.
Going back to AI, OpenAI is losing money on its premium subscriptions, and it lost around $5 billion in 2024. I’m not saying they won’t pull through, but at the moment it seems that whatever underlying technologies would make an AI company profitable may not yet be as developed as needed to build viable business.
I’m not sure I have a great answer on those technologies are – obviously, compute and power come to mind – but I suspect there may be more than that. I’ll have to leave a discussion of implications to another day, but just leave you with the idea that I’m not saying OpenAI will fail or that industry or that LLMs won’t have a transformational impact. Rather, I’m saying we may still be on the wrong side of the trough of disillusionment and should be thinking about what sort of underlying capabilities are going to allow this technology to take off and what we do in the meantime.