Welcome to my newsletter! I'm Dave Anderson, an ex-Amazon Tech Director and GM. I write this newsletter I've called Scarlet Ink, which is a weekly newsletter on tech industry careers and tactical leadership advice.

Free members can read some amount of each article, while paid members can read the full article. For some, part of the article is plenty! But if you'd like to read more, I'd love you to consider becoming a paid member.

FYI - If you missed the announcement last week, this newsletter has moved from Substack to Ghost. Hopefully this has no impact on you readers. If you do run into an issue, please reply to one of my newsletter emails, and I'll take a look.

I've been in the tech field for a relatively long time. I grew up playing games and coding on my Apple IIe. I learned C before C++ because C++ was pretty new. Anyway, my point is that I've been in the tech field for a while.

I find this AI boom/bubble absolutely fascinating, because so much of it reminds me of the 90s dot-com boom/bubble. And aging sneaks up on you, because I didn't think about the fact that many/most people in the tech field didn't live through the dot-com era.

What was my impression of the dot-com era? Everyone was incredibly excited about the global acceleration of commerce and business due to the explosion of internet access. Lots of capital was accessed, generated, and reinvested into the market. This incredible excitement skyrocketed stock valuations. Great ideas were funded, but also an incredible number of stupid ideas. The access to incredible amounts of money (making most hot startups oversubscribed) didn't just attract smart people with great ideas. It also attracted people more interested in getting funding than building something. It also encouraged otherwise successful companies to wildly overgrow, and burn themselves to the ground with overhiring and overspending. Companies that might have otherwise survived built such a high burn rate that they couldn't survive unless they grew to infinity.

Does this sound familiar? I certainly think it does.

I think we'll walk through a comparison of the two booms, including what we learned about building for the web and how that suggests we're getting things wrong today too.

The anatomy of a boom.

Dot-com company excitement was partially related to breaking down previous moats. As a small example, before internet access expansion, competing with existing big book stores would be hard. Most physical locations already had a store, people knew the brands, etc. However, Amazon was differentiated and could compete in an entirely new space as they approached customers from the internet rather than a crowded mall. I think this differentiation was a core element of the dot-com boom. "We can do everything on the internet now!"

This was largely true over the long term. But in the short term, customers weren't ready to do everything on the internet. And what was built was frequently crappy, as most V1 products are. So it required time for customers to change their habits, and for builders to make real improvements.

I think elements of this AI boom resonate in the same way. If you wanted to build yet another customer service outsourcing company 10 years ago, that would be tough. There were entrenched customer service companies with tens of thousands of agents. Plenty of companies had their own in-house offshore customer service groups. It was an extremely mature and established business model.

However, as soon as LLMs appeared, it was obvious that they were particularly suited to tasks like customer service. "We can do this with AI!" people proclaimed excitedly. And like Amazon stepped around the crowded physical space, AI can avoid all the downsides of using humans for various tasks.

And similar to the dot-com explosion, customers are also not necessarily ready for what we're building. We don't trust AI agents yet. When a chatbot answers my support request, I frequently groan. "Get me to a human!" I type with annoyance.

This is partially because we're also not good at building things yet. The agent can read help files to me, but how many agents are given the power to do refunds, or give me a discount code, or fix the address on my latest purchase? We'll get there, but it'll take time to get good at building.

As a side note, when I look at some real successes from the dot-com era, I see a pattern. In the gold rush era, it was profitable business to sell shovels. In the dot-com era, companies like Cisco made bank selling the equipment necessary for those thousands of internet startups. And these days Nvidia is obviously shining.

I'd also point out that while Cisco made a ton of money, their stock price also fell 88% after the crash and took forever to recover. Because forever growth rarely happens. And I'm suspicious about Nvidia's market capitalization. Not that the hardware business is bad (again, selling shovels absolutely makes money), but there's an incredible amount of assumptions tied into that stock value.