I always get nostalgic when the market crashes. Some of my best memories of the tech industry were after the crash in late 2001. I’d lost my job (in AI, actually!) and didn’t have much to my name besides a big Linux desktop computer. People decided that the web had been a passing fad. New York was a mess for reasons I don’t need to explain. 

Then suddenly I got an email: Did I want to come to the apartment of someone I didn’t know well and talk with other nerds about taxonomies? Did I! I brought a six pack of beer and hung out for a few hours. It was good to see people. We talked about XML, Topic Maps, information architecture, and search engines. The rest of the world might have found us irrelevant or out of the loop, but we just wanted to stay connected. Do I still work with those people? Yes, actually, I do! I throw them work, they throw me leads, and we catch up from time to time.

When we get together, me and my fellow tech industry folks, we talk about what excites us in our industry. We’ve seen a lot, but the reality is—AI is exciting. We’re all finding new ways to use it. That doesn’t mean there aren’t issues with it—copyright, energy usage, bias, overpromising—and tons to figure out. But at the same time, I keep finding uses. I’ve been using Claude from Anthropic to help me write code to learn music theory. I like it a little more than ChatGPT.

More relevant to this newsletter, in Aboard, we’ve recently added AI-powered search and filtering, and it lets you write really complicated queries in plain English and turns them straight into interfaces. It’s an incremental, easy way to build…well, software. We’re working on new stuff that lets you do larger, stranger programming tasks using AI. We’re able to prototype and experiment at a pace that would have ranged into months in the past, but now it’s almost conversational. Not the interface, but the pace. You can build stuff almost as fast as you can talk about it. I’m not surprised the market hasn’t figured out what a big deal this is, because no one has fully figured out what a big deal this is.

So I bet when the market crashed, I wasn’t the only person who felt a moment of excitement. Specifically, I thought, “Wow! Now there will be less venture funding, and a real incentive to lower the cost of training runs to build LLMs for people who want to experiment.” Which I know is sort of a strange prediction, but right now, the market is rewarding giant organizations that build enormous models at huge economic and ecological cost, and then we all use them by downloading (like Meta’s Llama 2) or chatting with them (like with OpenAI or Anthropic). 

What I remember the most about that period from 2001 through 2006 or so was that there wasn’t a lot of money to go around, but there was a thrill in building and seeing what you could get away with. I get why the technology industry is totally into making billions of dollars, but it was also extremely exciting to think about what things might look like if people started to think smaller—and to think harder, too.

Of course, the recent downturn didn’t actually turn out to be much of a crash. But it’s just something to keep in mind for the next correction. Because there will be one, of course. It’s not great when crashes happen, but it can lead to new kinds of thinking. And that can be exciting.