There was a time - not long ago - when AI coding assistants were glorified autocomplete toys. They helped you write console.log("hello world"), occasionally guessed what you meant, and only cost $19/month. For a while, that felt right. After all, $19 was less than your Uber Eats service fees.

But then the pricing got weird. What used to cost less than a large pizza started creeping up, then shot straight into enterprise-grade pricing models where costs scale aggressively and justification feels optional. Some tools now charge hundreds of dollars per developer per month for something that, just a year ago, felt like a neat autocomplete party trick. That’s not just inflation, that’s a whole new economic reality. The question is: what exactly are we paying for, and when did $19 become the opening bid?

In this talk, we’ll unpack what’s really happening under the hood of modern AI assistants:
- How the economics of token usage and scale are breaking old pricing models;
- Why the tools you're using today might be dramatically more powerful than the ones you started with (even if you're not using them that way yet);
- And what it means when your AI bill suddenly starts to rival your cloud spend.

We’ll also peek into the future of autonomous agents. These things won’t cost $19 - or even $190 - but if they can spec, code, test, and ship with minimal oversight? Maybe that’s a bargain. Or maybe it’s just the latest chapter in Silicon Valley’s favorite game: reinventing the wheel with more layers, less control, and a higher subscription fee.