There’s a long overdue “vibe shift” happening in technology and the broader business landscape.

A recent example is this episode of LaBossiere podcast, which I highly recommend listening to. Delian Asparouhov (Founders Fund & Varda Space) breaks down why zero marginal costs of software are a myth.

In short:

As entertaining as dunking on the 932nd generic B2B SaaS tool might be, nobody in their right mind would disagree that SaaS is a beautiful business model in isolation. Worked amazingly between 2005 and 2015. I love Slack, Notion, Shopify, Stripe et al. as products - and their business side is even more impressive. I don’t even doubt that there will be similar stories in software every year for the next ten years.

The upcoming decade will have way more capex than the previous one though.

To all of you accountants out there, I won’t differentiate between capex and asset-derived opex here. You can almost always outsource the long-term capex to another entity and have only short-term costs. Let’s use capex as a term encompassing all derivative costs from capital investment.

Might be premature to say, but so far the application layer of the AI hype cycle is a dud. Most value is accruing into extremely capital intensive foundational models. Foundational players like OpenAI have been able to outsource experimentation via their API to startups, traditional SaaS products integrating AI features, and indie devs. They can just observe what gets traction and verticalize it themselves.

Regardless of this, I encourage anyone willing to tinker around with AI, maybe it’s even an awesome side hustle! Just don’t expect defensibility at huge scale.


A site by Joose Toiviainen