Korva is a scroll of the internet’s best ideas.
AI startups win at the edge
Harvey, in legal, is a leading example: it has done well, but it picks out pieces of the work and has not gone after the core system. Companies on the edge of a system are more exposed, because people will experiment with AI decisions there.
Nobody is willing to do the work yet to explore rebuilding a core system, since no one knows what happens downstream when a piece is pulled out. An AI-native player will probably succeed there eventually, assuming its team appreciates capital efficiency, which many startups don't. That was the downfall of a lot of SaaS companies in 2021.
At a European group that rolls up software companies, many of them smaller, no customer has left or even threatened to leave over an AI-native solution.
Why is the core of a system hard to replace?
The core is the system of record, the authoritative source that other tools read from. Geoffrey Moore drew the line between systems of record and systems of engagement in a 2011 industry white paper. Edge tools can be swapped without touching that source, but replacing the core means moving its data and rewriting every integration built on top of it. A failure there spreads to every process downstream.
Why did startup funding tighten after 2021?
Venture funding peaked in 2021, when interest rates were near zero and investors rewarded growth above almost everything else. The Federal Reserve began raising rates in March 2022, which made safe investments more attractive and cut the value of long-dated growth forecasts. Venture funding fell sharply in 2022. Startups that had planned around cheap capital had to cut costs or raise money on harsher terms.
More from Joys of Compounding
