Accel KKR thoughts

It's happening in the LMM and "micro" markets already, mostly for horizontal application software supporting SMBs (but I'm not exactly sure that VMS is safe either given enough time). It took me less than 8 hours part-time in under a week to spin up a basic scheduling app for a repair service shop and the extent of my coding knowledge is a handful of python and basic systems courses I took at college back in the 2010s. There were headlines last year about applications like Workday being displaced at large enterprise accounts (Klarna) with self-built tools using OpenAI and with the speed at which these apps self-improve that's inevitably going to become more common across all market sizes. Apps like Lovable (not even 2yrs old and has 1000s of SMB customers already) code at or above the level of entry-level engineers while obviously much faster + at orders of magnitude lower cost than even foreign engineers, so any business with IT spend IMO is going to gradually adopt them. Those savings drop right to the bottom line. The biggest hurdle I've observed so far has been getting more senior (and I'm referring to age here, generally 35+) engineers to actually adopt them because they're more used to managing people and aren't as quick on the uptake for learning prompting/trusting its work.

I follow a bunch of folks on X who talk about this stuff and constantly share new use cases/examples of apps folks are spinning up with little to no coding background. What's happened in just the last 6 months has been wild to watch, it's hard not to imagine that in another 2-3 years these applications becoming much more ubiquitous and being able to produce work that's far more complex. It will massively benefit actual software companies of course because it lets them lower their engineering costs which are by far their biggest line item, but it will simultaneously put a lot of pressure on their pricing leverage because there will always be the substitute threat available to customers to just do it themselves using these apps + a good engineer who can piece it all together if the customers don't like the tradeoff of high cost in exchange for a service partner and are willing to take on the execution risk themselves. 

I don't know a single person in software PE who isn't thinking about AI with respect to their own portfolio companies because if they don't adopt and implement it reasonably quickly they're going to get left behind because of the momentum that becomes available to their competition. That seriously threatens investments they in retrospect probably overpaid for if they can't execute on pricing and feature increases they underwrote on their base cases.