A Few Thoughts on AI
Software's eating the world even faster
Here are a few views on the emerging AI landscape:
AI will expand the software industry just like the cloud did. While some legacy on-prem software companies were slower to adapt to the cloud than others, it was generally a boon across the industry. Even relative laggards (Adobe, Autodesk, Microsoft) were able to thrive once they embraced the cloud.
We expect today’s software industry to move faster to embrace AI than yesterday’s software industry embraced the cloud. This means it will be marginally harder for a new AI startup today vs a new SaaS startup 15-20 years ago.
The recursive momentum of AI’s progress seems greater than the cloud, which means the market expansion could also be much greater. AI will accelerate software eating the world.
The best SaaS companies don’t sell “cloud,” they sell benefits derived from leveraging the cloud. We expect the same with AI. Savvy SaaS startups will use AI to turbocharge their customer benefits.
Developer enthusiasm for AI lends itself to creating open source companies. We think AI will boost open source’s pace of taking share from closed source. There’s plenty of room for closed source companies to thrive, but we think the median open source AI startup will have an edge. Drop in on LangChain’s Discord to take a glimpse of what thriving open source can look like in AI.
AI has lowered the bar for creating certain types of compelling products. But it hasn’t lowered the bar to building great companies - only the very best founders will be able survive the rigors of scaling to create iconic companies.
Agree? Disagree?
AI is moving insanely fast. I’m sure some my views will look foolish in a few years’ time (or sooner). If I only knew which…
The comparisons to cloud are interesting — just like in that era, AI/LLMs introduce a non-linear disruptive change-agent available right now to any company regardless of size. And probably even moreso than the cloud migration due to its instant availability in the application layer to anyone building tools for end users.
A major difference is that most of the benefit of cloud didn't accrue to end users, but to someone in IT or management within their orgs. Predictive "AI assistant"-type tech benefits buyers (by increasing productivity, decreasing costs) AND end users.