Anything that can be copied will be copied in AI. I get a lot of pitches from teams who are building AI infra & apps with plans to own a niche in the market and expand beyond that wedge. They’ve been working on technical development for their companies for a few months & sometimes even weeks - then trying to raise a seed round. You can tell that they are technically talented & are moving fast. Maybe they even have a friends at other companies who have agreed to be "design partners" which they market in the deck. But these teams haven’t spent enough time thinking about product strategy & have a hard time explaining how Anthropic, OpenAI, and other big dogs won’t copy them when things go well. A basic rule if you’re building in AI right now is that "Anything That Can Be Copied" will be copied & that you have zero leverage beyond potentially being obsessed with some random niche that big players aren't prioritizing. You need to spend less time on building an MVP in 2026 & way more time on product strategy.