Natural diamond prices are in free fall, down 43% over the last four years. Why? They now have a chemically and visually identical lab-grown substitute that is also collapsing in price.
The eroding premium for natural diamonds results in a perverse market for lemons where the bigger and shinier the rock, the more likely an observer will assume it’s a fake stone and worthless commodity. When production is automated and it’s impossible for a consumer to discern the difference between natural and artificial, all production is worth less.
This is the same commoditization we’re seeing as AI takes the reigns of creation. It’s harder to quantify the value of well-written technical blog post or polished marketing site or a well structured PR, but I’m sure the value is declining as everyone stops engaging in the material.
Emerald prices, on the other hand, have been much more resilient. Unlike a diamond’s highly standardized 4Cs, emeralds are prized for their inclusions and color zoning (imperfections) that are difficult to recreate in the lab.
Every founder should aim to build something that is difficult to recreate in the lab - sell emeralds, not diamonds. Opinionated software that highlights the domain expertise and preferences of the team building it. Have the courage to make your world view legible in the software you create.
Ivan Zhao wanted to build paper for the internet and Notion works the way he thinks the world should work. Ian Kalin wanted to keep soldiers safe no matter what hardware they were deployed with, and TurbineOne’s product is defined by his military experience. Kareem Amin wanted a spreadsheet that fills itself in from the context of a a GTM team and you can feel this focus in the product at Clay, even at current scale. Chris Ellis and Adam Stevenson want to build a healthcare system people love and it shows in every product choice with Thatch that reduces complexity and tailors care patient by patient.
In this world where anyone can instruct the model to build anything, you also need a stronger signal that what you’ve built is not something that can be vibe coded by your end user. Time to value has to be instant and early integrations must make it clear how your approach, ergonomics and method get stronger as the models get 100x better.
Commoditization of the means of production means you have to fight harder to generate unique outputs. Product leaders need to ship applications that are obviously different in how they create and capture value. The best software will absorb our idiosyncrasies and manifest more value as a result. Personalization that gets deeper and irreplaceable with use. Applications that respond to realtime data streams to deliver increasing value that’s specific to a customer over time. Software that defines new workflows, imbued with intelligence that creates a widening advantage over the generic and lab grown alternatives.
The more founders who have the courage to express their unique ideas in software, the more likely we see intelligence flow where it is most valuable and the benefits of AI have the most impact.
Will advise folks to think about emeralds from now on