Discover more from Ben There Building That
Builders Weekly | 2/24/23
AI and venture capital, Duolingo's DAU growth model, and AI that interacts with everything on your computer
👋 Hey, I’m Ben! I write weekly about how to grow products and companies. I go deep on growth strategies, how to build products users love, and what actionable lessons can be learned from what best-in-class companies are doing and industry experts are saying.
You made it, it’s Friday!
I hope you’ve all had a great week. I’m back with the third installment of Builders Weekly where I’ll share a few of the things that got me thinking this past week and unpack how they can help you build better products.
What I’ve got for you this week…
Will AI be a venture capital dud?
Duolingo’s model for growing daily active users
AI that interacts with everything on your computer
Let’s get into it.
📉 Will AI be a VC dud?
Tyler Tringas had an excellent tweet thread this week that outlines a case for why AI will be a “dud for venture capital (VC)”. He touches on a lot of the same themes I covered in my post on building a winning business with AI.
TLDR; most of what is being built right now has no moat because it is just a UI on top of an API call to GPT-3. Big players with established brands and distribution power will come into and own the market once the cost to do so is lower. Adobe will own AI-based image editing, Office / Google docs will own AI-based writing, etc.
Assuming these big players do win the space there will be limited opportunities for billion-dollar exits for layer 2 & layer 3 AI companies. If this is the case, your fundraising strategy should look a little different. Tyler outlines his thoughts below…
Raise less than $1m, stay lean, build fast, and exit in a few years for a life-changing $10-20m/founder is a winning bet.
You might read this opinion and be skeptical, but I’d challenge you with a simple question: would you rather use “an AI tool for X” or have AI seamlessly integrated in the toolset that you are already using? I'd take the latter.
🦉 Duolingo’s model for growing daily active users
Duolingo’s blog released an excellent article on how they built a model to help them tackle stagnating growth in their daily active user (DAU) count.
To approach the problem, they broke DAU down into smaller user segments that were meaningful to the business but more specific and targetable. Once they had these cohorts defined, they were able to track the rates at which members switched between them and model which segments to target in order to have the largest impact on the business.
Whenever you’re trying to solve a hard problem, it’s helpful to break it down into smaller, more specific segments that you can build solutions around. Stay tuned for a future newsletter article on how you can do this via an exercise called opportunity mapping.
🤖 AI that interacts with everything on your computer
The number of cool things being built with AI continues to grow; I stumbled across adept.ai this week and wanted to share it with you all.
Adept is working on an AI model that interacts with everything on your computer. In its MVP form, it is a chrome extension that allows you to put in commands in natural language and it will then go and take actions on the web page to accomplish that task for you. For example, in the screenshot below, they show Adept executing on a manual task in Salesforce that would normally require 10 clicks by the user. I’m slightly skeptical on how ready for primetime this functionality will be (can it do any task on any website?) but either way, I’m excited to see where it goes.
You can read more about Adept and what they are building here.
That’s all I’ve got for you this week. See you again soon!
If you enjoyed this article consider giving it a like, share, and if you haven’t, subscribe for free to get more insights on how to build strong businesses around products that users love.