A good litmus test for someone being or becoming "AI native" is assessing whether they have an intuition for what different foundation models are good or bad at. The only way to accumulate an intuition like this is to make many things with these state of art tools. And to spend real time partnering with them to solve problems. No doubt, getting started is easier than ever, but the setup still deters a lot of people. It's tedious and you kind of need a sherpa.
There are a lot of dependencies and quality-of-life applications you need installed before you can start comfortably working with models like Opus, Codex, and Composer.
We started running a program at my company called the apprenticeship, which onboards new team members to these tools and helps them build a deep intuition for what the models are capable of.
Every time we onboard someone new, their machine is fresh and we hit all the same slowdowns. By the time we've finished explaining everything they need to install, their head is already spinning. So we built onboard.computer to make it easy on ourselves. It's not a new idea, but it gets people building way faster.
The application reads a software config file, compares it to what's on your machine, and shows you what's missing or out of date.
For devs, we created a companion Claude skill called "/onboard-computer" that you can use to produce YAML config files from directly within your project root or repo (we call these .onboard files). Run the skill and it will find all the dependencies, prompt you for any QOL apps you recommend to new users, and produce a file that you can give to anyone.
Drop a .onboard it a repo or give it to your team, and most anyone can go from zero to building in minutes instead of hours. It's not complicated. It just works well and it runs entirely on your machine. The project is open source and free.
There are so many people who can and should be building with these tools. Fighting to setup their dev environment shouldn't be a barrier.
We hope you share it with someone who wants to make something!
— Andrew & Benj