Table of Contents
Scott recently posted his take on building a company in the age of AI. In it, he discussed generating website content and code, doing market research and various other tasks that AI has made easier through the building and launch of Essembi. If you haven’t read it, check it out here.
I tend to move from task to task, initiative to initiative, fire-drill to fire-drill without much reflection. It’s fair to say that my rear-view mirror is foggy, if non-existent at times. That said, reading Scott’s thoughts offered the perfect opportunity to reflect on how AI has helped me day-by-day in my half of the yin-yang of leading Essembi.
Scott asked me what impact I felt that AI had on our business. My answer was that I felt AI made me 33% faster with 33% higher quality of output. This is an average across all actions, some where I’m 100% more efficient and others where I’m 0% more efficient because AI doesn’t have an application there (yet).
There are tons of examples we could discuss like competitor analysis and feature research, but the biggest needle mover for me has been: code, code and more code.
Check out GitHub Co-Pilot Chat reformatting code based on a prompt:
This tech has a long way to go. It makes a lot of mistakes. Corollary to that, the fact that this is the floor for this functionality is incredible. I picture a world where the AI is catching bugs in real time, authoring unit tests and yes – outright building features on the product roadmap.
There’s a natural progression that happens as tools improve over time, where the user of the tools is allowed the opportunity to think more about the big-picture flow of things rather than the nitty-gritty minutia. I remember when I first saw IntelliSense giving me real-time documentation and function call signatures while writing code. It was mind-blowing at the time, as AI is mind-blowing today.
Tools like GitHub’s Co-Pilot exponentially continue this trend of development tools that shorten the commute between problem and solution. This is especially true when interacting with third party libraries and APIs that have large adoption. Suddenly, rather than muddling through API calls to find the right combinations of functions to use, I’m spelling out pseudo code logic in comments and the AI is filling in the blanks.
Don’t take my word for it, though. Here’s what Thomas Dohmke, the CEO of GitHub, has to say:
Since our Copilot X announcement in March, our team has knuckled down and evolved the most profound breakthroughs I’ve seen in my lifetime.
— Thomas Dohmke (@ashtom) October 13, 2023
I am a skeptic German, I can 100% say: don’t miss GitHub Universe. Tectonic shifting stuff ahead.
I mean seriously, check out Co-Pilot correctly generating dozens of lines of Klaviyo API call code. How incredibly cool and powerful is this?
How about generating Luxon function calls based on a comment. Say goodbye to looking up documentation.
Or a task that many codebases need to have solved, weighted averages. Why bother writing this by hand when you can hit Tab and a magic solution appears:
Yup. The game has been changed forever. There’s no going back.
At Essembi, we’re all in on AI. This is not a buzzword or the “flavor of the month” for us. Not only do we want to deliver the most powerful, configurable, one-stop-shopping solution for software shops, we also want to deliver the most powerful management co-pilot for all roles in these types of businesses.
I said earlier that I felt 33% more efficient. Just like with Co-Pilot chat, I think this 33% is the floor. If this is the floor, what’s the ceiling? Technological progress marches on from fire and the wheel, through the steam engine, the computer, the internet, and now AI.
Nobody likes dealing with issue management software, let’s be honest. It doesn’t have to be that way. People view tickets and tasks in management software as administrative work that gets in the way of the “fun stuff.” Our goal is to change that by bringing at least 33% more value to each role with AI.
Whether it’s authoring test cases, doing competitor research, spotting bugs, etc., our goal is to remove the minutia and let smart people focus on cool and interesting problems.