Exponential Operator is a 100% human-written newsletter by Sawyer Middeleer, founder of Revi Systems. If you were forwarded this email, join some of the sharpest AI operators by subscribing here:
Hi all,
So I was gone for a little over 3 weeks. Family stuff, yada yada. I only had bandwidth for the barest of bare minimum work.
When I reopened LinkedIn and X for the first time in nearly a month on Sunday afternoon it was like the first day of school after a long summer break.
Oh hey, a new Claude model dropped. Everyone’s using Hermes I guess. Codex has Tamagotchis now??
Not sure if I need Clippy 2.0 but they’re cute
It was disorienting.
I realized this is a mini version of what it must feel like for most people who haven’t worked in the AI startup vortex like I have the last 3+ years.
Just when the mainstream discourse focuses on one tool, the tech world has already switched its attention four times to newer shiny objects. And they all want you to believe that your career and your company are cooked unless you also master this automation tool or that vibe coding agent.
Take a beat right there.
It’s okay to be behind the hype cycle. In fact, many of the most successful AI operators I’ve worked with aren’t up to date with all the latest product releases and models.
They understand that always chasing the new hot thing is a recipe for wasted time and burnout.
Instead of spending hours online debating whether MCP is dead, they focus their attention on the tools that get the job done.
You don’t need to buy every new $200 subscription. You don’t need to use Claude Code and Codex and OpenClaw and Hermes and …
Instead, there’s two paths I would recommend to anyone who aspires to integrate AI tools deeply into their work, but feels overwhelmed.
First is start from a problem you really need to solve. For me, that’s reaching new potential clients. For you, maybe that’s triaging support tickets or optimizing PPC ad spend.
Before you touch any tools, start by using the Jeff Bezos working backwards thought experiment. Imagine the outcome you’re looking for - maybe that’s handling more tickets without hiring more CS agents.
Then break down the gap between that outcome and your current state into discrete steps, such as
identify the most automatable tickets
activate an AI agent in our support platform
teach it our support policies
start automating tickets
optimize the system
At this point you can go test out AI tools to find the best one for the job. Get really good at using that tool and you’ll be ahead of 95% of people who know all the latest stuff but can’t do anything practical with it.
The second path you could take instead is to begin from a curiosity you’re dying to investigate. Maybe you’ve always wanted to build an iPhone app, or you’re really interested in cyber security. Go deep down that rabbit hole and try out one or two of the leading products in that space. Build an app, read the research (or have an AI summarize it for you).
Working with AI tools in different domains is actually very similar. So if you get really good at creating videos with AI, for example, you’ll also simultaneously become better at building spreadsheets with AI. The principles of context management, prompting, and all the rest carry over pretty well.
Regardless of which path you take, the core principle is the same: focus your attention on what matters and you can’t go wrong.
That’s all for today. Thanks for reading!
江湖再见🫡
-Sawyer