Axios C-Suite: Confessions of an AI lab rat

• In short, AI is way better, more accurate and mind-expanding than most think. But it's colliding with hard human and business realities — making it confusing, clunky and chaotic for most of our people in its current form.

My one-year-in lesson: Be more curious than ever and more humble than ever.

• Experiment and explore aggressively, but remember that NO ONE has a clue about how AI will unfold outside and inside our companies in the coming months, much less years.

A quick primer on my experiment parameters: Starting a year ago, we've aggressively tested AI (mainly OpenAI's Codex and Anthropic's Claude Code) across every layer of every department at Axios.

• We provided access and instruction to every employee. Most of my leadership team operates with chief-of-staff agents, and we're knee-deep in agent-to-agent prep.
• On a personal level, I use ChatGPT or Claude for one to two hours daily, usually in the early mornings, and control an AI personal operating system via my phone. That's connected to an always-on computer that runs several agents, including one that scans daily for CEO-related data and trends.

Here are my blunt takeaways for how AI impacts CEO-level work after all this experimenting:

1. It's exceeding expectations at the individual level. I've spent six months with my head buried in this, while talking to the smartest people in tech, politics and business. And I can tell you: AI is smarter than 95% of the people I know on 95% of topics. Even for someone using it obsessively with real discipline, I'm still discovering it's way better than I thought possible. Its ability to think creatively and research deeply is extraordinary. But only if you know how to use it. The point isn't to outsource your thinking to the machine — it's to expand your own. Used right, AI doesn't think for you. It makes you think bigger.
2. It takes work. You can't wing it or outsource it, even at the CEO level. You need to work at it daily — so AI learns you and you learn AI. This is when the magic happens. You have to feed it copious information and tell it what works and what sucks. This feedback loop creates a new form of super-institutional knowledge made up of the accumulated context and memory you build into it over time.
3. You need to go beyond chat. The life-changing work happens in projects or with agents. You need to feed in files of information and direct it with specific skills on how you want it to write, think, fact-check and cite actionable research and intel. This is where your company is headed. Only by doing it yourself will you discover new ways to apply it internally. And you can't delegate this. Your team copies what you do, not what you mandate. Adoption rises and falls based on whether the CEO is hands-on, especially at smaller orgs.
4. The last mile sucks. As good as it is, AI hits internal walls when it comes to security, connecting to other systems and deciding what data it can access. This is a massive issue right now, which is slowing any internal gain. In most cases, it's simply not ready for scaled deployment. This problem is getting WORSE, because ...
5. Agent-to-agent work is a mess. If AI transforms our business — and I think it will — agents need to work flawlessly with other agents. This is the unfolding frontier. My exec team has chief-of-staff agents, but we hit constant walls in determining what they can know, share and act on once they collaborate. Just think about all the permissions you need to nail down so this doesn't backfire. The AI companies know this and presumably will solve it. Until then, efficiency gains are overhyped. Our job as CEO is to live in that uncomfortable middle: don't believe the demos, don't dismiss the trajectory.
6. A new class of superworker is born. Here's the best news: We're spotting rank-and-file workers daily whose brains are wired for AI. It's been easier than expected to train them to be AI enablers on their team or companywide. If you don't have an AI enablement team now, build one ASAP and create a mechanism to spot and elevate these AI-wired brains. These people are NOT technologists by training.
7. You'll discover new business lines. A year ago, I thought the big win would be cost savings and productivity. Automate ruthlessly, cut expenses, expand margins. That's real — we've done it. But over the last three months, my view shifted. The bigger opportunity isn't efficiency. It's new business lines that were economically impossible before AI. We have at least three new revenue-generating projects underway that simply weren't possible without AI. And here's what separates the durable ones: They sit where raw AI output still needs a human-trusted layer with judgment the machine can't supply alone. Many jobs will change, but I now believe we'll hire more people over time than I'd have guessed a year ago.

The bottom line: I've never been more personally productive, more enthusiastic and more curious about hands-on experimentation.

• By the end of this month, I'll have personal agents helping with my day-to-day work, researching a book, building out a new business line and helping solve a medical mystery for a family member.
• And don't forget: I went in as dumb and blind to how tech works as anyone reading this!

📈 If you're a CEO or on a CEO's team: Ask to join Jim's new weekly Axios C-Suite newsletter.