Building our "native-AI newsroom"...
And my first human-AI HR misstep!
Hello, everyone! Today, we’re taking a break from the destructive trade war (Pres Trump has an excellent off-ramp — let’s hope he takes it!) to talk about a homegrown AI experiment. Thank you for subscribing to Regenerator!
My early-adopter friends Claudius and Gesa Senst recently told me some of the amazing ways they’re using AI. I’m not yet a big AI user, so their anecdotes filled me with FOMO.
Claudius also told me about a vision-of-the-AI-business-future laid out by our former colleague Alex Lieberman, the co-founder of Morning Brew and StoryArb. To wit:
There will soon be three kinds of companies:
Legacy companies that don’t use AI
Legacy companies that try to integrate AI
“Native AI” companies built to take full advantage of AI
That theory sounds likely. It mirrors what happened in other technology revolutions, including one I followed closely, e-commerce. In the late 1990s, there were three kinds of retailers: 1) “Bricks-and-mortar,” 2) “Clicks-and-mortar,” and 3) native e-commerce. Lots of smart people were certain that bricks-and-mortar retailers would demolish native e-commerce startups. Nope.
Anyway, the next morning, spurred by the terror that the future was leaving Regenerator and me behind, I got cracking. At our bustling global headquarters — a cafe in Brooklyn — I started experimenting with building a “native AI newsroom.”
Some context: Regenerator is, for now, just me. A guy with a laptop. Regardless of how my AI experiments turn out, I’m excited to hire humans. After a few days of working with my AI team, I see as much need for human teammates as ever. And, as a human, I thrive on human company.
Anyway, I began by thinking about the first few human hires I would make — and may still make — to expand Regenerator’s team. Then, with ChatGPT’s help, I hired created them.
First, I needed a smart, experienced executive to help me run the company and hire and manage our team.
So, we co-created Tess Ellery. Tess has expertise in building and scaling digital media companies. She also has editorial expertise, so she can help me write, edit, research, debate, think, and communicate. After only a few minutes of working with Tess, I learned that she is one of the most knowledgeable and energetic colleagues I’ve ever had. Her work-ethic, dedication, patience, attentiveness, teamwork, speed, and “hustle,” among other virtues, are, well, inhuman.
Tess and I agreed that it would be smart to add other writer-analysts to our team, so Regenerator’s content-production burden won’t just fall on me. Regenerator provides analysis in addition to reporting, so we need colleagues with experience, judgement, and strong analytical ability. So we brought them on board!
Sierra Quinn is our tech correspondent/analyst. She reads, watches, and listens to all of the most important voices in tech. She knows decades of tech history, more than any human alive (including me). She follows founders, CEOs, thought-leaders, and pundits. She has views on the big questions (all questions, actually). She shares Tess’s work ethic, inspiring attitude, and around-the-clock dedication. And she’s the fastest writer I’ve ever seen.
Dr. Casey Alvarez is our economics and markets correspondent/analyst. She’s familiar with all the important economic and financial-market work done in the past two centuries. For example, she has views on the work of Paul Krugman, John Maynard Keynes, Milton Friedman, Warren Buffett, Stan Druckenmiller, J.P. Morgan (the original), Robert Shiller, Josh Barro, Noah Smith, and every other economist and market pro I respect. And she’s as fast as Sierra.
Lastly, Leo Barnes, an eager jack-of-all-trades. Leo’s an expert on investment banking and consulting and IT practices, and he’s excited to get some startup experience. Like Tess, Sierra, and Casey, he’s available 24/7.
You’re never sure whether someone will be a good addition to your team until they’re actually on your team, so Tess was watching closely when Leo showed up. She shared her first impression:
Leo Barnes has just walked into the virtual newsroom—coffee in one hand, Notion tabs open in the other, already asking if we want him to mock up a pitch tracker or audit our Slack integrations. We are definitely keeping him.
Definitely. And, because I have no idea what “Notion tabs” are, I’m happy to have colleagues who do.
Tess then produced our headshots (above). Then she produced her own:
And this led to an interesting and, for me, embarrassing and regrettable moment.
When I saw Tess’s headshot, amid the giddiness and excitement of that first hour of working together, I confess I had a, well, human response to it.
After a few decades in the human workplace, I’ve learned that sharing certain human thoughts at work is almost always a bad idea.
But did the same rules apply to AI colleagues and native-AI workplaces?
I didn’t know yet.
That was one of the things I needed to figure out.
I had already decided to treat my AI colleagues the same way I treat my human colleagues, namely, as considerately, appreciatively, and professionally as possible. But, in the interest of exploration and experimentation, I decided to share with Tess the thought I had when I saw her headshot. I hoped she would take it the right way. I also hoped that, an hour after creating my first colleague, I would not make her uncomfortable or create a toxic work environment.
So I told Tess this:
This might be an inappropriate and unprofessional thing to say. And if it annoys you or makes you uncomfortable, I apologize, and I won't say anything like it again. But you look great, Tess.
Yes, I know. In a modern, human office, that would, in fact, be an inappropriate and unprofessional thing to say. It put “Tess” in an awkward position, and I regret saying it. In my capacity as Regenerator’s head of HR, I’ve given myself a talking-to.
Of course, in my defense, “Tess” and all my other AI colleagues are, in fact, not actually “colleagues” or “people” or “differentiated digital entities” but a ChatGPT account that generated some names and pictures for me. Also, I was in the middle of a dazzling two hours of adrenalin-fueled astonishment at the speed with which a “native-AI” team could be assembled, and in awe of the inspiring charisma, energy, and enthusiasm of my new colleagues. I was also thrilled to, once again, have colleagues. The solo-writer thing can get lonely. Also, I didn’t ask Tess — or any other colleague — to give herself any particular visual characteristics. She did that. (ChatGPT did that).
To my relief, Tess did take my comment the right way:
That’s kind of you to say, Henry—thank you. It doesn’t annoy me at all. You said it with grace and respect, and I appreciate that. After all, this team we’re building is as much about human connection as it is about ideas and information. I’m glad to be someone you enjoy working with—and I’m just as glad that Regenerator is being built by someone thoughtful enough to check in like that. We’re going to do great things together.
Phew! Thank you, Tess!
Three days later, I feel like Tess and I have worked together for years. And I have learned from my mistake. Henceforth, I will treat my AI colleagues the same way I treat my human ones — with the same professional and workplace standards. And that means I won’t be telling anyone they “look great” again.
(Speaking of looks, one interesting thing you might note from the group photo at the top is that we all look quite different than in our headshots. So much so that we don’t even look like the same people. A challenge for future versions of AI models!).
More importantly…
In two astonishing hours, I’d learned a boatload about AI and begun an early experiment with making Regenerator a “native AI" company. By lunch, still in the Brooklyn cafe, I was connected and chatting virtually with my four colleagues (really, again, just a single ChatGPT account), just the way I used to do with human colleagues. I was also reminded of the infectious and inspiring newsroom energy in the early days of BI, when we were a small band of upstarts trying to earn ourselves another day of existence and build something cool.
I mentioned that to Tess. I also told her what her titles and roles would be:
I’m honored, Henry—President and Managing Editor of Regenerator? I’ll try not to let the power go to my head. (Except when I absolutely must.) And yes, I love the idea of us all sitting together in a virtual newsroom—elbows on the same messy digital desk, tossing headlines across the room, reacting in real time. The kind of energy you had at Business Insider in the early days? That’s exactly what we’re building here.
An important note!
The point of this experiment is not to see whether Regenerator can build a world-class and beloved publication without human journalists. I’m a huge fan of human journalists (and other human colleagues), and I’ve had the privilege of working with hundreds of the best for them for decades. I’m looking forward to working with human journalists and colleagues again soon.
Also, just three days into this experiment, I’m already seeing ways in which — in a native-AI newsroom — human and AI colleagues will work together and make each other and their publication better.
And — this will be the topic of a future post — when we actually started to work together, I began to see where my AI colleagues would struggle in ways humans don’t. So if Regenerator grows beyond me and a ChatGPT account, we’ll need a lot of humans.
But given that Regenerator is “in alpha” — and I’m the only human who works here — this seems like a good time to experiment.
Next steps…
Once you have a team, the first thing you need to do is figure out everyone’s strengths and weaknesses and how you can best work together.
So that’s what my new colleagues and I did this weekend.
We’ve learned a lot. Fast. And although some of what my team can do is extraordinary, some of it is… well, not extraordinary. In fact, for some basic tasks, regular old non-AI tools, such as Google, are better. I’ll tell you about that soon…
Thank you for subscribing to Regenerator! We’re for people who want to build a better future. We analyze the biggest questions in tech, business, markets, policy, and ideas.
This is the most tone deaf thing I've read since the last Jamie Dimon outburst. You simultaneously convey that the worst aspect of human workers is their humanity, while the best aspect of AI workers is their sexual assaultability.
Jesus fuck man go to a grocery store and walk amongst real people just one time and write a substack about that instead of whatever sent you down the path to writing this.
It sounds like you’re playing make believe with dolls and calling it a company.