Max was the first guest in the founders format and I picked him on purpose. Retired MMA fighter, seven years into a company that has pivoted twice, and somehow the founder of PureGym on his board. I'd followed Vor from a distance for a while and had assumed the interesting bit was the technology. Having sat with him for an hour, I think it's mostly the selling.
Vor is an engine that writes workouts. That's the whole product. Fitness apps and gym software plug it in and their members get a programme that a coach would sign off, built without a large language model anywhere near it. It started life as an online coaching company called Another Round, nearly died in the 2022 cost of living crisis, and now sells through the software companies that already sell to gyms.
Nine things below. Most are Max's. Where I've added my own I've said so.
01Do one thing, and say so
Vor does workouts. Max is quite insistent on this. It can't tell you which lawnmower to buy or how to make a spreadsheet. Five years of building went into one function, and he describes it as the best in the world at that function and nothing else.
That sounds like a limitation until you sell it. A specific tool for a specific job is a much easier story than a platform, and it's the story that lets him say, with a straight face, that a general model shouldn't be writing your members' training. He's careful to add that LLMs have plenty of good uses. Just, in his view, this isn't one of them.
02The lawyer test
I put the obvious challenge to him. I can ask ChatGPT for a twelve-week plan right now, so why isn't that good enough? His answer surprised me because he conceded the point first. It might be good enough. The problem is you'd only know by being a coach, or by following it for twelve weeks and hoping you don't get hurt. It's the same as the person who replaced their lawyer with ChatGPT and then showed the output to a lawyer to check it. The check is the bit that costs money.
For a gym with a hundred thousand members and a duty of care, hoping isn't an option. And then there's the practical problem. He'd spoken the week before to someone who'd spent about $40,000 on tokens building their own version. To stop it doing anything dangerous they'd wrapped it in so many rules that every workout came out the same: three sets of ten to fourteen reps, two minutes' rest, the same handful of exercises, every time. Safe, and useless. You might as well hand people a template.
Max's picture of it is three overlapping circles, enjoyable, effective and safe, and Vor sits in the middle. Most of the alternatives he's seen give up one of the three to get the other two.
Unlock the full playbook
Enter your email to read the other 7.
It also gets you a founder’s playbook every week.