I spent years inside a security company as it scaled from scrappy to enterprise — then chose to run the loop again, on purpose, at an AI startup.

Not uncommon in tech careers: you join something small, it works, it scales, and the enterprise stage is reached. Quite often in hyper-growth phases, it happens quite quickly, like 5 or 6 years in. Many people then left, telling you on the way out that this is no longer the environment they joined, not what they wanna do anymore, and need to go back to smaller team and structures.

The business journey

I'd say that's fair in most cases, as many many many startups turn bureaucratic quite fast. While it doesn't happen without reasons, it does happen, and not everyone likes it (not sure if anyone does really, but I can't say). Whatever the needs driving this transformation, it is often necessary and the execution can be rough, usually at the expense of what I call "innovation efficiency" in favor of actual cost reduction. Paradoxically, the business becomes leaner in terms of dollar amount, but tends to slows down its pace of evolution. Maturity & stability is preferred for the now enterprises customers, the biggest payers. Achieving this balance is a challenge for every growing company. Senior leaders from large companies then join, replicating the same patterns they applied at their previous company, usually a global scale enterprise. The transition is often painful as it disrupts the culture that made the company successful in the first place. Not everyone agrees that is the right move. And others understand, generally agree with the "why", but question the jump in scale and speed to be too drastic.

In that context, keeping the innovation efficiency while scaling is a challenge. Growth remains the lifeline of the company that is eyeing an exit. And acquisitions usually make up for the innovation drag. Most acquisitions are really painful, both on the human and technical level, requiring the adaptation of different technologies, cultures, ways of working, and so on. Sometimes I wondered if rebuilding it all from scratch wouldn't be faster than going through all that. It's trading one type of pain for another, really. Except that innovating is extra hard as stability is the primary goal of the existing business.

Innovating

Innovating while the business is maturing is hard. Hard to move fast when stability goals are making changes harder, more procedure oriented, without even talking about the technical debt you're fully paying at that time. Hopefully the business built the right fundamentals to technically future forward, probably around a platform that can be extended without breaking the existing features. While you're doing that, new startups are popping up, trying to solve the same problems with newer technologies, newer approaches, and often with more agility and less or no technical debt.

It takes a special kind of leadership to create an innovation arm within the business, insulating it enough from the rest to foster innovation and a faster pace. Still not as fast as startups, but fast enough to find to make the product evolve in a more natural way than acquisition, and keeping the culture alive to retain the talent that made the company successful in the first place.

That's how you break the loop for many people. You keep the setup allowing them to do what they like doing the most, researching and learning new things using their experience and expertise, shielding them from the bureaucratic overhead that slows everyone down. It is exactly this process that kept me happy at Snyk for such a long time. Learning new things, making an impact, while having fun doing it. Not quite the same freedom/pace as a startup, but close enough, and you get the scale of your impact significantly higher.

That's pretty much a build; scale; repeat. loop right there. That's why I hadn't considered joining another company out there for all this time.

Over the edge

So what pushed me over the edge?

4 things:

1. The People
2. The AI era
3. The Risk
4. The Building

The People

One of the perks of experience in the industry is the network you build. Working with great people, some more closely than others, give you a sense of who is great at what. I've met amazing people over the years, with many extremely capable at what they do. But not all are builders at heart. Operating in uncertainty, loose requirements, with less supporting structure, is not for everyone, including some of the top performers. It takes a bit taste for risk, entrepreniarial spirit, and the ability to thrive in ambiguity. You get to identify these traits in people you work with. You can feel who you could start something from scratch or close to scratch with. And that goes both ways. So when the right people call, you know they surrounded themselves with the right people too. By extension, you know the kind of crew you're joining for the ride.

The AI era

AI is changing everything, to level much deeper than most people can fathom. Much deeper than I can fathom myself. Models are becoming commodity, sure. The shift is the way to use them. Extracting as efficiently as possible every ounce of capacity from them, using expended energy as efficiently as possible. The Frontier labs and the surrounding industries are leaning hard into the infrastructure play. Like really hard, hard like we've never seen. The capital thrown into it is just insane. But the technologies involved are just incredible. It's really the closest thing to magic. And when you see how ASML is doing the ultra violet lithography, you start seeing the fabric of reality bend, or maybe it's just my mind. AI is infused everywhere, and jumping into a pure AI company is my way to drink from the firehose, without restraint..

The Risk

Startups are by default dead. The first challenge is to survive. The next challenge is to survive. The following one is to keep surviving. And so on. If it survives long enough, that means it's making enough money to keep going, and if it keeps going, maybe you can make some money. People start companies because they can't help themselves. They have to do it. They have to take the risk. Usually, it's a bit pathological. They want to produce, change things, improve the world. It's easier to join an existing company and join the ride. The existing company is also in a fight to survive, but the size is such that you don't feel the pressure as much. Delivering value to customers is what keeps it alive. It's what keeps any company afloat. When that value disappears, or is too little compared to the competitors, the company dies. Only governments can survive without delivering value to customers, until society dies supporting it. Then it's back to square one. Difference between the 2 is in that in the latter case, people usually die. Historically, a lot of people. I choose to be part of the attempts at improvements, to be part of the change, to be part of the future. It's fun, it's rewarding, it's challenging. And it's the right thing to do. The what humans do. Not the only thing, but one of the best things in my opinion. That's how humanity progresses, discovers new frontiers, and keeps moving forward.

The Building

I just love building things, figure stuff out, helping my peers, contributing, helping customers, all of it. Not that I wasn't doing that before, but there is a certain thrill about doing it closer to the metal, with less abstraction, less layers of indirection, less bureaucracy. Obviously it has to be for a vision in which you believe. If you don't, I actually don't know how you actually can build anything beyond a few months. I joined Tenzai because autonomous, pairable AI hackers is key for the world to stay secure, and good actors not to be outgunned.

Repeat

So this is build; scale; repeat.

If you are in the middle of a similar stage — staying for scale, leaving for speed, unsure which itch is real — the only advice I trust is the boring one: notice which kind of day you want more of, then optimize for that. Titles catch up later.

More things to come from here. Keep an eye on this blog.