How AI is 'terraforming' a new Grindr - the internal and external impact of a tech revolution in action

I oftentimes say that Grindr became successful in spite of its marketing rather than because of it. We, frankly, we're not really focused on marketing at all and not really paying attention to it. 

An interesting admission from George Arison, CEO of dating app Grindr, but one that is not about to distract him from his goal of “rapidly terraforming” the company }into an AI-native organization. Progress on this ambition is going well, he argues, with up to 70% of coding currently driven by AI agents so the firm is able to ship faster with higher quality without the company getting heavier and slower.

With the debate around the human intelligence/AI balance ongoing, Arison stakes a claim for having been arguing the case here for “a lot longer than almost every other executive in tech”. That’s a bold boast, which he backs up with:

I said at a conference in the fall of 2024 that there will be a time that when there are synthetic employees working alongside humans inside companies, and I got a lot of flak for that. But I think no one denies that that's going to happen anymore. Synthetic AI agents or employees are going to be a fundamental part of our work on a go-forward basis, and we are seeing the impact of that now.

At Grindr, we have been at the very, very forefront of adoption of AI in our day-to-day work, especially in engineering. I'm pretty confident in saying that we're probably in the top five percentile of companies in tech in terms of how quickly we're adopting to that and how quickly we're terraforming to being an AI-native organization inside the company.

That will mean a change in organizational demographics, he admits:

I believe that there's going to be a time when almost all the code that we produce will be written by AI agents. That does not mean that engineers don't matter. Engineers actually matter even more now and awesome engineers matter even more because the really great engineers are able to take advantage of these tools even more than anybody else's, making themselves even more valuable because they can do so much more.

The concept of a 10x engineer is now becoming 100x engineer because one 10x engineer can do 4x, 5x, 6x, 7x, 10x engineers' worth of work as a result of what the coding agents and AI-based synthetics are able to do for him or her when they are writing code.

And the way code is being written completely changes, he says:

Before, I would come in and I would have a new project and I would think about [how] I’m going to assign work for it. I want to give this much to this person and this to this person, and it would take each of these two individuals six weeks to write the code that they were going to be working on for that particular project. I would spend a lot of time helping them be more successful, plus writing my own code, and then bringing it all together.

Now, I sit down with an agent, I start talking to it about what it is that I'm working on, I send it off to have them do the work, and within two days, that whole thing is done.  So the speed by which we're going to be producing code is increasing rapidly. Internally, we've seen about a 1.5x increase in productivity per engineer.

This all supports Arison’s thesis that Grindr is uniquely positioned to be able to be an AI-first company, and AI-first product because it has so much data, he argues:

AI is good theoretically, but if you don't have the data, it can't really do very much. We do have a ton of data that we can utilize.

Outside view

That’s the internal impact of AI. From a customer-facing perspective, AI-enabled innovations include the firm’s Edge offering, says Arison:

The things that we're trying to solve with Edge are twofold. One is that people end up having many, many conversations inside Grindr, which don't go very far, partly because new conversations take over. So because we're an open architecture platform, people can talk to anybody and people have many, many conversations at once. Power users, in particular, have even more conversations, but an average user sends 50 messages a day. So you have many people that you're talking to all the time.

That's really magical, and that's where a lot of the excitement of Grindr comes from. But one of the negatives of that is that some of the great conversation you might be having get kind of pushed down and lost in the inbox. The thing we've been thinking about over the years is, ‘How do we avoid that from happening? How do we help the user have a better sense of,’Hey, these are the conversations that I'm having, and I want to maintain them’ and maybe they go somewhere beyond just the conversation over the long term.

That's especially true for users who travel a lot because you might be having conversations in many different places, he goes on:

When you're in New York, you have a bunch of conversations in New York, then those get lost when you go back to, say, Chicago and are having conversations in Chicago. What we've built within Edge is a product called A-List, which takes your entire chat history and builds on top of it a set of summaries of the richest and the best conversation that you've had with people that the AI believes are your best matches. And then you can go to your A-List and see those conversations and brings together those conversations. It brings together the summary, it tells you what you told them, what that person told you and why that's interesting.

What is the important information that you shared about each other, whether it's your name or other relevant information, etc. It brings together the other person's photos as well. So you can see all the photos that he has shared with you or anything that you've shared with him. That feature is a killer feature. The users really love it.

The second issue Edge tackles is discovery.  Arison explains:

In almost every location in the world, the number of gay people in a given geography is actually quite limited because we're about, what, five percent of population, maybe six percent, and that's not that many people. When you take half the population is male and then five percent of that is gay. Maybe in New York, that's an exception where you do have a critical density, but everywhere else, density is lacking.

So, through a feature we call Discover, we're able to identify and surface people to you that are the right matches for you, meaning we believe you will like them based on everything we know about you and everything we know about them, but you otherwise might not find, and hat’s less contained by geography where Grindr is very geography focused...That allows people to find new people that they otherwise might not connect with.

But because it is based on all this information that we possess, it's actually a very positive recommendation and because it's transparent because of insights, there's actually a desire on the person's part to engage in a conversation and take a risk on a longer distance because there's so much alignment of interest. So that's what Edge is doing.

My take

A recent analyst call saw a large chunk of the prepared remarks read by AI using a proprietary voice model trained on Arison’s voice. He says:

We did that deliberately as a small demonstration of how deeply AI is becoming embedded in both our product and our operations. 

Actually I found it rather annoying and a dodgy marketing stunt - see opening remarks! - but no-one can accuse Grindr of not being committed to its AI transformation.