Phil Libin exits General Catalyst for All Turtles, a new AI ‘startup studio’

AI is one of the buzzwords of the moment in the world of tech, with startups coming at the concept from all angles — computer vision, machine learning, unstructured data inference and natural language processing being just a handful — in a wider effort to create more intelligent machines. Now comes a new organization that hopes to find and foster the next wave of AI businesses and products, co-founded by the ex-CEO of Evernote, Phil Libin (pictured above), who has left his role as a managing director at General Catalyst to build it (but he tells me he’ll stay on as an advisor).

All Turtles, as the new company is called, is not your traditional startup incubator. In an interview with TechCrunch earlier, Libin (whose other co-founders are Jessica Collier (Product Design) and Jon Cifuentes (Research and Operations) described it as “startup studio”, more akin to Netflix’s push to develop original content than to 500 Startups. It will start out with locations in San Francisco, Tokyo and Paris.

The idea with All Turtles is not necessarily to seek out fully formed businesses, but to find and develop interesting concepts en route to them possibly becoming businesses (or at least good ideas that other businesses might like to buy), with people and startups collaborating together to collectively grow. (This is also where the “All Turtles” name comes from: an apocryphal story about how the world is formed on the back of other’s work, that it’s a group effort in the end.)

This is not Libin’s first attempt to foster a group of AI startups: last year, after he parted ways with Evernote, Libin joined General Catalyst and kicked off a new project there to find, fund and grow startups building bots — tools that chatted with you, a human, using conversational artificial intelligence, to help you find information, solve a problem, order a sofa, and much more.

In case you are wondering (I was), Libin said that his departure from General Catalyst and subsequent move to develop the idea further in a separate studio was his decision — a way to dive deeper into the development angle beyond the primary financial one of VCs.

“Phil’s an incredible product guy and we’re excited to back him at All Turtles where he can focus on bringing awesome AI-driven products to market,” a spokesperson from General Catalyst told TechCrunch in an email. “Phil will continue on with General Catalyst as a senior advisor and will continue to work with many of the young bot companies he funded as an investor here.

There will also remain several strands connecting Libin, the startups and GC.

For starters, several startups that Libin backed as part of that bot effort at General Catalyst are now joining All Turtles as foundation members.

The list of foundation members include Butter, a “teams assistant bot” that works in Slack that helps you find information across multiple files as you need it; Growbot, a “recognition” bot to commemorate achievements and milestones at work; Edwin, a Facebook Messenger bot that tutors you in English vocabulary; Replika, an AI “friend” that acts as a companion and journal; robot lawyer DoNotPay; Loic Le Meur’s talent finder Leade.rs; social selling bot Octane AI; and home security and drone startup Sunflower Labs. There will be more coming on that will be announced soon, Libin told me.

On top of that list, General Catalyst, along with Japan’s Digital Garage, are coming on as backers of All Turtles. Libin declined to disclose the amount they are investing, or any numbers: more backers are going to be revealed later in the year, when All Turtles will also disclose more details as well.

Libin said that as part of staying on as an advisor at General Catalyst, he’ll also continue to be involved in the investments that he led for the company.

I had a conversation with Libin about what led to him starting All Turtles, what he hopes to achieve, and what he thinks about AI today. An edited version of it is below:

Why exactly did you decide to leave General Catalyst? You were already running a seed fund for bots, so didn’t you just develop this within that existing structure?

I’d been thinking about this for a while, for years really. It’s the job I thought I would have when I joined General Catalyst. I want to expand what entrepreneurship is. Top tier VCs are not the best place to do that, being adjacent to VC is best. I think this has worked out great. Many of the portfolio and studio companies are the ones I invested in General Catalyst, so it’s a natural progression.

It was my decision to do this and they supported me as primary backers of the company. I just realised that it’s more than investing. Just investing is not really enough to change how the industry works. I wanted to create a hands-on, more hybrid model.

You started with an effort to seek out bots. Now this has moved to AI but you are still primarily listing bots in All Turtles’ initial list. What is the roadmap here?

The first theme for us is practical AI bots, and conversational AI is a starting point. But I think this will moves quickly. Sunflower for example is making AI-based sensors and drones for security. So there is also a hardware component there. There are also a couple of others that haven’t been announced yet that we will be backing. 

Do you feel like AI has become too hackneyed?

AI is a buzzword similar to what mobile felt like to me when we started Evernote. Today, everyone is still talking about AI in a way that is similar to how people used to talk about mobile. Years ago, it was still possible to seek out and invest in companies building mobile apps. Now mobile is a part of everything and every company develops a mobile version.

I think something similar is happening with AI. It means a specific set of algorithms and technology, but within three to five years it won’t be “AI companies” anymore. It will fundamentally be baked into everything.

And that’s what this is about for us: we want to catch something before it goes mainstream. 

You are opening from the start in three geographies, San Francisco, Tokyo and Paris. A lot of VCs in Silicon Valley tend to like to invest more locally. What’s the story there?

I wanted to do this internationally from the beginning. I miss being in other parts of the world and talking to entrepreneurs there. We wanted to be in more than just San Francisco, and it was important to get great investors in Japan and Europe and Asia to go into all those areas.  

Do you see other “studio” models a la Betaworks or Science or Expa as your models for this?

We are trying to learn as many lessons from the likes of them as we can but I think what we’re doing is pretty different. We’re modelling ourselves after Netflix rather than startup incubators. But rather than making the world’s best original TV shows, we want to make the best AI products. We want to attract the best people to produce these in a professional setting and help them with the distribution. It’s probably more Hollywood than Silicon Valley.

The other thing is that the goal in the tech industry is still to make companies. Accelerators and incubators have a company fetish. It’s all companies and a company-first mentality. But in AI, the company isn’t necessarily the interesting thing, it’s the product. We want to focus on that first. Some will become companies, and some will not. We want the best.