Why Bother to Bootstrap Your AI Startup?
A quick side note on nomenclature… Should it be A.I. as Google Experiments puts it, AI as Wikipedia puts it, or A.i., which I have often seen and makes me cringe a bit just thinking about how a Natural Language Processing (NLP) engine interprets the lowercase ‘i’. This article is going to be all over the place, so buckle up.
There is a weird dilemma emerging in the AI startup space when it comes to seed and venture funding. In this article I will take you through the mindset and decision space, and talk about my experience and that of my clients. And other stuff too.
I decided to write this article in response to a comment by Tom Marsh on a previous article:
Daniel, I would add the decision to seek VC money as part of your discussion for many of us in AI. When the field is this hot and VC money in large amounts is available for AI projects, the decision dilemma is similar. Do you pursue solutions/plans just to take advantage of VC capital or stick with a plan you’re passionate about that might accommodate a bootstrapped strategy? No easy answers, just have to pick the one that fits you best. I fall back on Jim Collins’ advice from Good to Great: Pick the one that you’re passionate about, that you can be the best in the world at and feeds your economic engine. If you compromise on any of those you will pay a price and probably be miserable. -Tom
My consulting firm is bootstrapping, in the sense that we are not venture backed, but that is more by choice than by necessity. We fund our own startups in our “spare” time. Stuff like clockrr.com, genrush.com, and others. As Tom points out, a lot of the decision to go bootstrap rests on your personal preference. As a 50% equity holder with no debt, I can do basically whatever I want. I can take a day off, refuse a client, or just do a pro bono project for fun. It’s up to me. I’m not screwing over any investors if I fall asleep at my desk. (Yes, that does happen)
The same goes for Mathieu Lemay. We get to have a good time.
My non-enterprise clients (i.e. startup, early stage, whatever) are faced with a bit of a tougher decision. They need money to scale up, or even to develop. There is plenty of capital pumping into the AI space, but there is also a low capital cost of getting into the AI space. You just need to know what you are doing and what you want to do… And then you do it.
The mantra has been for some time now to “ grow fast, lose money, go public, cash out” and to do it primarily with other people’s money. I watched a really great lecture by Dan Lyons, a Newsweek guy who landed at HubSpot.
At HubSpot, Dan observed the absurdity of all this growth mentality, and how unhappy people really were, grinding it out to the tune set in stone by investors.
Hopping in bed with investors is not always bad. It can be wonderful. You often have your visionary seed-stage angels who are already wealthy, and just want to make the world a better place. You also have your late stage players that come on in and place you at the grownups table with a public listing and a big exit. Unfortunately, in Canada we really suck at venture funding. Seed funding is very hard to get too. Mostly, Canada is just very economically conservative and risk-averse. In reaction, there is a real pull for Canadian companies to head south. Basically, to chase the money. In all fairness though, Toronto is #20 on the list, so it is not as if you can’t make it as a Canadian AI startup. Also, Canada is the world center for AI research. No reason to head south if all the talent is back up here in the snow and ice.
There is also a work-life balance to consider. Jeff Bezos calls it work-life harmony. Things are moving very fast, and you may not want the pressure of a high burn rate cash-destruction-machine nightmare keeping you up at night. There is stress in bootstrapping too. My time is always scarce. I feel the shallowing of my soul, as I am moving from full length paper reads down to 2 minute video summaries… Squeezing every ounce of efficiency from every day.
My motto these days is:
while(True)Think;DO
Just do what feels right to you.
Why are investors so crazy about AI? Fear.
Ask your blacksmith how technology kills jobs. It’s all about creative destruction, baby. Job conservation is not an option. The invisible hand of the market is a merciless optimizer, and inefficiency will not stand. The Luddites fear being left behind in the race for profits.
The fear of being left behind is strong in the minds of investors. AI is insane. When it works, it can change the world. And change means someone can profit.
I promised not to nerd out too much, but here is an example of the insane power of AI. Below is a table showing that even with 10 dimensional vectors (red circle), an AI can figure out by googling documents and reading them, that the word chicken refers to an animal, and not to an error message (green circle).
CNN classification has started exceeding human ability. There are examples in medicine, but healthcare jobs are safe from machines for now. Reinforcement learning / q-learning will kill truck driving and taxi jobs. The list goes on.
Instead of veering into the universal basic income discussion, let’s stay on track and get into the investor’s mind with a specific example.
A Specific Example: Byte Nutrition Science
I have alluded in previous articles to one of our clients, Byte Nutrition Science. I don’t have any equity in it, but my firm does the AI work for them, and they are owned by friends and family. So, with that brief disclaimer out of the way, Byte Nutrition is an AI startup in the food processing space. The device the company makes scans materials to see what is inside them (non-destructive chemical analysis). The company is in this same dilemma we are talking about: when do you go get outside money?
By some miracle, the technology is all the way in the customer validation stage without having raised outside money. The founders are happy to wait for the “price” to go higher and higher while the burn rate is low. However, once the sales and support costs start to kick in, the founders have to think about growth capital. Maybe a loan? Maybe venture funding? How much sales growth represents a hockey stick? How much of the pie should go into the sales channels? Lots of questions and few answers, but this is the good kind of problem to have. Each new compound the system detects requires more work, and automated configuration is still not implemented. As a result, I think Byte Nutrition will end up going for venture funding once the R&D scales up, and the cost of sales kicks in. But who knows. Time will tell.
To sum it all up, just do what feels right to you.
Happy Coding!
-Daniel
daniel@lemay.ai ← Say hi.
Lemay.ai
1(855)LEMAY-AI
Other articles you may enjoy:
