Why I Don’t Invest in AI and Blockchain Startups

Why I Don’t Invest in AI and Blockchain Startups

Pitch me a product that solves a problem, not a technology stack

Why I Don’t Invest in AI and Blockchain Startups
Image via www.vpnsrus.com. Licensed under cc-by-2.0.

When I began investing in startups, it seemed that every pitch was a company building artificial intelligence for something or another.

AI for software testing. AI for optimizing advertising spend. AI for chemical analysis. AI for monitoring toenail fungus.

Then came blockchain. Suddenly, every startup was pitching blockchain for this or that. Blockchain for distributed software testing. Blockchain for tracking advertising spend. Blockchain for hazardous chemicals tracking. Blockchain for monitoring toenail fungus.

While listening to the pitches, my eyes glazed over. I started checking email or Twitter. There was nothing wrong with the pitches, but I wasn’t interested.

What have I got against AI and blockchain? Nothing. They’re wonderful technologies. World-changing technologies. But they’re technologies, not products.

Let’s start with AI or machine learning (ML), to be more specific. At its cold robotic heart, machine learning is simply a programming technique. Traditional programming uses heuristics to determine the course of action. i.e., if you’ve rated comedy movies highly and if this is a popular comedy movie, then recommend it.

Then machine learning came along and simplified the process. No more figuring out complex heuristics. Instead, throw all the data at the machines and let them figure out what you like by pattern matching your data with millions of others.

Sometimes machine learning is a way to solve a problem that would be impossible with heuristics. Sometimes it’s just lazy programming to avoid the work of figuring out how to solve the problem yourself.

But whether your product uses machine learning or heuristics or lava lamps doesn’t matter. As an investor, I don’t care how you’ve built the product. What I care about is the product itself and whether it solves a real need.

Whether you use machine learning or complex heuristics is no different than whether the front end is coded in angular or react, whether the back end is java or python or even cobol. Each has unique capabilities and limitations that make it a good choice for what you’re building. Or not.

Similarly, once you get past the hype and the cryptocurrencies and the scams, blockchain is nothing but a distributed database. A traditional database requires a central authority like a bank to be responsible for the veracity and integrity of the data. A distributed database provides a way to build and maintain a database without a trusted central authority.

Blockchain is an amazing technology. There are probably many problems that can be solved with a decentralized database that are difficult to solve with a traditional database. (Though I’ve yet to see more than a handful where blockchain solved a real problem that couldn’t be solved simpler and cheaper using a traditional database.)

As with machine learning, blockchain is simply a programming technique. As an investor, I don’t care whether you’re using a distributed database or a centralized database. I care about the problem being solved, whether it works, and whether users need it.

Does this mean I won’t invest in your machine learning startup? Not all. I’ve invested in several machine learning products and even helped build one myself. But machine learning isn’t the pitch. It isn’t the product. It’s a detail of how the product works.

If you’re using machine learning or blockchain, you may have a fabulous product that I’d be interested in. You just have to pitch it right.

“Software testing is a $40B industry, and highly inefficient. Large enterprises run thousands of tests for every release, not because they’re needed, but because there’s no way to know which tests are needed and which aren’t. By finding and running the relevant tests, our customers are cutting testing costs by 95% and releasing software in a fraction of the time.”

That’s an interesting pitch. It doesn’t mention machine learning though that’s how it works.

“We’ve developed the infrastructure to share in-game goods between different video games. The founders led the in-game goods stores at Activision and Electronic Arts. We have thousands of users already and revenues are growing 50% each month.”

Another exciting pitch. Nary a word about blockchain.

There is a place for a discussion of AI/ML or blockchain in the pitch. It’s in a very simple “How it Works” slide that explains the basics of the product operation.

If I were specifically a crypto investor, I’d probably find a pitch for blockchain for video game goods exciting. If I managed an AI fund, I’d be very receptive to a pitch for machine learning for fetal heartbeat monitoring.

But I’m not. Outside of energy sustainability, I’m a sector agnostic investor looking for interesting startups. In other words, I don’t know anything about your technology, and it doesn’t matter so long as you have a killer product that will grow like crazy and get acquired for an insane valuation.

When we get to the due diligence process, I’ll have experts help me drill down into the full technology stack. But that’s later, not your 10-minute pitch that has to get investors like me excited.

So tell me about your amazing product, your incredible team, the huge market ripe for disruption, and the fantastic traction you’ve achieved, and leave the details of how it works for later.

Go to Publisher:

Entrepreneur's Handbook – Medium

Author: DC Palter