How to convince Venture Capitalists you’re an expert in Artificial Intelligence

If you like this article, check out another by Robbie:

Not sure if you’ve heard, but artificial intelligence is kind of a big thing. There is an endless stream of prognosticators () droning on about how AI and machine learning is going to revolutionize every industry and all the jobs are going away.

If you are an entrepreneur trying to raise money, you’ll benefit from riding this wave while it lasts. My friends in venture capital tell me that at least 75% of pitches they see nowadays include something about AI. More than 75% of the startups I talk to are worried that they aren’t using AI enough.

That means there is a gap. You need to be able to sound like you know what you are talking about without knowing much about AI. No worries, this article is here to help. I’ll present several things you can say during a pitch meeting that will convince a VC that you put the AI in fundrAIse.

Use this “AI Expert” sentence generator

Like many areas of technology, AI is filled with buzzwords and acronyms. New AI acronyms are created on a regular basis as new systems, models, and datasets are invented.

Next, I’ll present a parametric n-gram language model for creating a word distribution that sounds like it was written by an AI expert. See, I can even make mad libs sound complex.

Here is the first one:

“We’ve built a state-of-the-art [CNN, RNN, kNN, GAN, seq-2-seq, LSTM] model that took a week to run on a [GPU, TPU, FPGA] cluster that we got access to from a friend at [Stanford, CMU, MIT, Toronto, Berkeley].”

And another:

“We created a system called [Popular Gen-X first name] that is based on the work of [Socher, Karpathy, Lecun, Hinton, Thrun, Koller, Goodfellow, Bengio] and uses a [50,100,1000] dimension proprietary dataset.”

You could combine the two sentences together into an uber expert sounding passage.

Sporadically reference a paper on Arxiv as if everyone has read it

is the repository of scientific papers that most computer science researchers publish to these days. Dozens of papers are published over the course of a week and yet people like to refer to specific papers as if everyone has read them. Find a paper on arxiv and talk about it like everyone knows it as well as Jack and the Beanstalk. You don’t even need to refer to it by name. You can just say “Yoshi’s paper on CNNs” or the “SELU paper” for that added touch of misdirection.

Propose your own “Net”

Stemming from the term “neural networks” or “neural nets” it’s common in the AI community to name your new neural network approach with the suffix “net”. For example, we use “ConvNet” instead of saying “Convolutional Neural Network.”

I’d like to propose RobbieNet: the state-of-the-art self-aggrandizing machine readable language model for sarcastic blog posts.

Trivialize existing “AI” companies

Many companies have already jumped on the AI bandwagon and have gotten trounced by AI experts for overselling what they do. IBM Watson is the current poster-child for . Chatbots are another area that has been supremely overhyped relative to the value companies have gotten out of them. By talking down about their lack of AI, you sound like you know the difference between real AI and fake AI — with the implication that you are doing real AI. I can hear Trump now: “It’s all fake AI!”

Disagree with Elon Musk

Elon Musk has been on a rampage about the . He’s really worried about the future of humanity. The is that he’s overstating the risks, especially in the near-term (10–20 years). By saying you disagree with Musk, you sound like you are educated on the topic of killer robots and thus must know a thing or two about the space.

Agree with Elon Musk

There are more people that disagree with Elon Musk’s doomsday concerns than agree with him. That presents a great opportunity for you to agree with him and be in the minority of a complex topic that one of the most successful entrepreneurs of our time sides with you on. Talk about how you can see some of the “latest advances” leading to and the possibility of robots taking over. Show a social conscious by talking about how we need to get in front of these issues now before it is too late!

Say you like PyTorch because Tensorflow is too slow

The Google-based Tensorflow platform has gotten a reputation of . The Facebook-based PyTorch platform, on the other hand, has a reputation for being fast. Most VCs have probably heard of Tensorflow, but not as much about PyTorch. Talking about alternatives to Tensorflow will make you sound smart.

Just don’t do this if anyone in the room is from Google!

Mention that you signed up for the Google TPU Alpha

Google sent a ripple through the chip world when they . TPUs are optimized for running ML models. Mention that you are excited about what they are doing and you signed up for the alpha.

On your way out, make a comment about automating Venture Capitalists

When you are wrapping up your pitch meeting, you can throw out one more zinger that will leave no doubt in the mind of the VC that you are an AI expert. From my days at , I learned that people are fascinated with the idea of technology that can automate their job. At Automated Insights, it was journalists that were concerned, which lead to them writing lots of articles about us.

After your pitch meeting and if the mood is right, you can throw out a line about how you look forward to wrapping up your round because you want to start building a deep learning model that will automate venture capitalists. Trust me, they’ll love it!

Machine Learning in Practice

Practical insights for executives, managers, and project managers eager to deploy machine learning inside their company.

Robbie Allen

Written by

CEO @InfiniaML, Exec Chairman @Ainsights, Lecturer at @kenanflagler, Ph.D. Student @UNCCS, Writing a book:

Machine Learning in Practice

Practical insights for executives, managers, and project managers eager to deploy machine learning inside their company.