So, I guess I’m the “lifelong learner” type… I have a growing passion for data science, machine learning, statistics, math, physics, computer programming, philosophy, sociology, social psychology, communication, product management, growth hacking, photography, design, user experience (no particular order) and I’ve always wanted to invent stuff that could change the world.

Don’t worry, I know what you’re thinking…  “here goes another Homo universalis“, or how can someone be passionate about ALL those topics and still make a worthy contribution in his life.  I’m definitely NOT the best at any of those topics taken individually, and there is always someone who knows more, I know that, but each of those topics, taken individually, was in my top #3 interests for at least 3 years.

I think that if you are an autodidact (with a high IQ and hyper-focusing ability), who can dedicate enough time to a specific topic of interest, and if you do this for multiple related and non related topics at some point you’ll be able to connect the dots in an unpredictable new way, bringing highly valuable results.

I am not against mastering a single topic of interest (also called specialization, considered to be the best way to go if you want to do something remarkable in a lifetime), but I think that this method is overused, abused and limited while the world needs an expert-generalist mind to solve some of its most important puzzles.

I have a lot of admiration for people like Benjamin Franklin, Buckminster Fuller, Isaac Asimov, Howard Hughes, Blaise Pascal, Leonardo Da Vinci, Brian May, Richard Feynman, C.S.Lewis, Elon Musk, Thomas Edison, Steve Jobs, Walt Disney, Jeff Bezos, and Peter Thiel (no particular order).

Finally, I believe that the most important breakthroughs in the next few years will come from applied artificial intelligence. I have no doubt that we can today influence and advance most of our future through one single thing: artificial intelligence.

I believe that society as we know it will end very soon. My own simulations tell me that today is probably too late already to start fighting the bad future consequences of… today.

My main interests lately are AI and security.

One of my top ongoing objectives is to understand how to create and hack/fool any AI/ML based decision system (and algorithm). I think that, at some point in the future, it’s very likely that these systems will control the world, and people behind them will control the world through them. By that time it may be too late to start learning how these systems work and more importantly: how they can be hacked.


@raulpopa | linkedin

PS. In my day to day life I’m the CEO, cofounder and data scientist at TypingDNA (an AI based typing biometrics authentication startup).

Read More

How to name a successful startup, tips based on statistics/machine learning

After playing around with successful and less successful startup names from YC list of dead, live & exit startups, creating a free app to predict the potential success of a startup just by it’s name, I analyzed the data and tried to get an interpretation that would practically help someone name a new (successful) startup.

So, based on hundreds of startup names and machine learning/artificial intelligence I came up with the following basic “statistic” tips for you (there are others rules for naming good startups but are harder to interpret and pass over) (more…)

Read More

YCombinator 2015 predictions (based on machine learning)

I’ve created a machine learning algorithm that learns from the yclist of dead, active and exited startups. The tool only uses the names of the startups. Some have names that include extensions such as drop.io or about.me and they become part of their name since they use it everywhere, therefore I considered the full (domain) name where it seemed appropriate.

So, based on the yclist training set, and some research based on naming patterns, I created a free startup name checker tool.

It’s free and you can use it to check your app/startup name. Please don’t flood it so I don’t have to limit usage. Thanks.


Read More