Updated: Jun 4, 2020
Anyone want to talk about something other than Covid19? Scary, but manageable, times in which we’re living. Makes me miss the simplicity of mutually assured destruction from the Cold War days. We, as a society, will get through this. We always have, and we will continue to do so.
So, back to the topic at hand then…
Why did I make the switch from software to what looks (at first glance, anyway) to be a services company? Simply put, and to borrow heavily from an old ad campaign; Fourteen33 isn’t your father’s services company.
Fourteen33 (F33) was born from Google. The CEO is someone I’ve proudly called my friend for almost 20 years; Adam Massey. He spent the last 13 years at Google. As a result, he knows the Cloud and Machine Learning-Artificial Intelligence (ML-AI) worlds as well as anyone.
Our charter is to help you navigate, operationalize, and (if possible) monetize ML-AI. The promise of ML-AI is so very tempting: find insights of great value in the lakes of data you already have that, in turn, helps you run your business better. Essentially getting something from nothing.
That said, businesses are littered with ML/AI initiatives that have fallen woefully short of expectations. Why? Because this stuff is exceedingly difficult. Difficult to prepare, difficult to create, difficult to operationalize, and difficult to maintain. These are the problem statements that F33 was founded to address.
In short: yes; F33 is a services company. But one who’s deliverables are enabled by proprietary technology (called Whisky AI) – more on this in another article. To my mind, F33 bears striking similarities to another (much larger) pseudo-services company: Palantir. For those of you who don’t know who they are, I’ll give you a moment to do some quick research. But in short; they began as a services company chartered to deliver mission-focused outcomes. In so doing, they created some pretty staggering technology to help themselves deliver said outcomes. The same is true of F33.
We are here to help you address the steep demands of operationalizing ML-AI. They include, but are not limited to, the following: 1. Data is often, put simply, a hot mess. It may require months of very specialized data engineering effort just to get it into the form required to begin experimentation. 2. The core of ML-AI applications are trained data models and associated algorithmic libraries. Creating models and algorithms that actually accomplish something of substantive business value requires deep mathematical and domain expertise, not just database knowledge. Moreover, because things change, maintaining them requires ongoing effort that is no less complex than their original creation. 3. ML-AI is computationally resource intensive, and as a function thereof: expensive. How expensive? Training a single AI model can cost hundreds of thousands of dollars in compute resources. And these are recurring costs. Since models (and data) change over time.
In a short amount of time, we have been able to help a $2.3B network security company, a $24.8B bank, and a $113B technology company get the most out of their ML-AI initiatives. So that’s our business. And why I chose to make the leap, even in these challenging times.
Stay safe and stay sane.
By Chris Mucha
Vice President, Strategic Development
