Folding Up Wings, and Full Throttle Ahead

The fantasy of creating something from absolute nothingness, and to build it into existence with my bare hands, is a very appealing proposition. Truth be told, I am addicted to that fantasy. But when I dream of a fantasy, I tend to ignore a lot of intricate details, and that would be the journey of getting there.

One thing that strikes me the most is the conversation I had.

“How hard would it be, for someone to come in and create a Convolutional Neural Network to do exactly what I did?” I asked.

“Hm, probably not hard. Especially it has gotten a lot easier now.” That was the response from someone whom, I believe, is quite qualified to answer that question.

Peggy is a special place. There is no doubt in me of the ability of the founding team. There are a lot of rock solid credentials behind their experience and their achievements.

Certainly, Peggy will eventually outgrow the skill sets that I possess right now. One day, it will even outgrow my ability to comprehend. The core components of the Peggy would one day require a complex set of machine learning models that requires the involvement from an actual expert. I daresay this is not a mere hypothesis, but only an issue of time.

There is always the pathway for me to push myself toward that direction. There are always papers and researches that we can tap into. Although stress is always a catalyst of growth, the path forward will not be an easy one.

But the real question here is, am I doing this out of my own genuine interests and my own curiosity? Or am I doing this because of the necessity of my position?

I am not talking about GPU acceleration, heterogeneous computation and vectorization. I’m talking about things like correlation matrix, eigen decomposition, high dimensional embeddings. The fact that I had to google two out of the three words listed here shows the level of competency I have in this realm.

I think the answer is clear here. I never really had any enjoyment grinding through my statistics and algebra courses. The reality is, I have tried almost everything I can to avoid getting into any of those courses. I am not about to start pretending suddenly, I become suddenly interested, or even an expert in those subjects. In contrast, I have always found pleasure in digging into the very foundation of a computer, squeezing every single bit of a performance out of those poor, already over-subscribed processor cores and asking it for more.

The decision here is between a risky leap of faith, jumping off a cliff, hoping to grow a pair of wings out of my back in time to take me to far and beyond, versus accelerating on a flat and long but firm runway made on top of a solid foundation.

I envy those people, who are flying high up with the pair of wings they grew during their free fall. They defied the death that gravity was pulling them toward. However, there is a long path in front of me. Time will come again, where there’d be a chance for me to take that leap.

Cockroach Labs can’t give me the fantasy of flying up high with my own pair of wings, but it gives me a long runway to accelerate, with guidance from the people who came before me, breaking through the clouds, and heading into the unknowns. Unlike those tech giants that has outgrown their prime, coasting on the low earth orbits, firing up the only few thrusters that are left every few months to avoid crashing down and burning up in the earth atmosphere, Cockroach Labs has just passed Max-Q, and is about to throttle up.

Forget about space travel analogies. I suppose what I really want to say is, there’s always a boundary that needs to be pushed. But pushing such boundary requires a source of motivation that will last me for a very long time, and that motivation must come from within.