The Self Driving Product — Part 1: Background
A view from the trenches
This series is split into 4 parts:
- Part 1 — Background
- Part 2 — Orientation & Ecosystem
- Part 3 — Autonomous Products
- Part 4 — The Lived Experience
All views expressed are entirely my own.
TL;DR — A reflection of 6 years in Autonomous Vehicles, the ecosystem, moving from engineering to product, and what makes doing product in AVs different.
It’s been almost 6 years since I exchanged clouds for tarmac and switched from flying robots to driving ones. While I will concede an affinity with all things flying and underwater, the immersion into automotive (and detours into mining and other off road applications) has been exciting and educational.
I got into the Autonomous Vehicle (AV) industry at roughly the same point in it’s hype cycle as I did the drone industry and experienced the equivalent shift from Predators to Phantoms. Or, from a tech perspective, from flying in underground garages to the 2+. There are some similarities: military roots, synergy of hardware & software advances (and cost reductions), safety concerns, regulatory frameworks, new business models, etc. However the AV transformation is of a much larger scale with more inertia, with an increased regulatory and societal impact.
From a technical point of view both drones and AVs are fundamentally robotics problems. If anything, the past 6 years have just reinforced my respect for individuals and organisations that can make robots work.
(Mobile) Robotics is the wonderful endeavour of making complex hardware work with complex software under all the noisy disturbances that physics and the real environment generates. A robotics company’s task is then to do this reliably and at scale. It is hard to really appreciate the challenge this is until you have spent many hours in the mud fiddling with some loose connector or XML config file to get the darn thing to move again (or worse, get it to stop). Experienced roboticists will have endless war stories to tell. What adds insult to injury is that even if you get 99% right but only 1% wrong, to a layman the outcome can look like you achieved the opposite.
Organisations proving they can tackle this challenge at scale are noteworthy. Hiring excellent multidisciplinary technical talent is one prerequisite. Another is getting your organisation, processes, and culture right. For else Conway’s Law and the Allen Curve will gang up and be your downfall.
With regards to AVs, the industry has a track record of over-optimistic predictions. Predictions that always seemed somewhat unmoored to those in the trenches. It is not surprising they bred some scepticism and pointed video editing.
Personally I have seen the AV wave build from different vantage points. Transitioning from building self-driving tech, to selling it, and now (crucially) delivering it. It being early 2022 I would say the wave is breaking and there are definitely reasons to be optimistic as focus shifts towards delivery and deployment, vs development.
Hence it’s a good point to take a step back and reflect on where things are and the experience of my own journey from engineering to (currently) product. A decision I made as it connects the close interaction with technical teams with the customer needs and commercial impact. Not to say I’ve fully crossed over to horned beast territory though. Still an engineer at heart, I build and tinker with flying and driving robots in my spare time and continue to add to my own chest of war stories.
Much has been said about both AVs and product, less so about where both crossover. I did find a panel discussion from 2016 on exactly this topic. But unfortunately little salient content or predictions to cross brace with.
I will break the remainder of this post into 3 parts. First, Part 2, provide some orientation w.r.t. what we mean when we say AV. Then reflect on how the world of product and the world of autonomy intersect in Part 3. Then circle back with some personal reflections in Part 4.