This post was originally published on LinkedIn.
Today was my last day at Arm.
Back in 2015 I was developing software for an IoT product, and loathed it. Windows only-tools, paid debuggers, terrible documentation, and troubleshooting was terribly difficult due to a complete lack of community. Fortunately, at a hackathon one of my colleagues handed me a Nordic nRF51-DK development kit and showed me the Mbed online compiler. Within an hour we had our first BLE example running, and I realized that this was the way IoT development should be done. The next week I reached out to the Mbed team and the rest is history.
The past three and a half years have been a blast. We’ve built a great developer evangelism organisation within Arm, I’ve had the chance to shape the future of networking technology in the LoRa Alliance and the NB-IoT forum, and worked on some really exciting technology such as the Mbed simulator and JerryScript. I’ve also worked with some of the smartest people I’ve ever met.
But… I feel that it’s time to move on. The Internet of Things is still a fascinating field, but it also feels that innovation in IoT is stalling. I think that’s actually a good thing for the industry. The core technologies have all been created, the new LPWAN networks are being widely rolled out, we finally have a good understanding on security primitives, and now it’s time to implement and scale the thing. It is however, not the place where I do my best work. I’d rather invent the future than implement it.
Over the next few months I’ll plan my next move. I’m currently really excited about the combination of machine learning and IoT devices. uTensor and TensorFlow Lite Micro have shown that it’s possible to run neural networks in less than a 100K of RAM, and I believe this has the possibility to add some proper intelligence to edge devices. If you’re interested in the same field let’s have a chat!