In the summer of 2021, Andy approached me and shared the idea of establishing a start-up together with one of his friends, Brian, which didn't sell itself. At the time, I was deeply troubled by the summer internship programme and stranded in the sand of cryptography; I was in faint pursuit of certainty in my life, both academically and personally, yet Andy didn't even know what the company was all about. All he got was a name, SonicBoom, which clearly came from his favourite Hedgehog in a Japanese video game.
Despite my lack of interest, Andy and Brain went out of their way to register their new company legally around September and joined the start-up programme at CUHK PI Centre. They sent me the invitation again and promised that they had a solid idea about network optimisation with some advanced technology their partner generously shared. All sounded promising, though there were virtually two real people working in this company besides me. Got nothing else that's meaningful to do at the start of a semester, I hopped on, went through a not so notorious process of bureaucracy, of which Brain gracefully took care, and officially became part of the team in November.
The network optimisation thing started with the real pain of using the Internet inside borders where network traffic is heavily censored with limited non-domestic accessibility. Sending and receiving data packages are basically like how we handle parcel deliveries in real life - first by trying to route data through a more direct high throughput line, then sending data back and forth across multiple data centres. The combination of these two is the process of routing, or shall I say, an overly simplified version of which. The network involving jumping across multiple stops is usually called a multi-hop network, and it presents some serious problems, especially when used wirelessly. The transmission throughput drops drastically when the number of hops exceeds 5 or 6, due to the accumulation of packet losses.
Our would-be partner, n-hop technologies, has developed a network coding technology called BATS, to sustain the transmission throughput after even hundreds of hops. BATS is even universal, meaning that a loosy multi-hop network over any physical layer could be optimised, be it short-range wireless (e.g., Bluetooth, Wi-Fi, ZigBee), medium-range wireless (e.g., LTE), long-range wireless (e.g., VSAT for satellite communication), underwater communication (e.g., acoustic, optical), or powerline communication. n-hop technologies has deployed this technology in the Hong Kong Government's pilot smart lamppost system and was looking forward to collaborating with us to introduce this technology for personal use, like gaming accelerator.
All sounded promising until I found out that the collaboration broke up well before the business went serious, and the fault was actually on us for miscommunication. The best part of it was that I was not informed of this critical change until I figured out something was wrong and asked explicitly. Thus, the core of our business was missing basically from the first day.
A kind professor came to the rescue weeks later. Their team was going to participate in a business competition with real investors and real money waiting for their pitching. The schedule was tight, and they needed professional technical evaluation since they came more from an MBA background. Our first meeting was at the beginning of 2022; at the time, they had no solid idea yet, while their advisor set the scope into the metaverse. I happened to have some experience with cryptography, blockchain, etc., so it went smooth first.
Two weeks later, they came back with their plan. They are making a new YouTube but on metaverse.
The aftermath: we cut the connection ever since, and I have no idea if they got their funding; I highly suspect no. SonicBoom is slowly working on a new metaverse product, even though I personally despise the current metaverse.