Sometimes people contact me through my blog and it jolts me to remember that I don’t update it enough. I should really get into this more and set a card on my to-do list for once a month to get back into the swing of it. In terms of what’s going on, I wrote another book which is on async Rust which can be seen in the link below:

https://www.oreilly.com/library/view/async-rust/9781098149086

The book is essentially how to get a computer program written in Rust to handle multiple tasks in one thread that do not need constant computation. We do this all the time in our daily lives. For instance, it we turn the kettle on, we don’t just stand and stare at the kettle, we instead keep a mental note to keep track of the kettle’s boiling status and do other tasks such as getting a mug or putting bread in the toaster. In async programming, you code the program in such a way that the program keeps tabs on background tasks whilst executing other tasks.

This isn’t all that has happened in the last 8 months. I’m currently working on the third edition of my second book, and a database company approached me to help build out the AI engine in Rust and embed that AI engine into the database, below is a video of me giving a talk on this project:

The database can run on a server, over a distributed cluster, embedded in a device, or in the browser. This means that we can pretty much deploy this database anywhere, and seeing as I’m embedding AI into the database, it makes sense that we also try and embed Rust backed AI into surgical robotics and this is what I am currently focusing on.

Kings is the London center of bioengineering and is the largest bioengineering center in the world outside of the USA (it’s the 4th largest if you include the USA).

There is some serious talent behind these walls and we are trying to get the robot to pass the video feed into the AI engine in an async way so the robot can work out where it is and what stage of the operation is currently happening.

We now have to get the AI inference interfacing with GPUs as you can probably guess it’s very computationally expensive, so I’ve traded the laptop for this Linux-based workstation that runs Nvidia GPUs:

Sadly I can’t just take this anywhere but my code now has to interact with the GPU and CPU in an async way if we are to get the robot identifying multiple things in the video feed in real time.