This is incredibly cool. As someone looking to try out some ideas around applied ML, can I ask what your stack looks like?

Employee here.

React/Django/Postgres on AWS for APIs and websites. Terraform to manage the infrastructure. cortex.dev for serving some ML models, AWS Lambda for serving others. ML models are all PyTorch at the moment, with RDKit doing the chemistry heavy lifting. Data obtained through various means, including some tools from nextmovesoftware.com

There's a bunch more tech involved in supporting the scientists in the COVID moonshot, but that's basically everything ML-related.

Any particular part of the stack you were curious about?

I may have missed something, but I believe the T4s are running in the cloud, so it wouldn't have any effect on the car's power consumption.

To clarify, I mean indirectly the power consumption. I guess this will become an important point of discussion soon-ish, as we get more and more electric cars and ML stuff in them.

My view is, if your car uses electrons, and some of those are for compute, and you just offload the compute to the cloud, you haven't actually reduced the total electricity consumption.

Similarly here, the total "footprint" of the car is increased by adding this feature.

To be honest, there are open source ALPR (https://github.com/openalpr/openalpr) that could probably run lower framerates even on modern i7 processors (http://doc.openalpr.com/getting_started.html#hardware-requir..., https://docs.google.com/spreadsheets/d/1FNwEuJAgZ1LyM2GGd7VR...)

I think the comparison with an electric car is to illustrate that this cloud solution is a waste of energy.

For a project that's intended to be a part of a car, cloud only makes sense for prototyping. 900 ms latency? Uh-oh.

I will forever be grateful to MoviePass for pressuring movie theaters to offer competitors. AMC A-List has been a game changer for me.

I strongly second the use of Atom + Hydrogen.

I struggle with finding food that fits my plans that isn't boring and monotonous. To that end, this seems like a useful endeavor, though I'm not sure if it's all the way there.

I searched through the site and still couldn't find a demo hosted anywhere. The idea of making paged media a more pleasant in-browser experience sounds neat, but the lack of examples is troubling.

We'll need to make these easier to find, but there are a few examples mixed in with the Readme and more at https://paged.design