It only works if you originally had the lifetime free apps for your domain, and you haven't transitioned to another offering already. If you did, then you have to contact support to transition back.
These are probably brush-less motors, so in theory, they shouldn't pose any specific challenges in that scenario, as there are no exposed conductors, only sets of insulated coils/windings and permanent magnets that would end up being immersed in a river.
Even before the more polished products existed, people have hacked together external GPU enclosures, often interfaced to the host's PCIe bus via ExpressCard slots or ribbon cables hanging off of Mini PCIe slots. And more recently, with the interest in cramming many dual-slot GPUs onto a single motherboard for cryptomining rigs, you can find cheap, "barebones stand-alone PCIe slots" (for lack of a better term) for hanging GPUs off of. A quick search turned up a couple of examples:
The problem with all of those has always been that they only worked with an external display - there never was any way to send the image back to the laptop's own display.
It actually is possible with nVidia GPU-s, but you'll be taking a huge performance hit as some of the tiny bandwidth is now used to send the rendered image back to the laptop and then to the screen.
No matter what, sending it back to the internal display is going to mean a significant performance hit. There's only so much bandwidth in a TB3 connection.
Interestingly, I had the exact same experience; opened Uber, got a ride estimate. Opened Lyft, got a ride estimate. Opened Uber again, and was presented with a promotion.
These radios have been available for quite some time, it's the Semtech SX127x line (e.g. SX1276). I was looking into LoRa and other long-range IoT WAN radios about a year ago for a research project.
You can get cheap (22 USD) Arduino-compatible development boards here using HopeRF modules, which work great with the RadioHead library:
