Welcome to the AI Tools and Techniques Forum Category
Aristides Alvarez
Posts: 490
This category is for sharing and discussing LLM and AI tools and techniques for programming the Propeller (1 or 2) in any language or for generating documentation.
Parallax will host several AI-specific Zoom presentations and workshops with interested members. These meetings will be scheduled on their own dates and times.
This Category and the meetings were requested on the last PLF (Propeller Live Forum) and you can watch the video here:
https://youtu.be/X6tBaA0ZRdQ?feature=shared
Comments
cringe
One interesting thing is the ability to make apps for mobile devices using AI.
Getting it to create an interface for Wifi control of something would be nice.
Haven't heard that it can do that, but maybe it could.
I actually tested this but I requested the code to be B4A (Basic 4 Android) and it got the WiFi stuff right first time. I'll see if I still have it when I get to my other PC.
However, it doesn't get any easier than using RFO BASIC interpreter, no tool-chain required and one can create a single .apk by "compiling" (packaging really) right on the Android device.
Yeah, it's inevitable that the topic of neural networks and AI on microcontrollers would pop up here.
I came across this video on YouTube describing a neural network application running on an Arduino:
It's a relatively simple task: read the voltage value from a potentiometer, then illuminate different colored LEDs depending upon the normalized input value. That's something standard Spin, Assembly, C, or FORTH code could easily handle, but this neural network application appears to use PWM outputs to drive the LEDs, so it's able to partially illuminate them for potentiometer values that fall between the pre-defined 8 values.
This simple LED task is something the P2 should be able to easily handle. It won't work on the P1 since it lacks an internal A/D converter but can probably work if an external one is used.
The video also mentioned the availability (or lack thereof) of neural network libraries for microcontrollers. Overall, it might be worthwhile to investigate further.
One can do adc on P1 with some external resistors
https://www.rayslogic.com/propeller/programming/adc.htm
Registration for the Propeller Live Forum - AI workshops is available under Propeller 1 & 2 Events tab on this link:
https://parallax.com/category/events/
We have scheduled three events:
Please note that these events are independent of the Propeller Live Forums which were also posted through December.
Yeah, a phone would make a really nice UI, file server, internet connection, power supply, debug console, code downloader, and maybe even light development platform for debugging things in situ.
Since I know more about linux than Android (which technically runs the linux kernel), I would probably start with chroot installation of ubuntu or debian, and then build things from within the environment I know.
Or maybe I would use AndroWish.
AI, on the other hand could help me get started with something simple.
The possibilities are endless.