I can't read the article because of the paywall, but from what I can grasp it's sort of a "generic GPIO controller".
Offhand, I'd be a little worried that it could be replicated for way cheaper with any run-of-the-mill LLM agent and an esp32. I've done some esp32 dev with LLMs and they don't really struggle with it, and I'm suspicious this is going to be way more expensive than "esp32 + random GPIO peripheral from 30 years ago".
The examples in the article feel like something any semi-modern LLM could one-shot, but I can't read it so maybe I'm underestimating the complexity.
I bought it because it’s turnkey and to support the Home Assistant foundation. I want support, not a project I have to maintain myself, that’s what the money buys.
If you have tokens and a bunch of free time, your proposal is an option. I just want it to work with my Home Assistant, and so I’m happy to give them my money. $70 is a steal for the value proposition. “Price is what you pay, value is what you get.”
TLDR turnkey for some, diy for others, based on what you’re optimizing for.
From the piece:
> The Aux-2 doesn’t do anything you couldn’t do already with an ESP32 board and various accessories, or even other ESP-based devices. The Apollo Automations Cast-1, for example, costs $35 and supports the same music-casting and Bluetooth proxy features I’m actually using from the Aux-2 at the moment. You could bodge together something even cheaper with DIY parts.
> That’s not a knock against the Aux-2. That’s the whole point. The ecosystem is open by design, which I find endlessly refreshing. Both Nabu Casa and Apollo fund the Open Home Foundation, which develops Home Assistant and ESPHome (and Sendspin and Music Assistant), and both provide hardware that’s a little easier to use than DIY options. But there’s nothing stopping you from making your own devices. It’s actively encouraged.
I can't read the article because of the paywall, but from what I can grasp it's sort of a "generic GPIO controller".
Offhand, I'd be a little worried that it could be replicated for way cheaper with any run-of-the-mill LLM agent and an esp32. I've done some esp32 dev with LLMs and they don't really struggle with it, and I'm suspicious this is going to be way more expensive than "esp32 + random GPIO peripheral from 30 years ago".
The examples in the article feel like something any semi-modern LLM could one-shot, but I can't read it so maybe I'm underestimating the complexity.
I bought it because it’s turnkey and to support the Home Assistant foundation. I want support, not a project I have to maintain myself, that’s what the money buys.
If you have tokens and a bunch of free time, your proposal is an option. I just want it to work with my Home Assistant, and so I’m happy to give them my money. $70 is a steal for the value proposition. “Price is what you pay, value is what you get.”
TLDR turnkey for some, diy for others, based on what you’re optimizing for.
From the piece:
> The Aux-2 doesn’t do anything you couldn’t do already with an ESP32 board and various accessories, or even other ESP-based devices. The Apollo Automations Cast-1, for example, costs $35 and supports the same music-casting and Bluetooth proxy features I’m actually using from the Aux-2 at the moment. You could bodge together something even cheaper with DIY parts.
> That’s not a knock against the Aux-2. That’s the whole point. The ecosystem is open by design, which I find endlessly refreshing. Both Nabu Casa and Apollo fund the Open Home Foundation, which develops Home Assistant and ESPHome (and Sendspin and Music Assistant), and both provide hardware that’s a little easier to use than DIY options. But there’s nothing stopping you from making your own devices. It’s actively encouraged.