I was studying more about graphql, having used it for years. There has been many phases of critiques towards the technology/paradigm; however, there is something interesting about it's role in a world of agentic AI.
If queries can be dynamically written on the fly, we wouldn't need automation/agent to interact with the UI (as is the case with chatGPT operator or anthropic computer use). Instead, an intelligence layer for the frontend could dynamically query based on a graphQL schema and render with some basic html elements.
Philosophy: building on current platforms limits solution sets and only gives you primitives that they thought of. I built 15+ variants of software versions tangential to flowy. It always limits experience. This doesn't mean everything should be it's own hardware. I just believe that human to human commz should be put higher on the hierarchy than, say your calculator.
Practical answer: we want to do cool things like "hover over keypad and say a name" and then you can talk to them. Only possible when we build the whole stack.
Seems like a huge amount of capital to invest in something most people are already perfectly fine without. If I were in this company's shoes, I would build a SaaS and sell the hardware as an add-on. The venn diagram of people who need this type of VoIP/chat service but don't have an existing SaaS solution is small, even smaller if you also limit it only to people who need hardware and are willing to pay for it.
At least if you did a dual SaaS/hardware offering you'd cast a bigger net.
