Programming is simple. Judging the market, however…
First, some news!
I am very pleased to announce that Ghost Tractor is now a Sponsoring Community Member of the Writer’s League of Texas! I am hoping to gain valuable information from them regarding what motivates them, what pain points they experience across the entire spectrum of published work production. I want to be able to supply solutions and make the entire process a smoother, more enjoyable experience.
Programming relies on rules, methods and best practices. There’s a lot of them, but they are all “knowable.” What I find difficult in the development of Ghost Tractor’s products is evaluating the state of the software market and the capabilities of consumer hardware. There really aren’t a lot of up-to-date sources describing the hardware and resource capability of the “average” end user system. When you’re designing this new breed of application that relies upon VRAM, AI cores, standard memory, processor cores and thread counts to operate, it becomes a guessing game of how much of a customer segment you’re leaving on the table.
Go too far, and the product you make fails to run appropriately, is limited to an elite few that has the hardware, or even crashes generating a lot of bad customer experience, trouble tickets and refunds. Underestimating the resources leaves the leaves your solution underpowered, light on utility and quickly left behind on the market.
I am currently working on a new product family called Carina, and it involves organization and optimizing multimedia files, collections and professional workflows. In so doing, I am integrating small local models into the applications, and getting it right in terms of footprint and capability is a real balancing act. Enterprise and office PCs rarely have any VRAM at all, and budget/casual systems barely makes a blip, as well, unless they’re using Apple Silicon unified architectures.
My saving grace in all this is that my target market is professional creators: writers, authors, photographers, and others that naturally gravitate to higher end systems* which are more likely to support the features I want to integrate. So I try to keep the applications dialed in, but as minimized as possible. Really it’s all an educated guess, as most marketing endeavors are, and those who know me would affirm, I am not a marketer or a visual designer. But I do what I can.
