Thoughts on Operations: People, Process, and Technology
It doesn’t matter if you are talking about AI, Cyber, or food service, there are three things you need to foster to operate: People, Process, Technology.
People are your glue, a catalyst, and provide context.
Process is your guide rails, a foundation, and ensures consistency.
Technology is your tool, an enabler, and facilitates accomplishment.
Then comes the balancing act. Like a three-legged stool, you have some room in your measurements for each leg and still be able to call it a stool, but there is a tipping point. If one of these three legs becomes too long or short, you will no longer have a functional stool.
Currently, we are witnessing a race to build better AI. I predict that the software leading the pack will not necessarily be the fastest, most economical, or most accurate, but it will be the most balanced in addressing business operations and customer needs. It will be centered around people, processes, and technology.
People are your glue, a catalyst, and provide context.
Process is your guide rails, a foundation, and ensures consistency.
Technology is your tool, an enabler, and facilitates accomplishment.
Challenges abound in fostering and balancing people, processes, and technology. In SOC (Security Operations Center) environments, I see these challenges all the time. I saw this in the food service industry decades ago when I worked in that field. I see it now with AI (Artificial Intelligence).
The technology in Generative AI is advancing faster than people and processes can hope to keep up with. Leaders need to be aware and ensure their focus is on the balance of people, processes, and technology.
To use my own experience as an example, I have watched Stable Diffusion release 1.5, 2.1, and SDXL in less than a year. My company has used all SD versions, trained models for Stable Diffusion 1.5, and created workflows for SDXL (https://civitai.com/user/EightBuffalo/models). Each new version brings great improvements, but also new challenges. Often, the training of people and adjustments of processes will allow the older version to match the newest base version. This has been my experience in every field I have worked in.
I have learn to not constantly chase the latest and greatest. Work on balance, iterate on improvements, and focus on stability. That is what will enable you to plan properly and take those risky leaps that we all must take from time to time, but now with a lower chance of failure.
The images attached to this article are generated with Stable Diffusion 1.5, using the latest 8buff real SD model, a LORA for generating a 'nobody' face (not based on a real person), and high-res fix on local hardware. It takes about 3 minutes to generate from prompt. Consider what the latest versions and models will achieve once they are adjusted and trained to the level of our year-old 8buff_real 1.5 SD model. The same holds true for LLM (Large Language Models) and other AI models.
These constant improvements are the reason why you must push yourself to change things up, take risks, but realize that you do not have to make changes blindly. Using people, process, and technology as guideposts, and the analogy of the three legged stool to keep your operations balanced can teach you how to change your environment within tolerances. In other words, see how far you can change each of the three legs but still have a functional stool.
The technology in Generative AI is advancing faster than people and processes can hope to keep up with. Leaders need to be aware and ensure their focus is on the balance of people, processes, and technology.
I have learn to not constantly chase the latest and greatest. Work on balance, iterate on improvements, and focus on stability. That is what will enable you to plan properly and take those risky leaps that we all must take from time to time, but now with a lower chance of failure.
The images attached to this article are generated with Stable Diffusion 1.5, using the latest 8buff real SD model, a LORA for generating a 'nobody' face (not based on a real person), and high-res fix on local hardware. It takes about 3 minutes to generate from prompt. Consider what the latest versions and models will achieve once they are adjusted and trained to the level of our year-old 8buff_real 1.5 SD model. The same holds true for LLM (Large Language Models) and other AI models.
These constant improvements are the reason why you must push yourself to change things up, take risks, but realize that you do not have to make changes blindly. Using people, process, and technology as guideposts, and the analogy of the three legged stool to keep your operations balanced can teach you how to change your environment within tolerances. In other words, see how far you can change each of the three legs but still have a functional stool.
Currently, we are witnessing a race to build better AI. I predict that the software leading the pack will not necessarily be the fastest, most economical, or most accurate, but it will be the most balanced in addressing business operations and customer needs. It will be centered around people, processes, and technology.
Comments
Post a Comment