Powerful New Tech, Same Old Pattern

Using the 3DM6 Approach to Spot Opportunities to Use AI

Making Life Easier

One of the most persistent human motivations is the search for ways to make life easier.
Throughout history, we’ve invented new tools or found new ways to adapt existing technologies to reduce effort and improve outcomes. AI is the most recent example, and it seems destined to have the biggest impact yet.

The Role of Decisions

One useful way to recognize opportunities to make life easier is to look at the decisions involved: How to dig, fish, hunt, write, light a dark room, or calculate.
A shovel for digging. A net for fishing. A spear for hunting. A printing press for books. Electricity for light. Computers to calculate.
All of these were once innovative tools or technologies that made tasks easier, faster, and more efficient with better outcomes.
AI is no different. One way to use AI is to look for ways to improve decisions.
One of the decisions AI is being applied to right now is “what word should come next.”
It’s a tiny, granular decision with surprisingly powerful applications: summarizing, comparing, researching, writing, generating ideas, or promoting products and services.

A 3DM6 Perspective

Recognizing that level of granular decision reflects the thinking built into the 3DM6 Approach.
Setting a goal of communicating an idea you think would be helpful could lead to the decision to write a book, which leads to decisions about what audience it is for and how to structure it. Eventually, you get down to a simple but essential choice: what word to write next.
From there, you could look at the strength of associations between words. You might even segment those associations by publication type to help differentiate phrasing for different audiences. Then you'd start thinking about how to scale that capability and build it into a solution that’s genuinely useful.

The Four Domains of the 3DM6 Approach

These four steps mirror the activities in the 3DM6 Approach:
• Objective Analysis means getting clear about what success looks like.
• Decision Analysis helps identify the choices and their characteristics.
• Essential Analysis uncovers patterns that inform those decisions.
• Solution Design is where it all comes together into something real and useful.

I’m not claiming the teams building Large Language Models followed the 3DM6 Approach with intention.
I am saying that the same best practices that helped humans develop spears, nets, and calculators can be found in the development of these LLMs. And the application of those same practices can help guide how we use and develop AI going forward.

What Actually Happened

Interestingly, the idea of predicting the next word using a probabilistic approach had been around for decades as part of research into Natural Language Processing (NLP). From the beginning, NLP had broad goals: to enable computers to read, write, translate, and carry on meaningful dialogue with humans.
The technical breakthrough that enabled large-scale execution came in 2017 with the introduction of the transformer model. After that, the challenge became how to apply it effectively and scale it reliably.

Although the original goals of NLP were diverse and evolving, the development of large language models offers a practical example of how the 3DM6 Approach can help structure progress from objectives to solutions.

• Objective Analysis – The original goals included translation, whether between languages or from language to code and back. Viewing each word as a target event provided the clarity needed to connect objectives to decisions, analysis, and solutions as well as to track progress.

• Decision Analysis – Key decisions included determining what a given word should translate to or what word might come next. Recognizing words as unique tokens or token combinations enabled translation, even when the underlying meaning was not fully understood.

• Essential Analysis – This required identifying patterns and associations in massive amounts of data in order to predict the next most likely word based on context.

• Solution Design – The technical barriers to performing that level of analysis at scale must have once seemed insurmountable. But with the transformer architecture, new training techniques, and massive compute resources, those foundational ideas were transformed into practical tools like ChatGPT.

Each advancement was driven by practical constraints and trade-offs, leading to new alternatives and, ultimately, more scalable solutions.
Whether or not this exact process was followed, the development of LLMs reflects the practical, decision-centered best practices captured in the 3DM6 Approach. These are principles that can help guide how we deploy AI and other technologies going forward.

For a deeper dive into the 3DM6 Approach, including unlisted posts and more detailed applications, visit the [Table of Contents].