Hello all,
Just read Andrew’s note in The Batch…
After rapid prototyping “what’s possible”, and quickly answering “how good is this model” with AI coding, consider the robust field of Business Systems Analysis. Decades and large amounts of tools and methodologies used to clearly define requirements, business processes, workflow (systems/people), design, scope… An obvious growth path seems to be standardizing on tools for needs (Swim lane charts…), and categories of definitions AI needs to learn next. Why not leverage this vast area maturity in requirements definition and human expertise to clearly define a problem
Thanks!
there are already researches going on making agentive tool behave as human-like experience. I suppose you must have heard about advanced general intelligence which is focused on aligning the traditional deep learning, memory and adapting capabilities while safety and security being prime concern.
Now the core developers in ai are more focused on robust rag architecture, so individuals who can more strategies on balancing and evolving the shortcomings and feature building, would be future of AI.
Right, understood! That’s the development/solution side.
The highest value, best fit, testable delivered solutions, begin with highly robust (even standardized) problem/requirements/scope definitions. Define a problem completely first, then designing and building the correct solution becomes exponentially easier. Weak definition is a major contributor to project failures!
I’m certain this is a huge opportunity set (along with intelligent orchestration) for AI, but may take some time…