i first principled my way to a pretty fascinating custom agent harness architecture. funny how there's always a conceptual convergence that happens across teams, but when starting from the ground up, you seem to end up with a more complete system.
because i wrote hyperdesign to be timeless, i can reference the paper with an agent anytime i'm working on a really complex design problem and it pays dividends every time
`context` always has been and always will be a design term. the more AI automates knowledge work, the more powerful design becomes.
πΈ me speaking @ucdavis about my studio Direct Message's experiential work for Pepsi at the Super Bowl
@dflieb you're totally right. if we have free productivity (post-scarcity) all problems become creative design problems. it's the evolution from knowledge work to wisdom work. you can read the argument in this design philosophy here. also, quietly refutes AGI... https://t.co/rNJ0bAB1Oo
@karrisaarinen a graphic and motion designer on my team has become one of the best interactive developers i know in the course of a year. imo, this was not possible without the extreme automation of code productivity. is it quantifiable by the existing system? no. is it important? yes.
@asallen because, if design is the "intentional shaping of form to fit a given context" it actually sits above art, but in practice, design faces much more strenuous constraints
@lavrton from implementation -> storytelling. agents are abstracting the mechanics out of sight + mind. human docs need to orient around human intent and provide useful patterns or scenarios for application value. call them technical case studies.
if that's the case, design is the best way to bridge the gap, because as a discipline, it has always been society's interface to technology. no i don't mean graphic design, i mean capital-D Design, the act of intentionally shaping form to fit a given context.
so semantic computing is really just beginning and is probably a 30-year problem. the basic question it's trying to resolve is: how can humans and machines gracefully co-create meaning?
machine learning is fundamentally a social technology, so regardless of it's theoretical potential to solve problems, any product at the application layer is subject to the adoption curve, cultural constraints, and the pace of organizational change.