When breakthroughs mean ( and via the moniker that ALL design is a remix ) pushing past the current Ai models the question becomes can creative convergence, diverge? Here's how.
When tools are constantly ( and in the current initial rollout phases of Ai, and in it's first forms ) wed to " style " or alternative use cases being based on training data only - and non experimental actions and feedback loops ( say in soccer playing robots ) when the jump to minimal compute chip level architectures that more closely resemble the human brain ( in terms of raw electricity use ) ala neuromorphic architectures is where we see ( in one possible path ) and as detailed in " Multifunctional Devices ... Neuromorphic Applications " some of the 1st fully articulated architectures that might have a chance at non-remix models of " thought " or as one might say, imaginative divergence.
Currently ( and as commonly understood in binary Von Neumann systems ) and as articulated in the latest nVidia chip architecture / training data computation action and reaction system design, we see how sensor data ( what is happening now ) and what could happen ( eg. standard video analysis of soccer games and the way say robots can read this and then be trained directly ) are still a clear call and response based design. It is effectively and more importantly, non-creative. However where truly artistic breakthroughs must ( supposedly ) happening in an experimental environment and where current systems design ( even devoid of Ai ) can do this through machine learning with essentially learning computers it's the " now " sensor based and conversational architectures ( where suggestion and evaluation of what is being considered ) can be engineered to break the ideas of remixing ( preventing unknown and unexpected new directions to be investigated and immediately put into use ) mean a whole new computational architectural when chip level architectures are created. Neuromorphic computing can do this IF it is free of ALL constraints. That in itself is a tall order but well on the way.
More aptly in " The Road To ... Neuromorphic Technologies " the confusion becomes the standard better ( design ) faster ( compute ) and cheaper ( energy use ) tradeoffs again, not focusing on the divergence and convergence of external data that ( through specific behavioral lenses ) can transform what has not been expressed via possibly bread-crumbing, synchronistic, and cut-up ( just to name a few ) mechanisms for unexpected connectivity. In this case, directions for investigations that can produce breakthroughs comes from known logical connections. Where the leaps begin to show their direction the clearest is in the not-knowing of directions. Similar to Johari Window introspection and extraspection oriented wanderings.
Where it is then, the idea of wanderings ( past just experimentation ) that leaves the specific, almost detective level decision tree analysis incomplete means chip level and application layer combinatoric logic systems under current development almost obsolete when neuromorphic computing architectures can be combined with non-local sensor based input and fully separate from training data models of input. In the oil and gas industries as well as in the photonics industry, these are just but a few of the ways vertical industry specific operational functionality ( is being scooped up by training data intake firms ) as a way, possibly, to pre-train in any and all industries that can be soaked up. But, this is still not divergent and synchronisitc in it's design. Neuromoripic systems are expaning that horizon and thus a way for breakthroughs to occur once learning computers can be miniaturized and made infinitely affordable - in a hand held way.
#iGNITIATE #Design #DesignThinking #DesignInnovation #IndustrialDesign #iGNITEconvergence #iGNITEprogram #DesignLeadership #LawrenceLivermoreNationalLabs #NSF #USNavy #EcoleDesPonts #Topiade #LouisVuitton #WorldRetailCongress #REUTPALA #WorldRetailCongress #OM #Fujitsu #Sharing #Swarovski #321-Contact #Bausch&Lomb #M.ONDE #SunStar #USPTO #EUIPO #WIPO