innovation is a gpu

when people think about innovation, we often think of it as a singular work done by a person or a group of people at a particular point in time. we hear things like, "alexander graham bell invented the telephone in 1876." while that isn't technically wrong, it doesn't tell the full story. it boxes us into believing that the invention of the telephone can be attributed almost entirely to one person.

i think there's a better way to visualize innovation.

i've been learning more about gpus recently, and i think they provide a very good analogy for how innovation actually works.

at a fundamental level, a gpu is a type of processor that allows us to perform a large number of computations in parallel. instead of having one process execute every task sequentially like the cpu, we can divide the work across thousands of threads and have them work simultaneously on a gpu.

most people, and even i am guilty of this, tend to think of innovation as a cpu.

whenever we see an invention, what immediately comes to mind is one person starting with an idea, developing it, and eventually producing the final invention. our default assumption is that it's one process that's executed from beginning to end.

however, innovation is much closer in characteristics to a gpu.

imagine that every person working on any problem, regardless of "size" or "importance," is a thread. however, there aren't just a few threads. there are thousands, perhaps millions, spread across different countries, over different generations, disciplines, and cultures. each of those threads is working on its own problem, often without knowing what the other threads are doing.

so one thread leads to a discovery in physics. maybe another develops a mathematical technique. there's another thread that figures out how to manipulate and make use of electricity. some other thread leads to the development of a more durable material. the accomplishment of another thread might be making a previous invention better. someone else fails at an experiment and unintentionally gives another person information that becomes useful decades later.

each of these is a separate thread, but eventually, their results accumulate.

take for example, the telephone. alexander graham bell is the person history recognizes for inventing it, but bell did not start from scratch. his work depended on an enormous body of knowledge that had been accumulated before he was even born. there were people studying sound, electricity, communication, materials, mathematics, and countless other things that eventually made the telephone possible.

the same thing is true of the airplane. everyone remembers and perhaps rightly honors the wright brothers, but their achievement depended on generations of mathematical discoveries and tons of physics laws that had been discovered. they wouldn't have succeeded if a bulk load of engineering knowledge hadn't been perfected, multiple experiments hadn't been done on some things, and the failed attempts that came before them hadn't happened.

when we look at the final invention, we see the output, but we don't see the thousands of threads that produced it.

and i think that's where our perception and understanding of innovation goes wrong.

we see the result and attribute it to the person standing at the end of the process. this means we don't see the enormous network of ideas, discoveries, failures, and incremental improvements over hundreds of years that converged to produce that result.

so innovation isn't really a sequential process. it is thousands of parallel processes running across time.

however, there is an interesting twist.

the threads of the gpu are sequential. so each of our "threads" is sequential.

a scientist still has to conduct one experiment after another, and a mathematician still has to prove one theorem at a time. an engineer still has to build one prototype, test it, learn from it, and build another.

what does this mean?

it tells us that innovation is simultaneously parallel and sequential. it's parallel at the level of civilization, but sequential at the individual level.

and, the process is recursive!!

the result of one thread becomes the input for another thread. that thread runs on its own and then produces another result, which becomes an input for another. the output, results, failures, and knowledge of one generation becomes the starting point for the next.

this is why i believe innovation is an infinite recursive loop.

it has no obvious beginning, and it definitely isn't ending anytime soon.

so when we say alexander graham bell invented the telephone, we are just choosing a point in an enormous chain and giving that point a name. but if we asked: what discoveries made bell's work possible? then, we ask again: what made those discoveries possible? and what made those discoveries possible?

the further back you go, the more threads you find. there's no end to it.

and the same thing happens going forward.

the telephone contributed to future communication technologies. those technologies contributed to computers as we know them today. the internet was built on the knowledge accumulated from building computers. it didn't stop there. the internet paid its due forward and contributed to smartphones. smartphones contributed to the systems we use today. and today's systems will also become inputs for inventions that haven't happened yet.

perhaps this explains why innovation is so difficult to predict. because at no point in time can we truly say which thread matters.

just like how early computers were built on charles babbage and ada lovelace's ideas and papers a century after, what seems insignificant today might become a critical component of an invention fifty years from now. the person making that discovery may have absolutely no idea what they are contributing to. they, and the society at that time, might even think of it as a failure.

so the next time you hear that someone "invented" something, i think it's worth thinking beyond the name attached to the invention.

don't think of innovation as a single process running from beginning to end.

think of it as thousands upon thousands of processes running in parallel, each one moving sequentially, constantly producing outputs that become inputs for other processes.

innovation is parallel.

innovation is sequential.

innovation is recursive.

and perhaps most importantly, innovation is cumulative.

what we call an invention is often just the moment when thousands of previously separate threads finally converge into something we can see.

unconventional thinker