Mar 14, 2023
Peter Diamandis, years ago, summarized the effects of technological progress in his “Six Ds of Disruption” framework. The final stage in the progression is “democratization.”
Personally, I might not always agree with Diamandis, but in this case, his insight is spot on. We all had heard the curious case of the iPhone — affordable only by upper-middle-class consumers in affluent societies when it launched; today, you are hard-pressed to see anyone without a (typically Android-based) smartphone anywhere in the world. And democratization affects not only the sheer cost of technology but also its accessibility.
Here is a very recent example: On March 1st, OpenAI announced the availability of not only ChatGPT via its API and Whisper. The former receives all the press at the moment, but the latter is no less impressive: Whisper takes voice recordings and transcribes them into text. All it takes is one (!) line of code (“transcription = openai.Audio.transcribe(“whisper-1”, file)”) at the cost of a mere $0.006 per minute, and you can transcribe, with high accuracy, spoken word into editable text.
Within 5 minutes (literally 5 minutes), I wrote a proof-of-concept taking a voice recording from my iPhone, transcribed it with Whisper, and sent it on to ChatGPT to summarize the recording in bullet points. All for a few cents and near instantaneously.
At a far lesser accuracy, this type of technology was considered magical (and partly impossible) just a few years ago. This brings up an important question: What technologies are we seeing today which aren’t quite good enough, and more importantly, too expensive to use at scale, becoming democratized in the next few years? And what are we going to build with them? (via Pascal)
