GenAI is (or is not) Bullshit
There is much talk over what the role of Generative AI — software that can produce various types of content, including text, imagery, audio, video and other synthetic data — will be in the future. There are those that believe that GenAI will take over the world (and kill us all). And there are those that believe that GenAI will take over the world (and save us all). But more clarity is needed, and it is only from discomfort that we grow. So, with that maxim in mind, I wanted to make two cases. Case 1: GenAI is bullshit; It doesn’t work; It will not take over the world; We can all calm down. Case 2: GenAI is not bullshit; It works amazing; It will take over the world; We should all get hype.
Case 1: GenAI is Bullshit. Literally. I use the Harry Frankfurt definition of bullshit — “bullshit is speech intended to persuade without regard for truth.” This is what GenAI produces. Bullshit. These models have no relationship to accuracy — thus no relationship with the truth. So GenAI is producing plausible sounding text, plausible sounding audio, plausible looking audio, and none of it is accurate.
Further Reading:
Case 2: GenAI is not bullshit. According to McKinsey, “Generative AI is poised to unleash the next wave of productivity.” They say that “generative AI could add the equivalent of $2.6 trillion to $4.4 trillion annually.” That’s the equivalent to adding the GPD of an entire country with an economy between the size of France and Japan.
Boston Consulting group believes GenAI can be used to unlock value in your business.
And the European parliament seems to agree with these assessments from these large consulting firms. They quote a study from the Analysis Group (funded by Facebook), which finds that AI will have direct and indirect effects on the economy and they envisage modest gains of US$1.49–2.95 trillion over the next decade.
Forbes outlines some ways generative AI will be a boon for the world:
And accuracy is improving.