GenAI pessimism is widespread, often making headlines during market downturns. The 'virality' of these arguments is a compelling reason to re-evaluate long-term risks. There will be many strong forces amplifying negativity when the cycle reverses, potentially leading to severe market reactions. While this is a predictable outcome, an optimist must identify indicators that will determine whether GenAI and subsequent innovations will be a substantial force driving the global economy positively for decades or if they will be overhyped and lacking a 'killer app 'as the saying goes from a long-term viewpoint.
We think categorizing the doom-mongers will help us discern their motives and then pick any relevant lessons they offer. First up: history seekers, the bubbles-and-bust specialists. They view everything through the lens of history. Their intent is normally to display a grasp of historic events and the power of their gab.
These storytellers weaponize narratives, starting with seemingly irrefutable claims: "This time is never different." One viral commentary even declared that technological progress only happens when cheaper alternatives emerge. Such pronouncements, while impactful, crumble upon scrutiny or are based on such a puny number of data points that even their most obsequious statistician associates will cringe. Were cars, computers, or appliances cheaper than their predecessors?
Rigour is not the only thing they willingly ignore in promoting good stories. They make their claim almost unfalsifiable, like dogmas of any faith, giving them the license not to try to follow the technological nuances for what they are.
This analyst wrote decades ago, pattern-matching a famous quote from his history grazing days: “It's better to have bubbles and busts than never to have bubbles at all.” Every bubble-bust-prone system has not progressed consistently over the long term, but it is difficult to think of any society that has grown well over decades without boom-bust cycles. History is not just repetition; it's a story of progress, of the firsts and the lasts. The records are not for when it was the same as before, but when the machines first flew, or someone discovered something radical against all the forecasts of pattern-seekers.
While other Cassandras may offer better risk assessments, history seekers, though entertaining, are perhaps the least useful. They may be proven right, but not for the reasons they cite. Their accuracy, if any, stems from other risks we'll explore later. It is important to remember that whatever path unfolds will not be because of the need to match any real or perceived past pattern; reality does not owe either optimists or pessimists anything to emerge in any set way, particularly with technologies with no precedence. We will need our deductive skills far more than inductive skills to learn about risks.
