Junior employees at large firms will lose first, through a combination of hiring slowdowns and some firings. As AIs get better and better, AI will climb the corporate pyramid and replace workers one by one. Eventually, in many industries, there will be competitive pressures that force companies to stop hiring and start firing throughout the organization. This will begin in white collar firms, but will eventually impact every sector. We call this pyramid replacement.
If we do nothing, the intelligence curse will work like this: Powerful AI will push automation through existing organizations, starting from the bottom and moving to the top. AI will obsolete even outlier human talent. Social mobility will stop, ending the social dynamism and progress that it drives. Non-human factors of production, like capital, resources, and control over AI, will become overwhelmingly more important than humans. This will usher in incentives for powerful actors around the world that break the modern social contract. This could result in the gradual—or sudden—disempowerment of the vast majority of humanity.
The Soft Margin SVM For data that isn't perfectly separable, we introduce a "slack" variable ξ i ≥0 for each data point. This variable represents how much a point is allowed to violate the margin. The optimization problem becomes: Minimize 2 1 ∣∣w∣∣ 2 +C∑ i ξ i Subject to y i (w T x i +b)≥1−ξ i and ξ i ≥0∀i The hyperparameter C controls the tradeoff between maximizing the margin and minimizing the classification error. A large C heavily penalizes margin violations, approaching the hard-margin case. A small C allows for a wider margin at the cost of more violations.
The objective function is the Within-Cluster Sum of Squares (WCSS): WCSS= j=1 ∑ K x i ∈C j ∑ ∣∣x i −μ j ∣∣ 2 where C j is the set of points in cluster j and μ j is the centroid of that cluster. Let's look at the two steps of the algorithm: Assignment Step: Each data point x i is assigned to the cluster with the nearest centroid μ j . By definition, this step either reduces the WCSS or keeps it the same. It can't increase it because we are explicitly moving the point to the cluster that minimizes its squared distance term ∣∣x i −μ j ∣∣ 2 . Update Step: The centroid μ j of each cluster is recalculated to be the mean of all points assigned to it. The mean of a set of points is the unique point that minimizes the sum of squared Euclidean distances to those points. Therefore, this step also is guaranteed to either reduce the WCSS or keep it the same. Conclusion: Since each step of the algorithm is guaranteed to be non-increasing with respect to the WCSS, and the WCSS has a lower bound of 0, the algorithm must eventually converge to a state where neither step can further reduce the WCSS. This is a local minimum. It's not guaranteed to find the global minimum, which is why running K-Means with different random initializations is common practice.
By default, in the post-labor-replacing-AI world: Money will be able to buy results in the real world better than ever People's labor gives them less leverage than ever before Achieving outlier success through your labor in most or all areas is now impossible There will have been no transformative leveling of capital, either within or between countries This means that those with significant capital when labor-replacing AI started have a permanent advantage. They will wield more power than the rich of today. Upstarts will not defeat them, since capital now trivially converts into superhuman labor in any field.
In a worse case, AI trillionaires have near-unlimited and unchecked power, and there's a permanent aristocracy that was locked in based on how much capital they had at the time of labor-replacing AI. The power disparities between classes might make modern people shiver, much like modern people consider feudal status hierarchies grotesque. But don't worry—much like the feudal underclass mostly accepted their world order due to their culture even without superhumanly persuasive AIs around, the future underclass will too. In the absolute worst case, humanity goes extinct, potentially because of a slow-rolling optimization for AI power over human prosperity over a long period of time. Because that's what the power and money incentives will point towards.