AI
Researchers studied AI layoffs. Here’s their warning.
The problem is that companies operate in a competitive environment. Your competitors are using AI to cut costs, so your company is forced to do the same (i.e., trapped). Doing so has positive first-order effects and short-term benefits but negative non-linear outcomes for higher order effects. The solution: self-restraint, but companies likely won’t do this. Some alternatives are reducing the incentive to replace a worker with a machine — augmenting with AI is okay, but to fully replace a worker costs a significant amount. (The closest working analogy is the carbon tax.) Another solution is to subsidize firms that don’t fire workers. If we do nothing, we’ll continue to get massive wealth inequality and political instability.
State of AI Impact in Engineering: Q2 Report (from DX)
The 49-page report provides details about themes that have arisen from 500+ organizations. AI is helping to some degree with speed (e.g., PRs/eng/week) but the perceived rate of delivery is flat. Some aspects of engineering have improved (e.g., documentation, maintainable code, production debugging) but PR size has doubled (i.e., AI-driven inflation rather than disciplined, tested code). Code is easier to read but harder to trust. Spend has increased 28x, but time for innovation has only been 4-6 hours/week.