If you dramatically redefine random to include choices between useful outcomes (ie. the non random boundary conditions to the decision tree), then Occam’s razor works well here.

Amusingly, I suspect that supporters of both the teleological position and the non-teleological position think Occam supports them.

For what it’s worth, the only empirical macroevolution we’ve done, from domestication, to Monte Carlo simulations, to genetic algorithms, to LLM training, is all goal directed.

> Amusingly, I suspect...

Yes...though I'd say that the teleological folks have emotional reasons to judge an ever-so-specific deity (who they already believe in) to be a conveniently minimal hypothesis. Which is a good example of a usually-necessary heuristic - pushing complexity, costs, unknowns, and other unhappy things far enough away (in real or metaphorical space/time) that you don't have to worry about them now.

> FWIW, the only empirical...

All the budget requests for sextillions of organisms, simulated for billions of years, were rejected. So, unfortunately, some large corners had to be cut...

Out of curiosity do you have any kind of argument or supporting evidence for your position or just straw men and snark?