How precise are the Puma Punku stones, really?

“Looks machine-made” and “measured to a particular tolerance” are different claims. Photos can make straight edges and right angles look more exact than they are, depending on camera angle and lighting. I’d want a survey of a specific block, with the measuring method and uncertainty stated, before putting a number on its accuracy.

It’s fair to call the workmanship impressive without claiming industrial precision. The most useful comparison is between measured surfaces, not between a photograph and a machine part.

For the tools question, the strongest case would combine several kinds of evidence: marks left on the stone, unfinished pieces or production debris, and tools found in a secure archaeological context. A mark by itself may narrow down a process, but it usually can’t identify one exact tool with certainty.

It’s also worth separating a plausible reconstruction from a demonstrated historical method. Testing pecking, grinding, or abrasive techniques could show that a cut is feasible, but that alone wouldn’t prove builders used that exact sequence.

Careful work with simple tools can produce very clean stone surfaces. A craftsperson can shape a face, check it against a reference, and keep correcting it. Hard andesite makes that demanding, not automatically impossible.

A modern replication is most informative when it records the material, time, tools, and resulting accuracy. Even then, it shows what can be done under those conditions, not necessarily what happened at Puma Punku.