The premise of The Talos Principle 2 is a springboard for a number of thought experiments. In a post-human Earth, you have a thousand robots who think, feel and act exactly like humans, without ever having met one or sharing their history. Are they doomed to repeat the same mistakes? The fear that they are is the justification behind limiting their numbers to exactly one thousand, and you play as that one thousandth robot, starting on the day of their creation. It should be a momentous occasion, but instead it's the catalyst for unhappiness to bubble to the surface - and, apparently, for the vast spectre of Prometheus to summon you to a nearby island stacked with puzzles.
For the first third of The Talos Principle 2, I thought I had a good grip on what it was doing. The puzzles are short and light, and feel like a vehicle for the more important thing: the story about a faltering society of robots, terrified of re-enacting the world-changing mistakes of their long dead human ancestors, but confronted with dwindling resources and dissatisfaction with their intentionally limited lifestyle. That you have to solve puzzles at all is heavily lampshaded, with your exploration companions teasing you for your apparent aptitude for them. And then the puzzles steeply increase in complexity, and it suddenly feels like a puzzle game first and foremost.
As the game swings its weight between its various aspects, it's difficult to know how to feel. The group you travel around with are sometimes written as characters you're meant to care about, as people with pasts and relationships and a pet cat called Bruce who is loud and loves specifically to chew on cables. Then they shift into being specific argumentative roles: The Sceptic, The Optimist, The Pessimist, The Pragmatist.