I’m already exhausted by the deluge of AI-related news, products, and grifts that have followed Bing’s ChatGPT reveal—except for Kevin Roose’s piece in the Times about its alter ego, Sydney, which is fascinating and should be read. In addition to that exhaustion, my neck aches from whiplash. AI-related technologies are improving at a relentless pace.

I was just generating silly images of cats in kiddie pools with Dall-E 2 a few months ago. Now I can have a conversation with software that can provide me with whatever information I request because it understands what I’m saying. Over the last week, I saw two threads on Twitter about new Tiktok filters—a teenage one and a new beauty one—from @memotv. The near-flawless, real time changes it makes to any face are awesome in the most literal definition of the word.

That awe has morphed into a growing sense of unease. In a world where objective truth has been under attack, are we catching a glimpse at the death of reality?

Extrapolate image generation forward and assume it can create anything with convincing enough accuracy. Do the same for conversational language models that will only get better at understanding and providing information in more human-like ways. Combine that with real time video and audio modification and synthesis. Assume the augmented and mixed reality hardware that’s on the horizon will be as compelling as people much smarter than I expect; then, imagine all that content superimposed on the world around us.

Without disclaimers, how will we know whether who or what we interact with is real? Will knowing matter in such a world?