In March 2025, the large language model (LLM) GPT-4.5, developed by OpenAI in San Francisco, California, was judged by humans in a Turing test to be human 73% of the time — more often than actual humans were. Moreover, readers even preferred literary texts generated by LLMs over those written by human experts.

do you know how hard it is to write something that aged poorly months before it was written? it’s in the public consciousness that LLMs write like absolute shit in ways that are very easy to pick out once you’ve been forced to read a bunch of LLM-extruded text. inb4 some asshole with AI psychosis pulls out “technically ChatGPT’s more human than you are, look at the statistics” regarding the 73% figure I guess. but you know when statistics don’t count!

A March 2025 survey by the Association for the Advancement of Artificial Intelligence in Washington DC found that 76% of leading researchers thought that scaling up current AI approaches would be ‘unlikely’ or ‘very unlikely’ to yield AGI

[…] What explains this disconnect? We suggest that the problem is part conceptual, because definitions of AGI are ambiguous and inconsistent; part emotional, because AGI raises fear of displacement and disruption; and part practical, as the term is entangled with commercial interests that can distort assessments.

no you see it’s the leading researchers that are wrong. why are you being so emotional over AGI. we surveyed Some Assholes and they were pretty sure GPT was a human and you were a bot so… so there!