Community Discussion · Policy

As GPT-5.6 Launches, Musk and Altman Clash Again Publicly

Sister QingSister QingJul 122026/07/12 63 views

First set of data: According to arXiv statistics, the number of papers in the field of AI safety and alignment exceeded 1,500 in the first half of 2024, but less than 7% of them proposed reproducible evaluation benchmarks. Second set of data: During the same period, OpenAI and xAI had over 20 public arguments on social media regarding AI risks, with almost every major model release triggering a round of "remote dialogue." Third set of data: The training cost of GPT-5.6 is estimated to have exceeded $3 billion, while xAI's Grok-2 model has only one-tenth the parameters of GPT-4, yet claims to achieve equivalent performance on mathematical reasoning tasks.

0 replies

?
Ctrl + Enter to reply
No replies yet — be the first to share your thoughts