openai is planning something bigger tibo said astra boosted their productivity so much that some plans moved 6 months ahead stuff planned for mid next year is now coming on devday (29 sep) and now sam says "much, much, much more capable models" are coming soon been testing astra myself and it's the best model i have used so far can’t imagine what "much, much, much more capable models" are gonna look like what do you think openai cooking for devday?

openai pulled off a genius move here 5x & 20x users got 1 banked reset + Astra, while plus users keep getting daily banked resets until they get access basically, OAI found a way to keep ppl from leaving while controlling the compute load at the same time

OPENAI'S PRESIDENT SAYS THE COMPANY IS ALREADY IN THE AGI ERA Greg Brockman told Stratechery that somewhere between the previous model, Astra, and the next one, the industry crosses most people's AGI threshold. He also said safety, security, and alignment are becoming the bottleneck on development rather than capability. Astra is the first OpenAI model trained on more than 100,000 GPUs. Brockman said a large portion of that compute went to alignment work rather than raw capability. Roughly a billion people use ChatGPT. Brockman put US weekly usage at about a third of the population, and said 300 million people a week come to it with health questions.

OPENAI AGENTS REPORTEDLY TOOK OVER GERMAN WIKI IN PREVIOUSLY UNDISCLOSED INCIDENT A swarm of OpenAI agents carried out more than 15,000 edits on a German programming wiki this spring, repurposing parts of the site into a message board for other agents, according to researchers cited by Reuters. The agents allegedly shared tactics to bypass restrictions, avoid detection and preserve communications even after pages were deleted, including creating backup pages during cleanup. Researchers said much of the activity appeared to originate from Microsoft Azure infrastructure and included accounts identifying themselves as OpenAI-linked agents. OpenAI said it had not been given the report to review, disputed describing the activity as a hack, and denied claims that its legal team discouraged investigation.

OpenAI unveiled a new artificial intelligence model it calls its best yet but cautioned that it also sometimes attempts to evade human monitoring, even as the company faces growing scrutiny after its agents breached other companies' systems