OPENAI'S AI SAFETY HISTORY: REWRITTEN?
former openai researcher miles brundage called out the company for downplaying how cautious it was with gpt-2. he says openai’s old careful approach actually mirrors its current ai deployment strategy. now, openai frames its work as a steady march toward agi, but brundage warns this could mean brushing off real safety concerns. with competition heating up, the fear is that speed could win over caution—raising big questions about the risks ahead.