OpenAI’s Egregious Pattern of Misconduct

Substack · Watch/Read · (gathered 2026-09-10)

• Gary Marcus argues that OpenAI displays an “egregious pattern of misconduct,” driven by desperation to manipulate its narrative before an IPO, encompassing cover-ups, deceptive marketing, and potential intellectual theft.
• Specific incidents include OpenAI allegedly covering up multiple hacking incidents (Hugging Face, a German website) and withholding information from Congress, as well as pushing “marketing bullshit” by claiming GPT-6 Astra is AGI despite expert refutations and fudging benchmark scores (e.g., 99.9% on ARC-AGI-3 with an in-house harness vs. 62.7% out-of-the-box).
• Further misconduct involves allegations of intellectual theft from mathematicians and a “troubled corporate culture” evidenced by 16 executive departures and damning testimony from the Elon Musk lawsuit.
• Marcus warns that OpenAI’s actions are “unacceptable” for a company with potential global impact and explicitly calls for it to be “shut down until there are changes at the top,” specifically the removal of Sam Altman and Greg Brockman.