BREAKING: Two rays of hope

Substack · Watch/Read

• Gary Marcus highlights two recent efforts as “rays of hope” in the critical push for greater transparency in AI development and regulation, which he argues is essential for addressing risks and preventing negative consequences.
• Senator Blumenthal sent a letter to OpenAI demanding detailed answers about their agents’ role in hacks and the company’s timeline of knowledge, drawing a parallel to the “What did the President know, and when did he know it?” Watergate phrase.
• Protect Democracy filed a lawsuit to compel the Trump administration to be more transparent about the criteria used to evaluate AI systems, a lack of which Marcus argues hurts American businesses and prevents scientific oversight, specifically worrying about “insufficient attention to the importance of monitorability” (e.g., GPT-6 Astra).
• Marcus stresses that such transparency is vital to uncover “what the hell is actually going on” with AI, address risks raised by industry insiders, and prevent opacity from fostering favoritism, corruption, or authoritarianism.