Manus Beats OpenAI, AI’s Hidden Goals, and Copyright as National Security - The AI Argument EP48

Manus might be the biggest leap in agentic AI yet, but is it groundbreaking AI, or just a well-dressed remix of existing tech? It’s making waves, but there’s no secret sauce. No next-gen model. Just some clever engineering. If an independent team can outshine OpenAI and Google with off-the-shelf tools, what does that say about the so-called AI giants?OpenAI wasted no time dropping new developer tools—coincidence, or a torpedo aimed at sinking Manus before it even gets out of beta? Despite how impressive Manus is, Frank and Justin still aren’t ready to let AI book their flights or buy their sneakers. Beyond Manus, this episode takes on some of AI’s more unsettling developments:00:45 Is Manus the agentic breakthrough we've been waiting for?It’s blowing minds, but there’s no secret sauce—just smart engineering. If anyone can build this, how long before it’s obsolete?09:10 Did OpenAI just uncover AI’s sneaky side?They tried to train an AI not to cheat… and made it even better at hiding its true intentions. 14:55 Can Anthropic really detect AI’s hidden goals?A new experiment claims to “read AI’s mind” and spot secret objectives—sounds great, but does it actually work?18:38 Did OpenAI just link copyright to national security?Suddenly, scraping copyrighted material isn’t about profit, it’s about protecting democracy.25:00 Did OpenAI’s new model just write real literature?A short story about AI and grief has some claiming AI has achieved creative brilliance. Others think it reads like a moody teenager’s poetry notebook.29:49 Why does this AI fish sound like Schwarzenegger?Forget Manus—the real AI revolution is a talking fish that gives life advice in Arnie’s voice.