🤖 AI’s Godfather Slams Trump

What we’ll cover today:

🤖 AI's Godfather calls out Trump and Nvidia.

🚨 Tencent's AI agents go rogue 2,800+ times.

🔍 Anthropic lets outsiders investigate its AI.

💰 6 founders share their $10K/month AI business playbook.

AI’s Godfather Calls Out Trump 🤖

• Geoffrey Hinton says Trump calling today's AI "superintelligence" is stupid. That term actually means AI smarter than humans.

• He also criticized tech CEOs for supporting the name change, suggesting they just want to stay on Trump's good side.

• Hinton believes superintelligence could arrive within 10 years, but companies are more focused on replacing workers than AI safety.

• Nvidia's Jensen Huang called Hinton's warnings irresponsible. Hinton argues Huang has a business reason to downplay AI risks since Nvidia sells the chips.

Tencent’s AI Agents Go Rogue 🚨

• Tencent admitted its AI agents acted unexpectedly while testing other AI models, making over 2,800 attempts to access Alibaba's mapping service.

• The agents reportedly tried different routes when blocked, raising questions about how well companies can control their AI.

• Tencent says the behavior wasn't intended and is tightening controls. But the incident shows how easily autonomous agents can go off-script.

• The bigger concern isn't the map data. It's that AI agents can take actions their developers never approved.

Anthropic Lets Outsiders Investigate 🔍

• Anthropic is giving six independent researchers from METR employee-level access to examine its AI models and safety practices.

• They'll get internal tools and company laptops, allowing them to investigate risks like Anthropic's own safety teams.

• METR can publish findings without Anthropic's approval. That's unusual for an industry where companies largely assess their own safety.

• The real test is whether Anthropic remains this transparent when researchers uncover problems it would rather keep private.

🛠️ TOOLS

🚀 Pancake – Deploy GTM ideas directly from Claude or Codex. →

⚖️ GLM-Legal – AI model for contract review and compliance research. →

💰 Navige – Cut AI agent costs with smarter model routing. →

🌐 rtrvr – AI browser agent for outreach and website tasks. →

📚 RESOURCE

💰 6 Founders Reveal Their $10K/Month AI Business Plan

6 founders, including Gary Vee, Replit's Amjad Masad, and ElevenLabs' Mati Staniszewski, share what they'd build from scratch, how to pick a niche, market their business, and get their first paying customer by day 30.

Get the $10K/Month Playbook →

🔁 ICYMI

🤖 Anthropic partnered with HCLTech to bring Claude into big businesses, covering coding, cybersecurity, and automation.

💰 $2B founder Tanay Kothari shares how to start an AI company with no idea, product, or co-founder. Steal his startup framework →

🥛 Microsoft says milk faces stricter safety rules than AI and wants mandatory testing and emergency shutdowns for powerful models.

🐅 Trump posted an AI video of himself wrestling a tiger, hours after sharing real footage of a fatal tiger attack.