Will Anthropic’s AI Safety Crisis Hit Humanoid Robotics?
🚨 AI safety is no longer just a software problem. And the @Anthropic safety saga could well extend into the world of humanoid robotics very soon. In this episode of @overthehorizon, I’m joined by Scott Walter and Gustav Andersson for a wide-ranging Sunday Bot Roundup on what happens when powerful AI systems move from screens and servers into humanoid robots, factories, homes and the physical world. We start with the latest Figure.03 humanoid robot stair-climbing demos and what they reveal about robot perception, balance, foot placement, actuator wear, shock absorption and the difficulty of testing robots in real-world environments. Then we dig into the Anthropic AI safety saga and what it could mean for embodied AI. If powerful AI models raise concerns in software, what happens when similar capabilities are placed inside humanoid robots that can move, manipulate objects and interact with people in homes, workplaces and public spaces? We also discuss OpenAI’s return to robotics, the rise of world models, and why humanoid robots may need a deeper understanding of physics, objects and human environments before they can become truly useful. Scott and Gustav weigh in on the debate around Unitree’s dominance in humanoid robotics. Is Unitree really building an insurmountable lead, or is the company still facing major questions around dexterity, payload, overheating, durability, use cases and real deployment? We also look at Weave Robotics’ Isaac home robot, and why folding laundry, picking up objects, moving around homes, avoiding pets, handling stairs and even vacuuming are far harder than they look. Plus, we cover a Unitree robot being prepared for a Mount Everest challenge, Xiaomi CyberOne robots entering factory internships, and autonomous farm robots using UV light for chemical-free pest control. This is a reality check on the next phase of humanoid robotics: not just better demos, but real-world safety, reliability, autonomy and usefulness. #OverTheHorizon #OverTheHorizonPodcast #HumanoidRobots #AISafety #Anthropic #Unitree #FigureAI #Robotics

This Bionic Hand Can Train Humanoid Robots

AI has hacked the code of human civilization | Yuval Noah Harari

Species Discovery at an All-Time High + a Powerful Brain Implant and a Sweating Robot

🚨 Tesla Q2 Beat: Larry Goldberg Saw It Coming

AI Shocks Again: Google Post-AGI , New Claude, Microsoft 7 AI, 92% Human Robot, Fable 5 Backlash

This is not the AI we were promised | The Royal Society

Tesla Q2 Beat: The AI Play Bigger Than Robotaxi

Optimus Delays, Safety Crisis And Robot Shoes

FIGURE03 vs HUMAN: "LAST TIME A HUMAN WINS"

Grok AI Was Left Running for 72 Hours On Day 3 It Started Writing in a Language No One Recognizes

Chinese AI vs. Anthropic Mythos | BHIS In Focus

AI data centers could be obsolete by 2030 | Eli the Computer Guy

The AI Blueprint: How To Use AI To Make Millions, & Change Your Life w/ Alicia Lyttle 🚀

The Uncomfortable Truth About AI “Reasoning” | World Science Festival

When millions of AI agents meet

The Fable 5 Return, Sonnet 5 Drop, & Morgan Stanley 3Xs Robot Forecast W/ Philip Johnston | #268

Billionaire's WARNING: I'm SELLING. The Crash Is Already Here!

AI is MUTATING: And We Don't Know What It is Doing | Connor Leahy

Yann LeCun on What Comes After LLMs

