⚡️Can AI Think Like a Human?

Hey TEDchats Fam!

Happy Friday! Picture walking into a messy room and telling your AI robot to clean it up. Meta’s chief AI scientist believes it's possible, but we’re still a decade away from AI that truly understands the world like we do.

Before we dive in, Gamma's AI can now help you present your ideas in a completely new way! Check it out for free → here

🧑‍💻 DAILY PROMPT

🍕 QUICK BITES

LeCun: World Models Are Key to Human-Level AI—But We’re “Years to Decades” Away

Meta’s chief AI scientist, Yann LeCun, cautions that today’s AI systems, including ChatGPT, are not yet capable of true human-like thinking, planning, or reasoning. He points to “world models” as the next breakthrough, where AI can understand and predict the real world in three dimensions, much like how humans clean a messy room with intuitive actions. However, LeCun warns that developing these systems will be complex and could take “years to decades" to realize their full potential (watch full keynote).

Tesla’s Robotaxi Vision Faces Red Tape Roadblocks

Tesla is aiming to release its “unsupervised” Full Self-Driving (FSD) in 2025 and plans to start producing steering wheel- and pedal-free robotaxis by 2027. However, the ambitious rollout may hit regulatory snags, particularly in California where stricter autonomous vehicle testing and deployment permits are required. Plus, federal safety standards could slow down Tesla’s robotaxi vision, as the company hasn’t yet applied for exemptions needed to mass-produce vehicles without traditional driver controls. In the meantime, you can book a ride with Tesla's “supervised” FSD technology here (read more).

The New York Times vs. Perplexity: Battle Over AI Data Use

The Wall Street Journal reports that The New York Times has sent a cease and desist letter to AI search engine Perplexity, alleging the unauthorized use of its journalism in the AI’s responses. Although this isn't the first time Perplexity has faced criticism, this latest incident is part of a larger conflict between major publishers and AI companies over the use of copyrighted material (see examples here and here). Perplexity, however, argues that it only indexes factual content and that facts themselves cannot be copyrighted. The company has stated its openness to collaborating with publishers, but the standoff reflects growing tension over how AI tools interact with original content.

🧰 AI TOOLBOX 

🤯 MIND SPARK

Humanoid Robot Atlas Gets an AI Boost! Boston Dynamics and Toyota Research Institute (TRI) are teaming up to bring advanced AI intelligence to the Atlas humanoid robot. Leveraging TRI’s large behavior models (LBMs), this partnership seeks to revolutionize autonomous robotics, enabling humanoid robots to perform complex tasks with far less training (see Toyota’s announcement).

Lightmatter’s $400M Power Play: Lightmatter has raised $400 million to revolutionize data centers with its photonic interconnect technology. Valued at $4.4 billion, the company’s optical interface synchronizes hundreds of GPUs, reducing costs and time associated with AI model training, while eliminating infrastructure bottlenecks for hyperscalers like Microsoft and OpenAI (see Lightmatter’s announcement).

AI-Powered Spark in the Electrician Shortage: Amid a growing electrician shortage, Treehouse, fresh off a $16.6 million funding round, is leveraging AI to optimize various types of installations, such as heat pumps and EV chargers. By analyzing home data and anticipating job needs, the company removes the need for pre-installation visits, significantly enhancing workforce efficiency (see Treehouse’s announcement).

Fable Expands Focus on ‘Digital Accessibility’: Fable, a leading startup in accessibility, just raised $25M to expand its tools for cognitive and hearing impairments. With this funding, Fable is also tackling inclusivity in AI by improving datasets to better represent people with disabilities, ensuring future AI models are built with accessibility in mind (see Fable’s announcement).

New Leadership to Fortify OpenAI’s Security: Dane Stuckey, former Chief Information Security Officer (CISO) at Palantir, has taken on the same role at OpenAI. With Stuckey's deep expertise in securing AI technologies, he could strengthen OpenAI's relationships with government agencies as the company expands its focus on secure AI infrastructure and defense-related initiatives (see Stuckey’s announcement).

Anthropic's Billion-Dollar Utopia: Anthropic CEO Dario Amodei envisions a future where AI cures mental illnesses, eradicates poverty, and compresses centuries of medical progress into mere decades. But to unlock this AI-driven utopia, he says, we’ll need “hundreds of billions of dollars” in compute and infrastructure—despite no guarantee that Artificial General Intelligence (AGI) will ever fully materialize (read Amodei’s essay).

An entirely new way to present ideas

Gamma’s AI creates beautiful presentations, websites, and more. No design or coding skills required. Try it free today.

🎨 AI ART

In partnership with