Search for a command to run...

Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
In this episode of Hard Fork, Kevin Roose and Casey Newton visit OpenAI's third annual Dev Day in San Francisco, where they explore the company's ambitious plans to transform ChatGPT from a simple chatbot into the "new front door to the Internet." (02:27) The hosts discuss OpenAI's major announcements including 800 million weekly users, new apps integration with services like Spotify and Zillow, and the rollout of Sora Two video generation technology. They also analyze OpenAI's trillion-dollar infrastructure deals with AMD and NVIDIA, examining the massive financial commitments that could reshape the entire tech economy. (24:00) The episode concludes with an entertaining conversation with Business Insider's Katie Natopoulos about her viral AI-generated videos and the cultural implications of accessible video generation technology. • Main themes include OpenAI's platform strategy, massive infrastructure investments, AI-generated content creation, and the potential economic and social impacts of these technological developments.
Kevin Roose is the technology columnist at The New York Times and co-host of Hard Fork. He covers artificial intelligence, social media, and the intersection of technology and society with a focus on how emerging technologies impact business and culture.
Casey Newton is the founder of Platformer, a newsletter covering social media and internet platforms, and co-host of Hard Fork. He previously worked as a senior editor at The Verge and has established himself as a leading voice in tech journalism covering platform dynamics and content moderation.
Katie Natopoulos is a senior correspondent for Business Insider covering technology and internet culture. She's known for her innovative social media experiments and viral pranks, including convincing mainstream publications she was the editor-in-chief of Threads and creating elaborate fake engagement bait content that highlights platform dynamics.
OpenAI revealed its strategy to transform ChatGPT into a comprehensive platform that integrates with major services like Spotify, Zillow, Expedia, and Target. (07:01) This represents a fundamental shift from ChatGPT as a conversational AI to becoming what executives call "the new homepage for the web." Users can now accomplish complex tasks like creating Spotify playlists or browsing real estate listings directly within ChatGPT, with the AI leveraging conversation history for personalized recommendations. This mirrors Facebook's failed platform strategy from the early 2010s, but with potentially higher stakes given the intimate nature of ChatGPT conversations. The success of this approach will determine whether OpenAI can capture a significant portion of internet commerce and activity.
The integration of third-party apps with ChatGPT creates unprecedented privacy vulnerabilities that dwarf previous social media data breaches. (12:27) Unlike Facebook data that reveals social connections and preferences, ChatGPT conversations often include therapy sessions, intimate personal details, and deeply private thoughts. When users connect apps like Zillow or Target, they're potentially sharing years of personal AI conversations with these companies. OpenAI promises to share only "minimum necessary information," but users must explicitly consent to sharing conversation data and memories with third-party developers. The company's own warnings acknowledge that "attackers may attempt to use ChatGPT to access your data in the app," highlighting the significant trust users must place in both OpenAI and its partner developers.
OpenAI's trillion-dollar infrastructure deals with companies like AMD and NVIDIA represent an unprecedented financial commitment that could impact the entire global economy. (24:35) The company has committed to purchasing six gigawatts worth of AMD chips and ten gigawatts from NVIDIA - equivalent to the power output of 20 nuclear reactors. A Harvard economist estimated that investments in data centers and information processing software accounted for 92% of U.S. GDP growth in the first half of the year. Wall Street analysts warn that Sam Altman now has "the power to crash the global economy for a decade or take us all to the promised land." This creates a "too big to fail" scenario where the success or failure of AI companies could have systemic economic effects, making the entire economy dependent on AI's continued growth and adoption.
Sora Two has reached number one in the U.S. App Store while simultaneously triggering backlash from copyright holders and concerns about misuse. (44:45) The technology enables users to create sophisticated video content featuring anyone's likeness, leading to both creative expression and potential harassment. Katie Natopoulos's viral videos demonstrate the platform's entertainment value, but also reveal concerning patterns - the user base skews heavily male, with women largely avoiding the platform due to obvious risks of deepfake abuse. The technology still has limitations including ten-second video limits, voice synthesis issues, and content restrictions, but these technical barriers will likely diminish over time. This mirrors ChatGPT's adoption curve, where initial novelty gave way to genuine utility as the technology improved.
YouTube creators like MrBeast and Casey Neistat are grappling with how AI video generation could threaten their livelihoods while simultaneously offering new creative tools. (45:37) The technology creates a fundamental tension - it poses an economic threat to content creators who rely on unique video production capabilities, yet offers innovative ways to enhance creativity. As AI-generated content becomes more sophisticated, traditional creators may find their competitive advantage eroded by automated systems that can produce similar content at scale. However, the current technology still requires human creativity for compelling prompts and concepts, suggesting a future where successful creators will be those who can effectively collaborate with AI tools rather than compete against them. The challenge will be maintaining authentic audience connections in an environment increasingly filled with AI-generated content.