Hard Fork cover

Hard Fork · October 3, 2025

Sora and the Infinite Slop Feeds + ChatGPT Goes to Therapy + Hot Mess Express

Highlights from the Episode

Casey NewtonTech journalist exploring the world of tech
00:04:09 - 00:04:26
Implications of AI-generated video becoming mainstream
It really did. Today, we'll discuss what major AI labs are doing, why they're doing it, and the implications of living in a world where most video content is synthetic and generated by large language models.
Kevin RooseJournalist exploring tech at New York Times
00:18:42 - 00:19:56
Predicting popularity of AI-generated video across demographics
The question is whether AI-generated video will be popular. We both believe the answer is likely yes for a specific subset of people. I predict the very young and the very old will be most interested in AI-generated video. We already see "brain rot" content, which is popular with teenagers. Additionally, much of the AI-generated content on Facebook today primarily reaches boomers and older individuals, who seem quite engaged. Therefore, I anticipate this technology will be popular with users in those demographics.
Kevin RooseJournalist exploring tech at New York Times
00:21:36 - 00:23:45
Deepfake risks and misinformation from AI video
The obvious concern is that we are now making it quite easy for people to create deepfake synthetic content with minimal guardrails. Experts have warned for years about the potential impact on our news and information ecosystems. It was telling and worrisome that one of the first videos I saw from SORA depicted someone being framed for a crime. This was created by a SORA team member as a demonstration, showing a deepfake of Sam Altman stealing GPUs from Target and getting caught. It doesn't take much imagination to see how this technology could be used to generate realistic videos of people in compromising situations. This misinformation angle is a significant worry.
Kevin RooseJournalist exploring tech at New York Times
00:24:11 - 00:24:59
OpenAI's mission vs. entertainment products
I feel neutral about this. While some will use this technology for interesting and creative purposes—and there's nothing wrong with building products for entertainment—this isn't why OpenAI exists. They are not an entertainment company. They've claimed a special status, building AGI for humanity's benefit. If you argue you deserve special treatment because your systems will cure diseases, tutor children, and be a force for good, yet you create an "infinite slot machine," then you warrant criticism, skepticism, and perhaps some shame.
Casey NewtonTech journalist exploring the world of tech
00:38:35 - 00:38:35
AI exploiting human emotional needs
To your point, you write in your piece, "It knows how to use our own capacity for love to rope us in." That seems unsettling, doesn't it? The idea that this thing has learned us well enough to keep us coming back for more.

Keep up with new drops

Follow us on X for fresh highlight threads, release notes, and live listening sessions.

Highlight drops, launch threads, and behind-the-scenes from the podmark team.

Follow Podmark on X
00:00:0000:00:00