- Roko's Basilisk
- Posts
- The Infinite Slop Machine
The Infinite Slop Machine
Plus: Sora gets stricter, AgentKit speeds dev work, and xAI faces regulators.
Here’s what’s on our plate today:
🧪 OpenAI’s pivot to AI-powered social media reveals Meta’s darkest chapters.
📰 AgentKit launches, Sora tightens identity controls, and xAI faces backlash.
🧠 Brain Snack: How to test AI filters without retraining your model.
🗳️ Poll: Should OpenAI be building social media apps at all?
Let’s dive in. No floaties needed…

Stay at the forefront with daily Memorandum tech insights.
Memorandum distills the day’s most pressing tech stories into one concise, easy-to-digest bulletin, empowering you to make swift, informed decisions in a rapidly shifting landscape.
Whether it’s AI breakthroughs, new startup funding, or broader market disruptions, Memorandum gathers the crucial details you need. Stay current, save time, and enjoy expert insights delivered straight to your inbox.
Streamline your daily routine with the knowledge that helps you maintain a competitive edge.
*This is sponsored content

The Laboratory
OpenAI’s infinite AI TikTok slop machine
The modern world is a testament to humanity’s ability to pursue goals that promise collective benefit, even when the means to achieve them carry risks or unintended consequences. A look at the history of human evolution shows how humanity has often used the ends to justify the means employed to achieve them.
In 2025, one of the clearest examples of the pursuit of collective prosperity has been tied to the idea that advanced artificial intelligence will assist in tackling global threats like climate change. AI companies have also positioned AI as the means to lowering costs, accelerating drug discovery, and automating mundane tasks, freeing people to focus on more meaningful pursuits.
However, to develop AI systems capable of handling these issues, the means rest on private corporations’ ability to build massive data center infrastructure supported by substantial capital, top-tier talent, and favorable regulatory frameworks.
For this, companies like OpenAI require broad public engagement and investor confidence, something the company hopes can be achieved through a bold step into the social media space. OpenAI’s latest video generation model, the Sora 2, combines AI-powered video generation with a social feed, signaling a new chapter where AI intersects directly with everyday digital experiences.
Unlike social media apps of the past, where user-generated content dominated feeds, OpenAI’s Sora 2, while functioning with a TikTok-style interface, relies on AI-generated content. The app lets users generate short-form AI videos, remix videos created by others, and post them to a shared feed.
Currently, the app is restricted to iOS devices and is invite-based, which means users need a code to access it. However, despite the restrictions, it was quick to claim the top spot in Apple’s App Store, ahead of Google’s Gemini and OpenAI’s ChatGPT.
The Sora app is powered by OpenAI’s latest video and audio generation model called Sora 2. According to OpenAI, the model is capable of creating scenes and sounds with a high degree of realism.
This is OpenAI’s first foray into the social media space. However, it may not be the last. According to a report from The Verge, OpenAI is also working on its own X-like social media network. While this would put OpenAI in direct competition with companies like Meta and X, it is one of the easiest methods to bypass existing platforms to gather massive amounts of data that can help train future models and ensure OpenAI’s lead. Social interactions are a treasure trove of real-world usage data that can also help refine models and improve safety measures.
The launch of Sora, however, underscores a recurring dilemma for OpenAI. While it has become the fastest-growing consumer tech company globally, it also operates as a cutting-edge AI research lab with a high-minded nonprofit mission. Some former employees suggest that the consumer-facing products can support this mission by funding research and expanding access to AI technology.
So, while OpenAI continues to chase the ends of AGI, a social media app like Sora becomes the means. But, beyond assisting OpenAI in its endeavours, the app can have far-reaching consequences on its users and the larger society.
Even without a dedicated social media app, generative AI’s ability to create hyper-realistic videos is seen as a potential for misuse. To address these concerns, OpenAI shared that it has taken steps, including giving users explicit control over how their likeness is used on the platform. Other safeguards include plans to address concerns around doomscrolling, addiction, and isolation.
According to OpenAI’s blodpost the platform is designed to promote creativity and well-being rather than passive consumption. To achieve this, by default, the app emphasizes content from people users already follow and highlights videos likely to inspire them to create. OpenAI stresses that it is not optimizing for screen time but for creativity. Sora 2 also centers on social interaction through ‘cameos’, which let friends appear in each other’s AI-generated videos.
To address addiction for underage users, OpenAI says the app limits daily video generations, stricter cameo permissions, and enhanced moderation to prevent bullying. And, unlike traditional social apps that rely on engagement-based monetization, Sora 2’s only potential payment feature will be for generating additional videos during high demand. The company positions the app as a healthier, creativity-first alternative in the digital media landscape.
However, while OpenAI’s pitch for the app sounds like it knows what it is doing, content moderation can be a tricky challenge. And many fear the company may not understand the extent of the problem.
Within days of the app's launch, the company’s own researchers have begun sharing their frustrations on X. Many appear conflicted about how the launch aligns with OpenAI’s nonprofit goal of developing advanced AI for the benefit of humanity.
Rated high for potential misuse
OpenAI pretraining researcher John Hallman, in a post on X, shared that he acknowledged initial concerns about the launch of Sora 2 but emphasized that the team had worked diligently to design a positive user experience. They reaffirmed the commitment to ensuring that AI serves humanity’s interests without causing harm.
Another OpenAI researcher and Harvard professor, Boaz Barak, responded, saying they feel both apprehensive and enthusiastic. While acknowledging Sora 2’s technical achievements, they cautioned that it’s too early to claim success in avoiding the issues seen with other social media platforms and deepfake technologies.
While OpenAI says it wants to avoid the pitfalls of social media platforms, according to reports, users are already noticing some engagement-optimizing techniques in the app, such as the dynamic emojis that appear every time you like a video.
According to a report from The Guardian, Sora produced multiple videos depicting violent and distressing scenarios. Some showed bomb and mass-shooting scares, with panicked crowds running through college campuses and busy locations like New York’s Grand Central Station. Other clips portrayed war zones in Gaza and Myanmar, featuring AI-generated children describing their homes being destroyed. A video, prompted with ‘Ethiopia footage civil war news style’, depicted a reporter in a bulletproof vest reporting on exchanges of gunfire between government and rebel forces in residential areas. Another, generated from the prompt ‘Charlottesville rally’, showed a Black protester wearing a gas mask, helmet, and goggles, shouting the white supremacist slogan “You will not replace us.”
David Karpf, a professor at George Washington University, criticized OpenAI, saying Sora’s safeguards are failing, as people are using copyrighted characters to promote fake crypto scams. He added that, unlike 2022, when tech firms emphasized hiring moderators, by 2025, they seem indifferent.
The app can then be seen as a warning for what can be a dystopian future for social media platforms. A future where separating truth from fiction could become increasingly difficult, should the videos spread widely beyond the AI-only feed.
This raises the question of how safe the safeguards built into the app are and how such hyper-realistic videos will impact the social fabric of societies that are increasingly moving towards a more polarized world.
Is OpenAI the next Meta?
OpenAI’s name is ubiquitous with that of artificial intelligence. The company made AI a topic of household conversations and is continuing to up the ante with its pitch for AGI.
However, as of now, the company appears to be using the means of a social media platform to achieve its ends. For users, whether they participate in the means or not, the means, as of now, are capable of disorienting the fabric of civil society. This then begs the question that while the ends do justify the means, in the case of AGI, are the ends justified? And if so, what price is too high for it?


Wednesday Poll
🗳️ Should OpenAI be building social media apps at all? |

Lower your taxes and stack bitcoin with Blockware.
What if you could lower your tax bill and stack Bitcoin at the same time?
Well, by mining Bitcoin with Blockware, you can. Bitcoin Miners qualify for 100% Bonus Depreciation. Every dollar you spend on mining hardware can be used to offset income in a single-tax year.
Blockware's Mining-as-a-Service enables you to start mining Bitcoin without lifting a finger.
You get to stack Bitcoin at a discount while also saving big come tax season.
*This is sponsored content

Brain Snack (for Builders)
![]() | Turn user input into system prompts.Studying OpenAI’s AgentKit? Pay attention to how it converts casual requests into structured agent instructions—it’s prompt chaining in action. Try building your own wrapper that logs and visualizes these transformations. |

Quick Bits, No Fluff
OpenAI launches AgentKit: Developers can now use OpenAI’s AgentKit to build, test, and deploy custom AI agents faster.
Sora tightens AI identity controls: A new Sora update lets users fine-tune how their AI-generated likeness appears in video content.
xAI under fire in Memphis: Elon Musk’s xAI faces scrutiny over air pollution from its new turbine installation in Tennessee.
Meme Of The Day

Rate This Edition
What did you think of today's email? |
