- The Prohuman
- Posts
- GPT-5.4 aims at agents
GPT-5.4 aims at agents
Plus: Netflix makes a rare AI acquisition
Hello, Prohuman
Today, we will talk about these stories:
OpenAI releases GPT-5.4
Netflix buys Ben Affleck’s AI film startup
The new middlemen of the AI boom
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
OpenAI pushes AI deeper into real work

Image Credits: Open AI
OpenAI says its newest model is built for actual office tasks.
The company released GPT-5.4 in ChatGPT, the API, and Codex, positioning it as a model designed for professional work across documents, spreadsheets, coding, and multi-step software workflows. The system also introduces stronger computer-use abilities, allowing agents to operate apps and websites, with benchmarks showing 75% success on OSWorld, a test for navigating real desktop environments.
You can hear the keyboard clicks. GPT-5.4 also expands context to 1 million tokens, improves web research, and reduces hallucinations, with OpenAI claiming responses are 33% less likely to contain false claims compared with GPT-5.2.
This release isn’t about chat quality. It’s about turning models into workers that can move through software, run tools, and finish multi-step tasks without constant supervision.
OpenAI is pushing hard toward agents that can handle spreadsheets, legal documents, coding environments, and browser tasks in one system.
If the computer-use layer holds up in real deployments, software workflows could shift toward AI-driven automation instead of human-driven interfaces. The pressure on SaaS tools to become AI-operable systems will increase quickly. If models can actually run software, what parts of everyday office work stay human?
Netflix brings filmmaker AI in-house

Image Credits: Reuters
Netflix just bought a 16-person AI startup founded by Ben Affleck.
The company acquired InterPositive, an AI tools startup Affleck founded in 2022, and is bringing its entire engineering and research team inside Netflix. The system trains models on a production’s dailies and lets filmmakers adjust things like lighting, color, visual effects, and shot continuity during postproduction.
A quiet edit bay at night. Netflix says the tools are meant for filmmakers already working on a project, not for generating movies from text prompts.
This feels like a defensive move as much as a creative one. Netflix wants AI inside the filmmaking workflow, but it also wants the system shaped by directors and cinematographers instead of tech startups building generic video generators.
Affleck framing the tools around postproduction is smart because that’s where studios already spend a lot of time and money fixing shots.
If this works, studios may start building proprietary AI toolchains tied to their own productions and datasets rather than relying on outside platforms. The streaming companies would effectively control both the content and the software used to finish it.
If every studio builds its own filmmaking AI, will the next creative arms race happen in postproduction tools?
Power and land brokers step into the AI rush

Image Credits: New York Times
Electricity, not chips, is becoming the constraint.
Cloverleaf Infrastructure is buying or optioning large parcels of land and pairing them with guaranteed electricity supply, then selling the package to companies building A.I. data centers. The firm recently sold a 1,900-acre Wisconsin site with 1.3 gigawatts of power to Vantage Data Centers, which plans a $15 billion complex tied to Oracle and OpenAI.
It smells like manure out there. A.I. companies are hunting for roughly 85 gigawatts of new data-center power by 2030, which S&P Global says already exceeds what the grid can easily provide.
This business exists because the grid moves slowly while the A.I. buildout is moving fast, so someone has to line up land, permits, and power years before the tech companies are ready to build.
Brian Janous is basically selling preparation, and the fact that investors put $300 million behind the idea tells you how frantic the power search has become. The room gets loud.
These brokers will likely multiply as utilities, landowners, and A.I. firms all try to move faster than the permitting system allows. The tension is already visible at town halls where residents worry about water, noise, and rising electricity prices tied to huge data facilities.
If power is the real bottleneck, who ends up controlling the next wave of A.I. infrastructure?
Prohuman team
Covers emerging technology, AI models, and the people building the next layer of the internet. | ![]() Founder |
Writes about how new interfaces, reasoning models, and automation are reshaping human work. | ![]() Founder |
Free Guides
Explore our free guides and products to get into AI and master it.
All of them are free to access and would stay free for you.
Feeling generous?
You know someone who loves breakthroughs as much as you do.
Share The Prohuman it’s how smart people stay one update ahead.



