Joshua bought a Mac Studio—but the reason has a lot less to do with simply wanting a faster computer than it does with AI.

In this episode of Tek With Josh, Joshua explains why his growing use of AI agents pushed him toward buying an M5 Ultra Mac Studio. He breaks down how he currently uses agents for editing books, publishing, marketing, business ideas, analytics, and other recurring tasks, and why running more of those workflows in the cloud started becoming harder to justify financially.

Joshua takes a closer look at his experience using Genspark for AI-powered book editing, including how quickly a single project could burn through credits once promotional pricing disappeared. He also discusses alternatives like paying for higher-tier cloud AI plans and why spending that money every month made him start considering something different: owning the hardware and running more of his AI workflow locally.

That eventually led to the Mac Studio. Joshua explains why he considered an M5 Max configuration with more unified memory before ultimately spending more for the M5 Ultra. He breaks down the additional CPU and GPU cores, doubled memory bandwidth, and why he believes those advantages make more sense for the kinds of local AI models and agents he wants to run—even with less unified memory than the Max configuration he originally considered.

The conversation also looks at Hermes, the local agent system Joshua plans to experiment with once the Mac Studio arrives. His goal is to build agents that can remember his workflows, develop skills, create specialized agents for individual tasks, track publishing and website data, help manage schedules and social media, and automate work he currently handles manually. The Mac Studio will also take over some duties from his Mac mini server while becoming a machine for video editing, YouTube production, and future projects across Books By Josh and Tek With Josh.

Topics Covered

Why Joshua bought a Mac Studio
How AI agents have changed Joshua’s workflow
Using agents for book editing, publishing, and marketing
Genspark AI agents and credit usage
Why promotional AI credit pricing changed the equation
The potential cost of editing books with cloud AI agents
Cloud AI subscriptions vs. owning local hardware
Why Joshua wanted to run AI agents locally
M5 Max vs. M5 Ultra Mac Studio
Why 128GB of memory wasn’t automatically the better choice
CPU and GPU differences between the Max and Ultra
Why memory bandwidth matters for local AI
The Mac Studio upgrade program and monthly cost
Running Hermes as a local AI agent
Hermes memory, skills, and specialized agents
Automating podcast and website workflows
Using AI to track book rankings, reviews, and analytics
AI agents for scheduling and social media
Moving Mac mini server duties to the Mac Studio
Video editing and content creation
Plans for Books By Josh on YouTube
Potential future Tek With Josh video content
Why Joshua didn’t actually need a Mac Studio
Why owning the computing power made sense for his use case

Key Takeaway

Joshua didn’t buy the M5 Ultra Mac Studio because his current MacBook Pro was no longer powerful enough. He bought it because the way he uses computers is beginning to change. As AI agents become a bigger part of his publishing, business, analytics, and content workflows, continuously paying for cloud computing started to look less attractive than investing in hardware he could run locally. The Mac Studio is a major expense and certainly not something everyone needs, but for Joshua, it represents an experiment in moving from renting AI computing power to owning the machine that runs it.

Listen, read, and explore more at:
👉 BooksByJosh.com