I quit customer service to fund AI videos—costs hit hard

AI videos – Jonathan Laramy, 32 and creator of Chloe VS History, says making AI-generated YouTube videos is far more expensive and time-consuming than many people assume—leading him to quit his customer service job in June 2025. His channel’s workflow relies on multiple A

He quit his customer service job in June 2025 with a simple belief: he could turn his AI experiments into something bigger. Jonathan Laramy. 32. now runs the viral YouTube channel Chloe VS History—home to an AI-generated influencer named Chloe who “travels” through landmark historical moments. But behind the smooth, modern vlogger style, Laramy says the work is anything but quick or cheap.

Chloe VS History follows Chloe through scenes such as the Titanic’s maiden voyage and Pompeii on the day of Mount Vesuvius’ eruption. Laramy describes the format as a modern-day vlog, placing viewers inside historical settings and events. The character may be generated by AI. but he says the videos still demand creative direction—reviewing. adjusting. regenerating. and editing scenes until they work.

That reality collides with what he believes many viewers assume. A lot of people. he says. think AI videos are created by entering a few prompts. pressing buttons. and uploading the output. Laramy doesn’t deny the workflow uses AI. He just argues the process is far more complicated than the shortcut many people imagine.

Every video requires dozens of steps, and mistakes are expensive

Laramy’s production pipeline relies on multiple AI tools. For ideation and general script structure, he uses Claude. For images, he generates them using PAI 2 (an AI software by Utopai Studios), with Nano Banana Pro and ChatGPT 2.0. He then turns those images into video clips with Seedance 2.0. and uses an AI voice model to keep Chloe’s voice consistent.

Creating Chloe’s character came, he says, quicker than everything else. He generated between five and 10 versions of the character before choosing the one that felt most relatable. He then spent more time refining Chloe’s personality and dialogue so she would sound like a real influencer rather than an AI-generated character. Laramy says that attention to detail helped him answer one of the most common questions he gets: whether Chloe is real.

Laramy insists the channel doesn’t simply accept whatever the software produces. “Every scene needs to be reviewed. adjusted. regenerated. and edited until it works. ” he says—comparing it to how a real movie is made. Historical content adds its own hurdles, he says, because modern AI models can introduce anachronistic details. He cites examples from ancient Rome, where elements like street lamps and sunglasses can appear.

Because of that, a single long-form video can take weeks. The steps include scripting, image generation, video generation, voice work, multiple revisions, and post-production.

Laramy says he pays over $1,000 to make a single AI video

The production costs are the part viewers often miss.

For his channel, a long-form video typically costs between £300 and £800, which he translates to $400 to $1,070. He says the final cost depends on how difficult the topic is and how many times scenes need to be rerun before they look right.

Unlike traditional YouTube production—where a creator might spend on equipment or travel—Laramy says a large portion of the costs come from generating and regenerating content through AI models. He estimates that every revision has a cost attached to it. In practical terms, he says it works out to about $3 per 15-second clip.

But the real driver is the number of reruns. He says a single video can take 10–15 revisions, so costs accumulate even when the creative direction is already set.

And those expenses don’t depend on whether a video succeeds.

Even if a video underperforms, the bill still arrives

Laramy describes a second kind of risk—the financial one that exists whether the audience arrives or not. “And those expenses are there whether the video succeeds or not,” he says.

His early income also illustrates why that risk matters.

He initially tried making money on Instagram, TikTok, and YouTube, but only one platform took off. On Instagram, he says he wasn’t generating much income because the platform doesn’t pay for views. On TikTok. he says it wasn’t a major source of revenue because he wasn’t accepted into its Creator Rewards program. Even on YouTube, he says monetization took time.

For a while, he says he wasn’t making much money from the project.

The turning point came when he began producing longer videos. He says long-form content required a different approach: holding someone’s attention for 15 or 20 minutes is different from stopping them while they scroll for a few seconds. As more viewers watched his longer videos to the end and engaged with them, YouTube recommended his videos more aggressively.

Laramy describes a ripple effect: once the long-form uploads gained traction, they also helped drive attention to his Shorts. He says his Shorts hadn’t been getting much traction from YouTube on their own.

The expensive part isn’t just tools—it’s what creators build

Laramy says he has reached a conclusion about what separates successful creators from everyone else.

Access to AI tools, he argues, isn’t the divider. “Anyone can sign up for the software and generate images or videos,” he says. The harder part is figuring out what people want to watch—and then building a process that consistently delivers it.

For Laramy, that process has meant weeks of work, repeated revisions, and ongoing costs tied to generation itself. It also meant leaving his customer service job in June 2025, betting that a channel built on AI-generated history could grow into something that pays.

Now, he says, the channel earns him way more than his original customer service wage—but it took time to get there.

AI videos YouTube creator economy production costs Claude Seedance 2.0 PAI 2 Nano Banana Pro ChatGPT 2.0 Utopai Studios Chloe VS History Jonathan Laramy