A Founder’s Gamble on ‘Old School Web’ Proves Successful

Craig Campbell stepped away from the flood of investor cash pouring into AI to launch something unexpected: a website. With a background as a former engineer at Meta and a successful tech founder who sold his last venture—a Shopify e-commerce tool—just as the AI craze was taking off, he had plenty of opportunities to dive into AI. Instead, he had a different vision.

While many shy away from the website game, especially with the looming threat of Google’s dominance, Campbell forged ahead and built Past Maps into a thriving business through organic search. True to its name, Past Maps allows users to explore historical maps overlaid on modern-day maps, with adjustable opacity for seamless transitions. The maps are sourced from public data like the US Geological Survey, but the innovative tools for exploration were crafted by Campbell himself. He originally developed them to enhance his metal detecting hobby, helping him locate old structures and trails. After sharing his creations on Reddit, he discovered a community eager to use his tools, leading to the birth of his latest tech venture.

You don’t need to be a treasure hunter to appreciate Past Maps. It’s a goldmine of information for anyone curious about their surroundings. Personally, I’ve used it to understand the historical shape of the Duwamish River before it was straightened for navigation. Users leverage it for various purposes, from genealogy research to tracking old oil wells. It serves as both a research tool and a source of enjoyment.

The growth has been impressive. Campbell reports that traffic has surged from around 20,000 active users per month to over 300,000 in just three years. The revenue is sufficient to support him and his wife, who also contributes to the business. Yet, he can’t help but wonder how much more he could have made had he pursued those AI investments. “I’m making about the same as when I was an E4 at Facebook, which is a mid-level engineer,” he reflects.

The primary source of traffic for Past Maps comes from Google Search. Early on, Campbell noticed that his site was climbing the search rankings as people sought historical information about places of interest—like a church their grandmother attended or abandoned mines in a specific area. By optimizing his maps and webpages for Google’s algorithms, he initiated a positive feedback loop. “As I started to publish this data and make it accessible to Google, traffic just began to grow.”

“This is how the web is meant to function. It’s the old-school web,” he explains. “It’s still alive, but only in very niche areas.”

In the past, web publishers relied heavily on display advertising for revenue. While Past Maps offers a free account, deeper access requires a $9 weekly pass or a $52 annual subscription. This subscription model shields Campbell from the unpredictable nature of advertising budgets and the ad tech landscape, which is largely dominated by Google—deemed an illegal monopoly by the DOJ in 2025.

Despite the AI takeover of the open web, Campbell has embraced AI tools to streamline his operations. Previously, he spent one to two hours daily managing service requests and crafting detailed emails. Now, he utilizes a local agent model on his desktop for front-line triage. This automated task runs hourly, provided his laptop is on, filtering out spam and marketing emails, identifying urgent matters, and drafting responses. This has reduced his customer service workload to about 10 minutes a day.

“I do encounter some frustrated customers,” Campbell admits. “If they request a refund, it automatically processes the refund and subscription cancellation through Stripe. It handles everything, then notifies me.” He reviews the request, makes a decision, and sends a final message.

Campbell is also leveraging AI to develop an Optical Character Recognition (OCR) tool for old maps. “Cartographers can be quite challenging,” he jokes. Historical maps pose unique difficulties for existing OCR systems due to their curved labels, inconsistent spacing, and crowded text. Off-the-shelf tools often struggle, but Campbell has found success using modern LLMs that incorporate reasoning. However, he emphasizes that it’s not as simple as just instructing an agent to “OCR these maps.”

Instead, he combines human creativity and experimentation with the capabilities of LLMs, rather than relying solely on the technology. “It still lacks that human-level reasoning and creativity, the ability to connect decades of experience with these tools,” he notes. “You need to infuse that human touch into the process.”

While Campbell may have turned his back on the AI gold rush, he seems to have discovered a winning formula for an online business in the era of AI and automated summaries. By starting with a passion, creating something genuinely useful, and sharing it with like-minded individuals, he’s built a solid foundation. His daily operations may look vastly different from a decade ago, but the core elements that drive his success remain deeply human.