About — BorderPage | A 25-year bet against AI slop
About / Our Stand

Two builders.
25 years between them.
One thesis.

The web doesn't have an AI problem. It has a sameness problem. We built BorderPage because no one else was treating design and proprietary perspective as the actual ranking factors they've become.

§01

We're not anti-AI. We're anti-average.

→ IThe wrong fight

Half the industry is busy "humanizing" AI text or hiding their AI workflow from Google. That's a losing fight. The model isn't the problem.

→ IIThe actual fight

The fight is information gain vs. recycled commentary. Proprietary data vs. paraphrase. Designed pages vs. theme #743. That's where the algorithm now lives.

→ IIIThe compounding cost

Every templated article you publish trains crawlers to expect more of the same from your domain. Brand decay is invisible until it isn't.

→ IVThe way out

Train a model on what only you know. Design components nobody else has. Publish at human velocity. Repeat. There is no shortcut and that's the whole moat.

02 — The Origin

We watched the web turn into a monoculture. Then we got tired of watching.

Sometime around late 2023, both of us — independently, in different time zones, on completely different client engagements — started seeing the same thing. Every new client's site looked like every other client's site. Same hero. Same three-column grid. Same generic "About" page. Same blog with 200 thin posts written by a Zapier-and-ChatGPT pipeline.

The content marketing teams were exhausted. Their rankings were flat or falling despite shipping more than they ever had. The "we need to scale content with AI" mandate from leadership had quietly become a "why isn't this working" panic. The honest answer — "because you're publishing the same thing as everyone else, dressed up in the same template" — was hard to say out loud, because nobody had a tool that solved it.

We tried, at first, to solve it as a service. Custom design + bespoke content strategy + in-house writers. It worked beautifully for three clients. It also took 8 weeks per page and cost more than any sane SaaS company could justify. The unit economics were broken. Service alone wasn't scalable. Software alone wasn't enough. The mismatch between "what actually works" and "what's affordable to operate" was the whole problem.

The web doesn't reward AI content or human content. It rewards information gain. We just built the cheapest way to produce it.

BorderPage is what we wish had existed when we were trying to thread this needle for clients. A custom-trained AI on the things only your company knows — your client calls, your benchmarks, your founders' perspectives. A custom-designed component library so the output doesn't look like template #743. A middleware sync so your content team operates in Google Docs and Word, where they already live. A publishing scheduler that ramps slowly enough to look like editorial activity, not a content factory.

It's hybrid by necessity, not preference. The first month is genuine consulting work — designing components, mapping IDs, training the LLM project on your data — and nothing about that is templatable. After that, the SaaS handles the boring parts on autopilot. That's the deal. That's the whole product.

03 — The Team

A web designer and an inbound marketer walk into a content problem.

FOUNDER 01
a
The Designer
15 yrs · Web Design & Frontend Architecture

Started building sites in 2010 — back when "responsive" was still a debated term. Has shipped UI for fintechs, healthcare platforms, marketplaces, and far too many SaaS landing pages. The component library is his obsession. Believes that "design system" is the most overused phrase in tech, and that real systems live in the constraints, not the docs.

15years shipping
200+sites built
themes refused
FOUNDER 02
b
The Marketer
10+ yrs · Inbound, SEO & Content Strategy

Has been on the inbound side since the original HubSpot certification was a print PDF. Took three different companies through full-funnel SEO programs and watched two of them get smoked by Google's helpful content updates because they'd over-indexed on volume. Now writes about information gain like other people write about cardio — frequently, with conviction, and to anyone who'll listen.

10+yrs inbound
3programs led
2algorithmic lessons
04 — How We Got Here

A short, mostly-honest history of the sameness problem.

2003
WordPress 1.0 ships
A brilliant tool that became the world's most successful CMS — and also the foundation of a homogeneity problem that would take 20+ years to surface.
2010 – 2018
The age of "content marketing"
Brands learned to publish. Publishing volume became a vanity metric. The first generation of "scale content" tooling appeared. Templated thinking became normalized.
2022
GPT-3.5 + ChatGPT public release
Suddenly, anyone could generate plausible-sounding content in under 60 seconds. Within 12 months, the web was flooded with templated AI articles indistinguishable from one another.
2024
We started prototyping
Spent ~9 months testing whether a custom-trained LLM project + a designer-built component library could systematically deliver information gain at velocity. It could.
05 — What We Stand For

Three principles. No more.

i

Information gain over volume

We'd rather ship 4 genuinely original pages a month than 40 paraphrased ones. We measure ourselves on signal density, not output.

ii

Design is a ranking factor

The crawler now reads engagement, dwell time, and bounce as proxies for quality. A boring page is an under-performing page. Period.

iii

Hybrid because honest

We refuse to sell pure software that promises results we know require services. The first month is consulting. That's not a limitation — it's the work.

Read this far? Then you're probably our kind of marketer.

Skip the brochure call. We'll review your last 10 pages, score them on information gain, and tell you whether we can move the needle.

Get the free audit