Blair Harrison keeps returning to the machinery behind the picture. Long before a television channel could be assembled in a browser, he was working on computer graphics. Before viewers expected any clip to be searchable, he co-founded a company that indexed live television. Before “viral” became the house dialect of the internet, he ran IFILM, where programmers sifted the day's oddities into an early map of online taste. Today, at Frequency, he is working on software that can schedule, package and distribute streaming channels, then learn how to do the job better.
The formats change. The screens breed. The durable problem is organization: how to turn a chaotic library of moving images into something a person might actually watch. Harrison's career is less a sequence of reinventions than a long argument that video should behave like software.
He is an English-born engineer who studied Computer Systems Engineering at the University of Bristol from 1985 to 1988 and moved to Los Angeles in the late 1990s. His public X biography is wonderfully economical: “CEO Frequency. Programmer. Itinerant optimist.” The last two words do most of the character work. A programmer notices systems. An optimist believes the system can be improved. An itinerant one is willing to follow the problem through several eras of bad bandwidth, strange business models and conference acronyms.
A career with one persistent plot
In 1995, Harrison founded Second Nature Industries, a computer-graphics business later acquired by Autodesk. Three years later he co-founded FasTV and became its CTO. FasTV used video recognition to generate searchable clips from live broadcast television, a striking idea at a time when online video was slow, scarce and often the size of a postage stamp. The company anticipated a world in which pictures would need metadata, indexing and retrieval as urgently as text did.
Then came IFILM. Harrison joined as chief technology officer in 2001 and went on to lead the company. IFILM was an online destination for short films, trailers, commercials and the species of clip that arrived in your inbox with an imperative subject line. Its Viral Vault, launched in 2001, gave a durable name to a new circulation system: viewers sending video to other viewers because they felt compelled to share the surprise.
A 2007 visit to IFILM's Los Angeles office caught the business in its native habitat: brick walls, large rugs and a daily editorial meeting where programmers judged the strange weather of internet culture. Harrison said the company looked for people who had demonstrated passion for internet pop culture, which he distinguished from regular pop culture. The distinction was shrewd. The web was not simply a cheaper pipe for television. It had its own tempo, its own jokes and its own unlikely stars.
Viacom acquired IFILM in 2005, the same year Harrison received a Wired Rave Award. Yet another remark from that era is more revealing than the trophy. Asked about delivery infrastructure in 2001, he said building a content delivery network was not something he wanted to do. Prices were falling, specialist providers existed, and IFILM's value lay elsewhere. Even in a frontier market, he was drawing a disciplined line between the hard thing his company should own and the hard thing it should rent.
“Programmer. Itinerant optimist.”Harrison's public X bio, in five efficient words
Frequency finds its frequency
Harrison founded Frequency in 2010. Its earliest public incarnation helped people gather and discover online video from multiple sources. The mission evolved as television itself migrated into apps, connected sets and free ad-supported streaming channels. By Harrison's account, Frequency entered FAST around 2018 or 2019, when Pluto TV and Xumo were early anchors and the category still felt like “fun and games.”
Fun is a lovely prototype and a terrible operating system. Once major rights holders, broadcasters and advertisers arrived, a channel required dependable ingest, schedules, rights rules, graphics, ad breaks, live feeds, analytics and delivery to many platforms. Frequency moved into this operational middle. Its Studio platform became the place where a media company could turn a library into a channel and send it outward without rebuilding the plumbing for every destination.
The business model has been equally legible. In a 2022 interview, Harrison and colleague Jon Cohen described a flat fee per channel rather than an accumulating toll for every additional distribution platform. In 2023, Frequency partnered with Wurl: Frequency would concentrate on channel creation and complex distribution, while Wurl supplied monetization services. Harrison's formulation was collaborative rather than imperial. “Together we have a combined solution that is more valuable than the sum of the parts,” he said.
Own the workflow where your product creates leverage. Partner for the adjacent layer when a specialist already has the incentives, relationships and machinery to do it well.
That focus did not mean standing still. Frequency added managed services for organizations that wanted expertise along with software. Fusion, introduced in 2024, connected traditional broadcast schedules to cloud workflows. Its Program Templates promised to reduce scheduling time by as much as 95 percent. Live sports and local news entered the mix. The company was no longer arguing that streaming would replace television. Harrison described traditional TV, FAST and OTT as parts of one universe: “Streaming Television.”
One library, many moving parts
The autonomous media company
At the 2025 StreamTV Show, Harrison gave a keynote with a title that sounded like either a thesis or a mild threat: “The Autonomous Media Company.” When an interviewer asked why he had omitted a subtitle, Harrison said he wanted to keep people guessing. The answer was playful. The idea beneath it was concrete.
He was not proposing synthetic actors or algorithmic plots. He was talking about the work around content: processing it, packaging it into a viewer experience, programming it and distributing it. A cloud channel could react to seasonality, external events and performance data. If the software understood what was in a library and how viewers responded, programming could move at social-media speed instead of waiting for the slower cadence of traditional television operations.
“We're building software that lets channels essentially run themselves,” Harrison said. The phrase invites a picture of an empty control room, but his examples remained editorial. A movie channel might adjust its lineup for a season. Programming could respond quickly to a newsworthy event. The intelligence would operate within rules rather than merely filling blank hours.
Frequency made that model explicit in April 2026 with a platform-wide AI architecture for its Studio product. The company introduced Blueprints, natural-language definitions of a channel's programming logic, editorial rules and optimization goals. A Blueprint might begin as a paragraph, a spreadsheet schedule or a conversation. AI Scheduling was the first application, designed to assemble programming while respecting rights windows, audience patterns and the identity of the channel.
The clever bit is not an AI button. It is shared context across the whole channel.The platform argument behind Frequency's 2026 Blueprints
Harrison's criticism of scattered AI features was precise: a tool for one workflow and an assistant for another do not necessarily share context or compound in value. Frequency's claim rests on architecture. One platform and one version for customers can allow scheduling, metadata, distribution, graphics and live operations to understand the same channel. For builders in other industries, this is the useful part to steal. An intelligent feature saves a task. A coherent model of the customer's operation can improve a system.
Optimism, with wiring diagrams
There is a pleasing symmetry to Harrison's story. FasTV tried to understand live television well enough to make it searchable. IFILM learned to program the unruly tastes of the early web. Frequency began by organizing online video, then grew into software for programming and moving entire channels. Blueprints now attempt to express the character of a channel as rules a machine can interpret.
The aspirations have expanded, but the craft remains recognizably engineering: separate concerns, keep context, automate repetition, measure the result. Even Frequency's 2026 move deeper into monetization follows the operational logic. Harrison said the next phase of streaming would depend not only on growing audiences but on monetizing them effectively. A channel that cannot support its economics is merely a very elaborate playlist.
Technology careers are often flattened into a gift for prediction. Harrison's looks more practical. He stayed close to one medium while its bottlenecks migrated. First the picture had to be rendered. Then delivered. Then found. Then programmed across an explosion of destinations. Now it must react intelligently without misplacing rights, breaking the schedule or losing its editorial shape.
The television of the future may still look comfortably like television. It will have familiar shows, live events, ad breaks and the small luxury of not choosing every single thing. Behind the screen, though, the control room is dissolving into software. Blair Harrison has been walking toward that room for thirty years. He has finally reached the interesting part: teaching it how to think.