John Borthwick once drove four hours to use a browser. It was 1993. He was studying in Philadelphia, and the machine he wanted was in MIT’s AI lab. On its screen, pictures from the Louvre appeared slowly, line by line. A museum in Paris had become reachable from a corner in Massachusetts. The journey home could not compete with that.
There is something pleasingly inconvenient about this beginning. A technology that would make distance feel irrelevant first required a road trip. Borthwick’s excitement belonged to the act of access: someone could publish something, and someone far away could see it. The images were small; the implication occupied considerably more room.
That experience helps explain the career that followed. Borthwick founded Betaworks in 2008, building a New York business around experiments in how people find, share, and use things online. Its history includes short links, live publishing analytics, animated GIFs, and conversational software. The products differ. The recurring question concerns the surface between a person and the network.
A browser, then a boardroom
Before Betaworks, Borthwick had already worked on both sides of the internet’s early growth: the small studio making things and the large company deciding where technology should go. His first company, WP Studio, was acquired by AOL. He led AOL’s product development studio and later worked on technology strategy and alliances at Time Warner. He also served as CEO of Fotolog, an early photo-sharing service.
His education followed a similarly mixed route: economics at Wesleyan University, then an MBA at Wharton. These are useful credentials for a person whose work requires judging both a product and the organization that might grow around it. Yet the browser story supplies a less formal qualification. He wanted to see the thing himself.
By the time he started Betaworks, search was no longer the only way to discover what mattered online. Social connections could carry links, recommendations, and conversation. New habits created demand for new tools. Borthwick’s wager was that this changing traffic deserved a workshop of its own, with room to build products and invest in people building nearby.
The useful part of a failed idea
The early Betaworks story has the awkward turns that polished company biographies usually tidy away. Findings, an initial product, failed to take off. But work around it helped produce Bitly. Another experiment, Firefly, offered real-time messaging on websites. It ran into a distinctly human problem: showing how few people were present could leave publishers feeling worse about their audience.
Billy Chasen, who created Firefly, also created the product that emerged from it: Chartbeat. The same underlying activity became useful when presented as monitoring rather than conversation. A publisher could see what was happening now. The first version appeared roughly 30 to 45 days after Firefly closed. Much of the code survived its previous business model.
This is a revealing kind of thrift. Betaworks could abandon a proposition without throwing away every observation made while building it. Borthwick’s preference for a working beta made that possible. A presentation can explain why an idea ought to succeed. A product gives its users a chance to disagree, sometimes with an efficiency that no board meeting can match.
His demand was concise: “I want to see betas, I want to see products.” It puts the emphasis on contact with reality. The distance between an attractive idea and something people actually use is where a great deal of company building happens.
Three ways to make the same bet
Betaworks developed three connected ways of working: build companies, support cohorts through Camp, and invest at the earliest stages. Borthwick described this arrangement in 2017 as a way to explore particular themes deeply. A studio project could reveal a practical difficulty. An investment could bring another founder’s answer into view. A cohort could let several teams compare what they were learning.
The themes moved with the interfaces. Early social distribution gave way to mobile, conversational software, voice, and other experiments. The logic was to concentrate attention where habits might change, then use more than one business structure to learn. For a founder, that meant the people evaluating the product might also have experience trying to build in the same territory.
The recognizable names accumulated. Betaworks’ company-building history includes Giphy and Dots; its investments and accelerated companies include Tumblr, Kickstarter, Medium, Anchor, and Gimlet. These belong to different corners of culture, but each concerns participation, publishing, play, or distribution. Borthwick’s work sits close to the places where software meets an audience.
Before the age of agents: Borthwick in a photograph published with his 2010 product-development interview. Photo: Mary Hodder.
A chatbot arrives at Camp
By 2015, Borthwick and the Betaworks team were focusing on machine learning. The first Camp, BotCamp, followed in 2016, centered on natural-language processing and conversational interfaces. Hugging Face took part. That detail matters because it places Betaworks’ AI interest years before the current wave of general-purpose chat interfaces became an everyday subject of conversation.
Camp made a theme concrete. Instead of waiting for a category to settle into familiar company types, Betaworks could bring early teams together while the category was still taking shape. Its successive programs examined voice, computer vision, synthetic media, and other applications. There was a repeated attempt to discover what people would do with a capability once it became usable.
In July 2025, Betaworks announced a $66 million third fund. The announcement looked back at approximately 100 companies backed through 12 Camp classes since 2016. Those figures describe institutional persistence as much as scale: returning to early product questions through successive cycles, while the underlying technology kept changing.
The fund sizes also set useful boundaries around the story. These are funds managed for investing in founders, rather than a measure of Borthwick’s personal wealth. Fund 3.0 followed $48 million for Fund 1.0 in 2016 and $46 million for Fund 2.0 in 2020. The newest pool supported the same pre-seed and seed territory where an idea still has plenty to prove.
The club without the disco
Borthwick’s appetite for gathering people has also taken a more literal form. Betaworks Studios opened as a membership club for builders and people interested in technology. In 2019, he explained that members could work, find collaborators, meet potential hires, and attend events. His friends, hearing that he was opening a club, apparently imagined disco. The actual entertainment involved rather more product conversation.
That joke catches an important feature of his working style. Community was something to organize in a room, with opportunities for useful collisions. A person who spends his career considering digital connection also found value in having people turn up in the same place. The network could have chairs.
He continues to participate in conversations that stretch beyond portfolio introductions. A 2025 Betaworks panel brought him together with Snipd founder Kevin Smith and The Next Big Idea’s Caleb Bissinger to discuss AI and podcasting. The questions included discovery, creation, editing, and asking questions of audio. It was another version of the familiar problem: a medium exists, but the way into it can change.
Later that year, Matt Mullenweg interviewed him about early investments, New York’s technology scene, and open versus proprietary AI. The subjects connected the city where Borthwick works with the larger question of who gets to build on emerging systems. The local room and the wider network have remained parts of the same conversation.
The internet needs repairs, too
There is a complication in Borthwick’s optimism: he has spent years criticizing what the connected world became. His 2018 essay about building bicycles for the mind argued that platforms had compromised their usefulness while pursuing attention and scale. He questioned the terms on which people surrendered their time and data. The disappointment came from someone who had helped fund and build the social web.
The bicycle image, borrowed from Steve Jobs, gave him a way to describe the alternative. A well-designed tool extends what its user can do. It should make a person more capable, with a purpose the person recognizes. That is a demanding standard for an industry that can count activity more easily than usefulness.
In 2020, Betaworks announced Betalab with Lupa Systems. Its proposed areas included authenticating synthetic media, addressing abuse and disinformation, improving privacy, supporting news, and strengthening identity tools. Borthwick also acknowledged that technical fixes alone would be insufficient. The plan combined investment with convening and research.
His current biography lists board work with New York Public Radio and Tech Policy Press. Those connections fit the questions he has raised about public information and technological power. He remains involved in making new products while also engaging with the institutions that consider their consequences. A builder can be excited about the next interface and still ask what the last one cost.
“Technical solutions alone will not save us.”John Borthwick, introducing Betalab, 2020
What happens after the prompt?
In 2022, Borthwick was writing about tools for thinking: software that could help organize notes, connect knowledge, and support sharing. He identified natural-language processing and large language models as enabling technologies. The ambition concerned ordinary intellectual work, including the untidy business of turning accumulated information into something one can use.
By spring 2026, Betaworks’ Camp theme was Agent Systems. The program ran from March 2 to May 15 and sought systems capable of perception, memory, planning, execution, and self-evaluation. The proposition shifted toward completing work across a whole process. It raised practical questions about reliability, recovery, and when a system should act.
The autumn 2026 program, The New Agentic Economy, extended that inquiry to company design. Its published schedule runs from September 14 to November 20. The thesis considers organizations built around AI, verification between people and agents, marketplaces with non-human participants, and systems that interact with the physical world. These are directions for experimentation, with outcomes still to be earned.
The pattern brings Borthwick back to his original curiosity. The browser offered a new way to reach a museum. A social feed changed the route to an article. Voice offered another way to navigate. Agents introduce the possibility of asking for an outcome and letting software work through the steps. Each changes the relationship between intention and action.
The four-hour drive remains a useful image because it preserves the effort in that curiosity. Borthwick went somewhere to look at a possibility before it was convenient. His career has repeatedly made room for people doing the same thing: building an imperfect version, watching what happens, and deciding what deserves another attempt. The future, in this telling, still needs someone to try the browser.