A scrappy startup called Speak — founded by former Thiel fellows and hatched out of a South Korea “sandbox” before moving its headquarters to San Francisco — is now trying to do what big-name apps have promised for years: teach people to speak another language without enduring awkward human tutors. Its origin story reads like classic American entrepreneurship mixed with aggressive global ambition, but the real headline is that Speak is now heading into the U.S. market to take on household names like Duolingo.
Speak’s pitch is simple and unapologetic: get people talking. The app uses voice-based AI coaches built atop OpenAI models to simulate real-world conversations, correct accents and build what the company calls “language muscle memory.” The numbers are not trivial — the company says it has about 15 million downloads and has crossed a milestone of more than $100 million in annualized revenue while sitting at a roughly $1 billion valuation.
But let’s be blunt: Speak is marching into a ring already dominated by a giant. Duolingo last year reported hundreds of millions in revenue and has aggressively pivoted to an “AI-first” strategy, rolling out hundreds of AI-created courses and even moving to replace some human contractors with automated systems. That matters not just for market share, but for what kind of economy and culture we wind up with — one run by engineers in Palo Alto or one that still prizes skilled human work.
Conservatives should be perfectly comfortable celebrating innovation, but we should also be fierce defenders of workers and human dignity. The rush to replace teachers, tutors and contractors with silicon “tutors” risks hollowing out meaningful work and transferring power to a handful of dominant platforms that claim efficiency as a moral good. When big companies cheerfully substitute algorithms for people, communities pay the price in lost jobs and eroded commitments to craftsmanship.
That said, Speak’s story is the kind of startup energy America should applaud. Founded by dropouts-turned-founders who learned their trade in the real world and then raised capital from respected investors like Khosla, Accel and the OpenAI Startup Fund, the company has hauled in roughly $160 million in funding to date. Venture capital backing matters because it lets challengers build better products and keep incumbents honest — competition beats regulation when it comes to delivering value to consumers.
We should also not romanticize incumbents. Duolingo’s scale is undeniable, and its revenues dwarf those of newcomers, which is why challengers like Speak must fight smart to win trust and users. Competition from nimble, mission-driven startups is exactly how markets correct themselves — consumers get better options, pricing stabilizes, and innovation favors the product that actually teaches people to do useful things. The last thing we need is complacency in the face of a single platform’s dominance.
There’s also a practical business point conservatives should latch onto: employers are already buying into AI language coaching. Speak reports that about 500 companies, including global firms, now offer subscriptions to employees — a sign that corporate America values job-ready language skills and is willing to pay for what works. That’s good news for anyone who believes in skills, upward mobility and real-world results over gimmicks and engagement bait.
So here’s the bottom line for hardworking Americans: back the entrepreneurs who build tools that make people stronger, not the platforms that simply turn everything into another engagement metric. Support competition that preserves jobs, demands accountability from AI firms, and insists that innovation serve citizens rather than replace them. If Speak can deliver real conversational ability without sacrificing privacy, worker dignity, or human judgment, then let the market reward it — and let the rest of us hold Big Tech to a higher standard while we cheer on American ingenuity.
