The team that convinced Mark Zuckerberg to spend $2 billion on virtual reality goggles in 2014 has returned with a thesis that sounds almost quaint: what if the future of AI is less about typing prompts and more about having a conversation?
Sesame, the conversational AI startup founded by Oculus veterans, has launched its iOS app, entering a market that appears saturated but may be more fragmented than it looks. The company is betting that the current generation of AI assistants—ChatGPT, Claude, Gemini, Siri—have optimized for the wrong interface. Text boxes and chat logs, in Sesame's view, are a transitional technology, not a destination.
The Oculus playbook, applied to language
When Brendan Iribe and his co-founders pitched Oculus, the conventional wisdom held that VR was a solved problem—or rather, an unsolvable one. Headsets existed; they just weren't good enough to matter. The Oculus team argued that the technology had finally crossed a threshold where consumer adoption was possible, and they were right, even if the mass market took longer to materialize than anyone expected.
Sesame appears to be running a similar play. Voice assistants have existed for over a decade, and they have remained stubbornly mediocre. Siri, Alexa, and Google Assistant can set timers and play music, but they collapse under the weight of anything resembling nuanced conversation. The Sesame founders believe that large language models have changed the underlying economics: voice AI can now be genuinely useful, not just a party trick.
Why voice, why now
The timing is not accidental. Apple is reportedly preparing a significant Siri overhaul, and OpenAI's voice mode has demonstrated that LLM-powered conversation can feel remarkably natural. But neither company has fully committed to voice as the primary interface. ChatGPT still defaults to text; Siri still feels like a feature, not a product.
Sesame is positioning itself as voice-native from the ground up. The app is designed for extended, ambient interaction—the kind of always-on companionship that smart speakers promised but never delivered. Whether users actually want an AI companion that listens continuously is an open question, one that will be answered by adoption metrics rather than investor enthusiasm.
The crowded middle
The challenge for Sesame is that it occupies an awkward competitive position. It lacks the distribution advantages of Apple, the model sophistication of OpenAI, and the enterprise relationships of Anthropic. What it has is a team with a track record of identifying inflection points and a willingness to bet on interfaces that incumbents have neglected.
That may be enough. The AI assistant market is large enough to support multiple winners, and voice-first interaction is a genuine gap in the current landscape. But Sesame will need to demonstrate that its product is meaningfully better than the voice features that OpenAI and Apple are already shipping, not just philosophically different.
Our take
The Oculus founders have earned the right to be taken seriously when they claim a new interface paradigm is ready for prime time. But conversational AI is a harder problem than VR in at least one respect: the competition is not asleep. OpenAI, Apple, and Google are all racing toward the same destination. Sesame's bet is that they are racing in the wrong lane. It is a contrarian thesis, which is usually a compliment in Silicon Valley—until it isn't.
