Oculus Veterans Skip Meta's AR Dream, Build $250M Voice Companion Instead

I watched Brendan Iribe demo the first Oculus Rift in 2012. His eyes lit up describing presence - that magical moment when your brain forgets you're wearing a headset. Now he's chasing the same lightning with voices instead of visuals.

Iribe left Meta in 2018, right as Zuckerberg pivoted hard toward the metaverse. While Meta burns billions on cartoon avatars, Iribe's new company Sesame just raised $250 million from Sequoia to build AI companions that live in your pocket - and eventually, lightweight smart glasses.

Those numbers reveal something profound. People crave connection with AI, not just utility. ChatGPT feels like talking to a search engine. Sesame feels like talking to a friend.

The 200-Millisecond Gamble

Here's where it gets technically fascinating. Sesame built a Conversational Speech Model (CSM) that processes text and audio tokens simultaneously, targeting response times of just 200-300 milliseconds.

That's the difference between:

• Traditional AI: You speak → transcription → text processing → text-to-speech → response
• Sesame's approach: You speak → immediate audio understanding → direct speech response

The result? Natural interruptions, emotional responses, actual laughter. Not the robotic "I understand you're frustrated" garbage we're used to.

But speed isn't everything. The real breakthrough is emotional bandwidth. Sesame's AI can shift tone, respond to context, and maintain conversational flow. It's designed for the kind of back-and-forth that feels less like querying a database and more like texting your smartest friend.

The Oculus Reunion Tour

Iribe didn't build this alone. His co-founder Ankit Kumar was CTO of Ubiquity6 (remember spatial computing?). Nate Mitchell, another Oculus co-founder, runs product. They've assembled the Avengers of VR to tackle voice AI.

This team knows how to:

1. Build hardware that doesn't suck

2. Create software people actually want to use

3. Navigate the hype cycle without face-planting

Their plan? Start with the iOS app (launching now), prove the AI works, then move to lightweight smart glasses for all-day wear. It's the opposite of Meta's strategy - software first, hardware later.

The Uncomfortable Truth About AI Companions

Sesame's early success exposes something we don't like admitting: humans are lonely. Five million minutes of conversation with artificial beings suggests we're starving for interaction that feels genuine.

But here's the twist - Sesame might succeed because it's artificial. No judgment, infinite patience, always available. Your AI companion won't ghost you or get tired of your problems.

The competition is fierce though. Meta has billions to throw at AR glasses. OpenAI has the best language models. Friend (the controversial AI companion hardware) already launched. Apple's building Siri 2.0.

My Bet

Sesame wins the premium AI companion market but struggles with mass adoption. Their Oculus pedigree and voice-first approach create a genuinely differentiated product that tech enthusiasts love. The $250M war chest gives them runway to perfect the glasses.

But breakthrough consumer hardware is brutally hard. Even brilliant teams fail. I predict Sesame becomes the "Porsche of AI companions" - beloved by early adopters, acquired by Apple or Google before reaching mainstream scale. The technology survives. The company probably doesn't.

The real winner? All of us, finally getting AI that feels less like talking to a computer and more like talking to someone.