Get live statistics and analysis of Hermes ᯅ's profile on X / Twitter

I spent the last 12 months building voice AI agents. Here’s the 10 things I wish someone had told me about when you move from “demo on localhost” to “real users.” (ranked in order of importance) 🧵…

If you're still building chatbots instead of conversational AI, you're already behind the ball. A16z in their 2025 update on Voice Agents calls out Conversational AI as the primary way people will interact with AI.

Good morning. I got up early and came to watch you sleep.

San Francisco Halloween sucks, thankfully we celebrate on the east coast. Happy Halloween 🎃

Join me today at 12:30 PM EDT / 9:30 AM PDT for a candid discussion on what it takes to build your own real-time live streaming voice and video solutions. 👉 Reserve your spot now 🔗👇

This is Doug. Not much to look at, but at home they really keep things together. It’s not AI, it’s tele-operated robotics. Don’t call them animatronic; Doug doesn’t like that.

The artificial intelligence arms race is transforming industries and fundamentally changing how we interact with technology. While chatbots and text-based AI interfaces have revolutionized communication, the need to type out our thoughts forces us to translate our natural instincts into a more structured, less spontaneous format. As the AI industry evolves at lightning speed, so do the ways users interact with these new models and agents. Not long ago, interacting with AI meant typing prompts and waiting for a written response. That static interaction made sense in the early days — language models were still learning to handle context and semantics. However, as the industry aims for mass adoption, it faces high expectations from users accustomed to seamless, immediate communication. Real-time voice interactions are emerging to reshape a landscape once dominated by text-based prompts. What began with Siri, Alexa, and Google Assistant has now given way to a generational shift in how people expect to interact with AI. Since their launch, these assistants have seen only incremental improvements in handling off-script tasks. But innovations like OpenAI’s Real-Time SDK are disrupting the field, introducing agents that far surpass the capabilities of legacy voice assistants. Shifts of this magnitude affect not only end users but also developers and businesses. It’s no longer enough to offer an app or website; users now expect AI interactions to feel natural and deeply contextual. Designing UX for voice and video requires a new mindset. Pacing, intonation, and other cues replace written prompts, creating a steep learning curve. Yet, the rewards are significant. Those who embrace this shift can differentiate themselves with more engaging, immersive, and intuitive AI-powered features.

Your Chatbot sucks. And deep down you know it. Typing out messages to a computer program feels about as satisfying as a gas station sandwich. Chatbots are the new Clippy—annoying, dated, and despised. Don’t even start with that Siri and Alexa garbage. We’ve gone from "How can I help you today?" to "I’m sorry, I didn’t catch that"—on repeat. Here's the thing, voice-based AI could revolutionize how we interact with technology—if only products could get it right. Imagine having a natural, free-flowing conversation with an AI that understands every word and handles complex tasks within the app it’s embedded into. So, founders and developers alike, listen up... If you’re not building with voice-driven AI, you’re already behind. And guess what? The smart ones are already ahead. YC reported that 22% of their latest cohort is building with voice-driven AI. The rest? Playing catch-up. Conversational AI is the next big disruptor—letting humans engage naturally with voice-powered integrations that 10x productivity without the friction of typing. The future of AI isn't typed. It's spoken. Don't get left on mute.

Join me today at 12:30 PM EDT / 9:30 AM PDT for a candid discussion on what it takes to build your own real-time live streaming voice and video solutions. 👉 Reserve your spot now 🔗👇

This is Doug. Not much to look at, but at home they really keep things together. It’s not AI, it’s tele-operated robotics. Don’t call them animatronic; Doug doesn’t like that.

I spent the last 12 months building voice AI agents. Here’s the 10 things I wish someone had told me about when you move from “demo on localhost” to “real users.” (ranked in order of importance) 🧵…

San Francisco Halloween sucks, thankfully we celebrate on the east coast. Happy Halloween 🎃

Good morning. I got up early and came to watch you sleep.

Your Chatbot sucks. And deep down you know it. Typing out messages to a computer program feels about as satisfying as a gas station sandwich. Chatbots are the new Clippy—annoying, dated, and despised. Don’t even start with that Siri and Alexa garbage. We’ve gone from "How can I help you today?" to "I’m sorry, I didn’t catch that"—on repeat. Here's the thing, voice-based AI could revolutionize how we interact with technology—if only products could get it right. Imagine having a natural, free-flowing conversation with an AI that understands every word and handles complex tasks within the app it’s embedded into. So, founders and developers alike, listen up... If you’re not building with voice-driven AI, you’re already behind. And guess what? The smart ones are already ahead. YC reported that 22% of their latest cohort is building with voice-driven AI. The rest? Playing catch-up. Conversational AI is the next big disruptor—letting humans engage naturally with voice-powered integrations that 10x productivity without the friction of typing. The future of AI isn't typed. It's spoken. Don't get left on mute.

If you're still building chatbots instead of conversational AI, you're already behind the ball. A16z in their 2025 update on Voice Agents calls out Conversational AI as the primary way people will interact with AI.

The artificial intelligence arms race is transforming industries and fundamentally changing how we interact with technology. While chatbots and text-based AI interfaces have revolutionized communication, the need to type out our thoughts forces us to translate our natural instincts into a more structured, less spontaneous format. As the AI industry evolves at lightning speed, so do the ways users interact with these new models and agents. Not long ago, interacting with AI meant typing prompts and waiting for a written response. That static interaction made sense in the early days — language models were still learning to handle context and semantics. However, as the industry aims for mass adoption, it faces high expectations from users accustomed to seamless, immediate communication. Real-time voice interactions are emerging to reshape a landscape once dominated by text-based prompts. What began with Siri, Alexa, and Google Assistant has now given way to a generational shift in how people expect to interact with AI. Since their launch, these assistants have seen only incremental improvements in handling off-script tasks. But innovations like OpenAI’s Real-Time SDK are disrupting the field, introducing agents that far surpass the capabilities of legacy voice assistants. Shifts of this magnitude affect not only end users but also developers and businesses. It’s no longer enough to offer an app or website; users now expect AI interactions to feel natural and deeply contextual. Designing UX for voice and video requires a new mindset. Pacing, intonation, and other cues replace written prompts, creating a steep learning curve. Yet, the rewards are significant. Those who embrace this shift can differentiate themselves with more engaging, immersive, and intuitive AI-powered features.