Siddharth Sharma (@sidroopdaska) header
Siddharth Sharma (@sidroopdaska) profile picture

Making voice AI as familiar as talking to a person @familiar_io

San Francisco
Joined October 2014
Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)Photo by Siddharth Sharma (@sidroopdaska)
We measure success by outcomes on real sales calls, not benchmark scores. Our breakdown of three popular duplex speech benchmarks, what they measure, what they miss and why we don't publish scores.
Outcome maxing > benchmark maxing. A model can top speech benchmarks and still struggle to qualify prospects and book sales meetings. We broke down three popular duplex speech benchmarks, what they measure, what they miss, and why we don't publish benchmark scores @familiar_io. Instead, we focus on business outcomes. We've already seen our duplex speech model match human SDR performance on outbound lead qualification. Full breakdown: familiar.io/blog/benchmarks-…
There's a new version of this post
19
[1/3] One thing we keep hearing from sales leaders is that the biggest value of AI in outbound sales isn't just scale or cost savings. It's removing the management burden of SDR teams. Hiring, training, coaching, motivation and attrition are a constant cycle. As teams grow, maintaining consistent performance becomes increasingly difficult.
1
3
2
159
[2/3] AI voice agents offer a different approach. They can qualify leads, handle follow-ups and run outreach at scale. Feedback from live calls can improve every agent. But none of this matters if the AI can't hold a good sales conversation. @familiar_io, we've matched human SDR performance on outbound lead qualification calls.
2
76
[3/3] How we're doing it: - A full-duplex speech model built in-house that listens while speaking and handles interruptions, overlaps and background noise naturally. - Training on calls from top-performing sales reps to adapt to prospects, ask follow-up questions and handle objections. More here: familiar.io/blog/familiar-vs…
33
Exciting to see @familiar_io's research translate into real-world outcomes! We matched the performance of a human sales rep on lead qualification calls. More to come as we rollout our duplex speech models.
1.8% apart. We compared Familiar with a human sales rep on outbound lead qualification for PolyborgAI. We saw similar disqualification and booking rates. Familiar also surfaced insights that helped Polyborg improve its lead targeting. And it delivered these outcomes at roughly 20x lower cost. Full results, plus what we're improving next: familiar.io/blog/familiar-vs…
4
129
Action + sales solve everything.
1
6
118
It was clear from my first conversations with the @BMWiVentures team that we saw voice AI the same way. For it to be widely adopted, talking to it has to feel effortless and familiar, like talking to a person. It's rare to find investors who see the interface itself as the frontier. Welcome aboard, and excited to build with you all!
Replying to @BMWiVentures
@BMWiVentures is backing Familiar. What drew us together is a shared belief that the quality of a product is the experience of using it. BMW has built a century of engineering on that idea. It's the same standard we hold for voice interactions with AI. They should feel effortless and familiar, like talking to a person. That's what we're building: a full-duplex speech model that adapts to how people already speak. It listens while it speaks, and understands not just what was said but how. Live on revenue calls today. As Kasper Sage, Managing Partner at BMW i Ventures, put it: "Full duplex is the difference between a demo and a voice agent you can trust in high-stakes workflows." Read announcement below.
2
1
10
349
Talking to AI should feel familiar, like talking to a person. It's the future we've been building toward. From today, our name says it too.
MetaVoice is now Familiar. Voice AI is growing fast, but most interactions still feel like a walkie-talkie. That's why voice agents today are largely limited to scripted, QA-style conversations. For voice AI to move beyond that, it has to adapt to the way people already speak. Not the other way around. Talking to AI should feel familiar, like talking to another person. That is the standard we named the company after. At the core of Familiar is a full-duplex speech model built to make that possible. It listens while it speaks, for a continuous, two-way conversation. More on the rename and where we're going: familiar.io/blog/metavoice-i…
1
6
277
I was an early engineer @wayve_ai , building end-to-end AI for self-driving. The lesson: hand-built pipelines lose to models that learn the whole task. Voice AI is at the same moment. Every voice agent today is a pipeline: transcribe → think → speak, with a separate component guessing when you're done talking. This turns the conversation into a walkie-talkie exchange, people get frustrated and hang up. So at @metavoiceio we're building the other thing: one duplex model that listens while it speaks. Calls feel natural, like talking to a person, even through interruptions, overlaps and background voices. I spoke with Theo @TheBreakAI about my journey, why we're building MetaVoice, how it differs from today’s turn-based solutions, and why revenue calls are the first use case. youtube.com/watch?v=dy2GTNHZ…
1
5
7
648
Announcing @metavoiceio's duplex speech model purpose-built for revenue calls! Talk to it at metavoice.io
Today we’re launching Mia & Leo: AI personalities backed by purpose built duplex speech models for revenue calls. It listens while it speaks so voice agents can handle interruptions, overlap, and background speech without turning the call into a walkie-talkie conversation.
2
1
18
1,956
I second that. I start all long complex tasks with a dense, detailed explanation. I share as much context as possible through voice notes and attachments, without worrying about typos. LLMs are great at interpreting noisy input to infer the overall intent.
One pattern I find useful for working with LLMs is a nice long ramble session. Sometimes the LLM needs more bits to understand what you're trying to achieve, but you're too lazy to type them. In these cases I like to lean back, switch to /voice and just ramble for like 10 minutes, total mess, anything goes, full stream of consciousness. Sometimes I declare it up top, something like "switching to speech recognition sorry for any typos...". Sometimes I turn it into a small interview of a few turns. But I find that the LLMs are somehow very good at reconstructing long incoherent rambles and often their echo of your own tangle of thoughts comes out quite a bit cleaner than what you started with. The result is that you improve the mind meld and have to correct things less from that point on.
1
228
Natural conversation maxxing! A sneak peek of @metavoiceio's duplex speech AI. Launching in 2 weeks.
4
5
29
1,889
Just moved to SF from London. Excited to reveal @metavoiceio to everyone soon! PS: skipping the customary photo next to golden gate bridge.
2
11
1,471
Accidentally said "I write code" instead of "vibe code" and they kicked me out of SF
Accidentally said "how are you" instead of "what are you building" and they kicked me out of SF
1
2
1,280
Siddharth Sharma retweeted
1/ Our belief @metavoiceio is that voice interactions with AI don’t go far enough. What if talking to AI felt indistinguishable from talking to a person, even after years of daily interactions? Fluid back and forth, empathetic responses and a consistent personality.
2
2
1
3
1,888
Why doesn’t Chrome have a (toggleable) side panel with Gemini always available? Would save me the trip to chat.com and back.
1
229
Siddharth Sharma retweeted
Big news: @wayve_ai is partnering with @Uber @dkhos @andrewgordonmac to launch the first-ever public road trials of Level 4 fully driverless vehicles in London 🇬🇧🚗 London is our home and one of the most diverse cities - where I can't wait to bring the benefits of autonomy!
25
42
16
375
52,603
Siddharth Sharma retweeted
AGI = All Ghibli Images?
52
157
21
1,594
120,939