@MIT alum | Formerly @NASA @NASAJPL | Current CEO & Co-founder of @Modulate_ai
Joined August 2015
- Tweets127
- Following102
- Followers38
- Likes45
Pinned Tweet
I used to teach spacecraft to see.
Now I teach software to listen.
@modulate_ai recently raised $25M, led by Future Ventures, to build the listening layer for AI.
Cosmology → JPL → voice. Here’s the path 🧵
Coolest thing someone’s built with our tech: a prototype that listens for early warning signs of dementia.
It checks 4 things:
→ Sounds confused
→ Then frustrated
→ Can’t recall a name
→ …of something most people wouldn’t forget
The first two live in the voice, not the words.
“Is this really you?” is about to become a normal question on phone calls.
Voices can be cloned now. A familiar voice isn’t proof of identity anymore.
I’m on a panel at VapiCon 2026 about what it actually takes to defend against voice fraud.
Code HUFFMAN20 = 20% off your ticket.
Reply with the link
Full agenda: vapi.ai/vapicon-2026
#VapiCon
In case you missed our @modulate_ai @LinkedIn Live this week, you can check out the replay below⬇️
tinyurl.com/52bxyr3s
William and I had a fantastic conversation on how to spot synthetic voices as the call happens, how to route bots and humans where they belong and how to take it from prototype to production!🤖
#VoiceAI #AgenticAI #ContactCenters
AI agent: “Is there anything else I can help you with?”
You: “No, I’m good, thanks!”
AI agent: “Great! Is there any more information I can give you?”
You: “No, really, I’m all set.”
AI agent: “Wonderful! Is there anything else…”
Rude people hang up. Polite people get trapped.
The agent doesn’t know it’s stuck in a loop. Someone has to be listening for it.
One of the coolest thing someone’s built with our tech:
A prototype that listens for early warning signs of dementia.
Not “forgot a word once.” That happens to everyone. (Me, constantly.)
It’s the pattern: confusion, then frustration, increasing over time.
That signal lives in the audio. Not the transcript.
Going live in a few hours 🚨
Your next inbound call might not be a person.
William and I are demoing real-time synthetic voice detection: spot the bot mid-call, then route it where it belongs.
Bots one way. Humans the other.
Join us: linkedin.com/events/75190943…
Carter Huffman retweeted
~$0.20: a 6-minute voice AI interaction.
~$2–5: a human-handled call.*
So why are AI callers still reaching your humans?
Modulate scores whether a caller is likely human or AI from as little as 2.5s of audio. Humans → humans. Bots → bots.
*Modulate analysis, varies by operating model
Most deepfake defenses only screen the first few seconds of a call.
Fraudsters know this.
Pass the check up front, and the rest of the call goes unwatched.
If you’re not listening to the whole call, you’re not really listening.
New series: Velma in the Wild.
We take famous scenes, run Velma on the raw audio, and watch what it hears. No script. No transcript. Just the sound.
Episode 2: A Few Good Men, the courtroom scene.
Everyone remembers “You can’t handle the truth!”
But the shout isn’t the real moment. The confession that comes after is.
Watch what Velma flags on the way there.
Words give you the script. Voice gives you the performance.
Most of AI is betting on one bigger model.
We went the opposite way: 100+ specialized models.
Different models listen for different signals. The words. The emotion. Cadence. Pronunciation. Whether your audio is degrading mid-call.
Then they combine into something that actually understands the conversation.
Your voice carries far more signal than most people think.
Thought experiment: early 2020. What if a phone call could flag a likely COVID case?
Could we have seen the spread sooner?
Huge caveats. But think about the lives that might have saved.
As a first time CEO, it’s pretty awesome to watch @Google’s former CEO @ericschmidt talk about leadership and #AI.
Leadership doesn’t happen by accident. It’s a privilege and with that privilege comes responsibility.
The question isn’t just what you can accomplish with your position, power and influence. It’s what you choose to do with them and who you lift up along the way!🙌
Another powerful reminder from the @mastersofscale Summit stage.
#MoSSummit @modulate_ai
Human connection matters!! And it shouldn’t be an afterthought. Perhaps my favorite recurring theme at the @mastersofscale Summit in #SanFrancisco: Keep humans at the center. Always.
As #AI becomes more capable, human agency, creativity and judgment become even more important.
Especially when the stakes are high, humans MUST remain in the loop.
The goal isn’t to remove people from the equation. It’s to empower them.
Technology should amplify human potential, not diminish it.
What do you think?🤔
#MoSSummit @modulate_ai
Another recurring theme at the @mastersofscale Summit: #AI is about choices, not inevitabilities.❗️
Almost all sessions so far have asked a version of these questions⬇️
What happens to our cognitive abilities when we surrender more of our thinking to AI? Do they atrophy? How do we prevent that?
What do we choose to surrender to technology…and what do we protect?🤔
Do we use AI to augment human potential or replace it? For example: Do we dismantle education as we know it, or reimagine it for the better?
These aren’t predetermined outcomes. They’re choices.
We have agency. We get to decide. And those decisions matter.
#MoSSummit @IuliaNandrea @modulate_ai
.@karaswisher and @netflix’s co-founder @reedhastings cracking jokes on the @mastersofscale Summit stage and reliving Netflix’s startup roots and how touch and go that journey can be.🙃
Love hearing what Reed has learned over the years and it’s pretty simple:
💭 Startups take a lot of imagination. Be creative. Be bold.
⏩ When you’re in a climate when things are constantly changing, you have to be willing to bet forward.
‼️ If you’re willing to be radical, you have to get some stuff wrong. And be comfortable with that.
#MoSSummit @modulate_ai
