Justin Uberti (@juberti) | nitter

Pinned Tweet

Nov 4

Just promoted our new GPT-Live outbound telephony API to GA, allowing developers to make phone calls from their Live voice agent, via any SIP trunk provider (e.g, Telnyx or Twilio). Calls can use HD Voice and can send and receive keypad events (DTMF).
developers.openai.com/api/docs/guide…

59

45

530

142,579

7h

ChatGPT supports shortcuts to start a chat, dictation, or call to your dot, which you can map to the iPhone action button for one-touch access.

4

2

32

2,224

Nov 6

Just finished Outer Wilds. Quite an interesting game, with a soundtrack I’ll probably never forget.

6

0

21

2,051

Nov 6

nice one

Chess Master
@xxChessMaster
Nov 5
Why is this not a checkmate?🤔😱

2

0

15

5,917

Nov 6

Dots calling was kinda slow, but we fixed it. It's pretty fast now, thanks to WARP and Instant Connect and some heroic refactors, and we think you'll enjoy it a lot more.

samir
@_samirism
Nov 4
we know people love calling their dots - connection times are now much faster.

also by popular demand - you can now hang up hands-free. just say “goodbye” to your dot and it’ll end the call.

26

4

184

8,683

Nov 5

Will be talking about the architecture for GPT-Live next week at VapiCon 2026!

Vapi
@Vapi_AI
Oct 28
On Day 1 of VapiCon, @juberti’s gives a talk on speech-to-speech models, “Skipping Turns: A More Responsive Architecture for Voice Agents.”

Justin, Head of Realtime AI at @OpenAI is a hands-on engineer and vision-based leader of large (150+) engineering teams. He is also an n, physicist, and musician.

Join Justin and other AI leaders in San Francisco this November.

📍 Fort Mason, San Francisco
📅 November 11–12, 2026

Explore the agenda and get your ticket: vapi.ai/vapicon-2026

2

2

43

2,060

Nov 5

Great question! Ideally the agent could know a direct interaction is possible before even making the call.

Amrutha Gujjar
@amruthagujjar
Nov 5
What happens when the outbound agent reaches another voice agent? Do they stay in speech + DTMF, or could the call negotiate a structured exchange?

5

0

7

1,856

Nov 5

Can I interest you in OpenAI Presence?
openai.com/business/opena…

Hasan Jilani
@mhasanjilani
Nov 5
The interesting question is where vertical integration stops. Telephony and session management are natural extensions, but orchestration increasingly touches business logic, external systems, application-specific policies, etc. Do you see the model provider eventually owning that layer too?

3

0

35

4,076

Nov 5

persistence pays off

Alex Kantrowitz
@Kantrowitz
Nov 2
New: OpenAI tells me 250 Million people talk with ChatGPT weekly (voice and dictation). About 20% of its user base.

7

0

61

3,940

Nov 5

The problem is that this stuff is very complicated and excruciatingly hard to debug. Vertical integration makes for fewer seams and helps ensure better quality for developers.

nutanc
@nutanc
Nov 5
So, does this mean OpenAI is going after the infrastructure layer(like LiveKit) and the orchestration layer(like pipecat).
So theoretically you can have OpenAI and the Sip trunk layer(like Ozonetel,Twilio etc) and that's all you need for the voice agent stack.
Interesting.

8

0

53

4,644

Nov 4

Updated the IETF WARP (WebRTC speedup) drafts:
- sped-01 adds explicit negotiation via the "sped" ICE option, and guidance on pacing and DTLS roles
datatracker.ietf.org/doc/draft-hanc…
- snap-01 clarifies a=sctp-init and allows INIT chunks w/ or w/o trailing padding
datatracker.ietf.org/doc/draft-hanc…

1

1

32

3,352

Nov 4

Replying to @juberti

snap-01 is backwards compatible, sped-01 is not (due to the ICE option in SDP, wire format is the same otherwise).

0

0

1

538

Oct 20

It's good to be cautious here, but these arguments are surprisingly hand-wavy:
1) intelligence doesn't imply consciousness
2) we think computational functionalism is "fatally flawed"
3) some theories require specific physical properties that LLMs lack

nature
@Nature
Oct 19
LLMs can mimic thought and feeling with uncanny skill, but intelligence alone is not enough for subjective experience

go.nature.com/3VSXWQ4

3

1

18

2,383

Justin Uberti retweeted

Oct 16

Dots, now on your Home Screen

112

33

991

371,266

Oct 16

Nonlinear activation functions can do some amazing things!

Cheng Lou
@_chenglou
Oct 15
My dear friends, I’d like to show you my latest toy project

Introducing Handwired: turn your regular JavaScript code into a neural net!
- no training
- exact outputs vs the JS program
- fully differentiable

All runnable in a browser, without GPU

1

1

11

1,935

Oct 16

Absolutely insane live demo. So many things that could have gone wrong. But they nailed it.

Tesla
@Tesla
Oct 16
Real life, baby

2

1

39

2,720

Oct 16

If your Live API app typically selects from a few specific tools, you can now use the new Decisions API to speed up delegation, while retaining the ability to use a full LLM for long-tail queries.

jiayu
@daidaijiayu
Oct 15
you can now use native decision delegation in your voice app.

a week ago i wrote this voice x decision doc. today i update the guide because now it is miles better

you can trigger a decision on every input or output frame (roughly every 200ms).

you can have a realtime decision engine judging what i’m saying while i’m still saying it!!
try realtime action × realtime voice 👇

developers.openai.com/api/docs/guide…

3

2

60

3,603

Oct 15

GPT-Live-1.1 offers much better multilingual support (including several new voices), more natural delegation, and improved steerability. Rolling out now to ChatGPT Voice, Desktop, and Dots.

OpenAI
@OpenAI
Oct 15
New voices. Better accents. And less "checking..."

GPT-Live-1.1 brings these updates to Voice in ChatGPT, Codex, and dots, with new voices in French, German, Hindi, Mexican Spanish, Korean, and Japanese.

Rolling out today to all users. Coming soon to the API.

15

2

140

6,792