Hacker News: underfox

What we need to make voice AI agentic

underfox — Tue, 03 Mar 2026 19:08:20 +0000

Article URL: https://www.ultravox.ai/blog/what-we-need-to-make-voice-ai-fully-agentic

Comments URL: https://news.ycombinator.com/item?id=47237186

Points: 1

# Comments: 0

Tackling the problem of "naturalness" in voice AI [video]

underfox — Fri, 13 Feb 2026 22:55:49 +0000

Article URL: https://www.youtube.com/watch?v=7zIEC8tmWkA

Comments URL: https://news.ycombinator.com/item?id=47008965

Points: 2

# Comments: 1

Ultravox's Breakthrough Voice AI Benchmark [video]

underfox — Thu, 05 Feb 2026 18:15:16 +0000

Article URL: https://www.youtube.com/watch?v=Z7l4w9RTQ_0

Comments URL: https://news.ycombinator.com/item?id=46902751

Points: 1

# Comments: 0

Why speech-to-speech is the future for AI voice agents: Unpacking the AIEWF Eval

underfox — Tue, 03 Feb 2026 18:44:24 +0000

Article URL: https://www.ultravox.ai/blog/why-speech-to-speech-is-the-future-for-ai-voice-agents-unpacking-the-aiewf-eval

Comments URL: https://news.ycombinator.com/item?id=46875296

Points: 3

# Comments: 0

The State of Voice AI Instruction Following in 2026

underfox — Thu, 29 Jan 2026 18:29:29 +0000

Article URL: https://www.coval.dev/blog/the-state-of-voice-ai-instruction-following-in-2026-a-conversation-with-kwindla-from-pipecat-and-zach-from-ultravox

Comments URL: https://news.ycombinator.com/item?id=46814299

Points: 4

# Comments: 0

New comment by underfox in "Our approach to advertising"

underfox — Fri, 16 Jan 2026 19:53:16 +0000

Obviously disappointing, but not entirely shocking given how much capital they've already burned through. Convincing individual users to pay $8/mo was never going to even out the balance sheet.

Ultravox Realtime is now available as a speech-to-speech service in Pipecat

underfox — Fri, 16 Jan 2026 18:00:59 +0000

Article URL: https://www.ultravox.ai/blog2026/introducing-the-ultravox-integration-for-pipecat

Comments URL: https://news.ycombinator.com/item?id=46649560

Points: 1

# Comments: 0

New comment by underfox in "History LLMs: Models trained exclusively on pre-1913 texts"

underfox — Fri, 19 Dec 2025 21:19:08 +0000

> [They aren't] perfect mirrors of "public opinion" (they represent published text, which skews educated and toward dominant viewpoints)

Really good point that I don't think I would've considered on my own. Easy to take for granted how easy it is to share information (for better or worse) now, but pre-1913 there were far more structural and societal barriers to doing the same.