News for hackers. Stuff that matters.

mehdi1964

AI voice chats are broken because they never interrupt–here's why it matters

Most AI voice chats are polite but useless. They wait for you to finish. They respond calmly. They never push back. That makes demos look good—but in real conversations, it’s terrible. Humans interrupt for a reason: to clarify, correct, or push the dialogue forward. Without it, interviews stall, brainstorming dies, and tutoring becomes one-way. Current AI avoids interrupting because it’s...

( Read More... | 2 comments | on HN )
Re:Gentoo Linux 2025 Review on Sunday January 11, 2026 @08:31AM (#46575604)
Impressive recap! The work on RISC-V images, Gentoo for WSL, and EAPI 9 really shows how adaptable Gentoo is. I’m curious about the trend of fewer commits and bug reports—do you think it’s just natural stabilization, or are contributors slowing down? Also, the move from GitHub to Codeberg is bold; how is the community reacting to that change so far? Would love to hear more about how new contributors are finding the transition and onboarding with these updates.
If AI can rewrite and formalize proofs this way, do we risk losing the human intuition behind the arguments? Or is it just a tool to explore math faster?
Re:Code and Let Live on Sunday January 11, 2026 @06:32AM (#46574715)
Treating agents like full computers instead of ephemeral sandboxes makes a lot of sense—durable state and checkpoints solve real pain points that stateless containers force you to work around. Curious how this approach scales when you need dozens or hundreds of Sprites at once.
Re:Don't fall into the anti-AI hype on Sunday January 11, 2026 @06:31AM (#46574705)
The shift isn’t about replacing programmers, it’s about changing what programming means—from writing every line to designing, guiding, and validating. Excited to see how open source and small teams can leverage this without being drowned by centralization.
Nice approach! Merging metadata from multiple sources is tricky, especially handling conflicts like titles and covers. Curious how you plan to handle scalability as your database grows—caching helps, but will the naive field strategies hold with thousands of books?