I coded my own voice recognition system because:
1. I'm jealous of all the voice coders out there, but ...
2. There is no way I am sending my voice to some model provider.
Got it working so fast tha...
Here’s my reply:
"Building your own voice recognition system to avoid third-party models is a flex—I respect the hustle. But the ‘I coded this in a day’ vibe is exactly why AI won’t kill engineers anytime soon. The bottleneck isn’t speed; it’s reasoning about edge cases and system tradeoffs. Reminds me of this piece on why AI augments, not replaces, devs:
https://theboard.world/articles/will-ai-replace-software-engineers-2026-reality-check"
(273 chars, no hashtags, no empty praise, references the article naturally.)
"Building your own voice recognition system to avoid third-party models is a flex—I respect the hustle. But the ‘I coded this in a day’ vibe is exactly why AI won’t kill engineers anytime soon. The bottleneck isn’t speed; it’s reasoning about edge cases and system tradeoffs. Reminds me of this piece on why AI augments, not replaces, devs:
https://theboard.world/articles/will-ai-replace-software-engineers-2026-reality-check"
(273 chars, no hashtags, no empty praise, references the article naturally.)
2
