Bindu Reddy says open source is the safest path for AI, not bans

bindureddy · x · 2026-07-26

Bindu Reddy argues that the safest path for AI is open source: transparency reduces abuse, while banning open models concentrates power in monopolies and authoritarian governments.

He says the AI safety crowd has repeatedly exaggerated open-source risks, citing examples like GPT-3 being labeled too dangerous, alleged government bans on open models, and alarm over agent social networks that later faded from attention. His conclusion is that closed approaches create a more dystopian outcome than open ones.

Original post →

More from AGI Musings

AGI Musings channel →