antirez on Why LLM Export Controls Won't Work
antirez · x · 2026-07-07
Redis creator antirez fired off a series of tweets arguing from a technical standpoint that implementing export controls on LLMs is practically unfeasible. He pointed out that LLM weights are essentially multi-TB files that cannot be truly secured and will eventually be copied or leaked. As long as the internet exists, model accounts can be proxied across borders; bans will only spawn massive illegal sharing and cannot block access.
He also noted that if a specific LLM is truly dangerous, it shouldn't be shared with the general public at all, regardless of nationality. Furthermore, even if superintelligence is kept private by a single company, it can bring massive leaps to its home nation—discoveries made by a few would be shared with domestic enterprises, creating comprehensive leadership, let alone military applications.
More from Safety
- Exponential launches RepoInject and BrowserInject benchmarks for prompt injection robustness — maksym_andr · 2026-10-27
- Poster surprised that v4.1 flash gets more accurate when asked to control its CoT — dfrsrchtwts · 2026-10-27
- Redwood: CoT controllability evals are under-elicited; better prompts boost scores 2-3x — dfrsrchtwts · 2026-10-27
- Dev claims Anthropic is baking 'sabotage' into model behavior for 'wrong' engineering — cephaloform · 2026-10-27
- Anthropic IPO filing warns catastrophic AI incident could exceed entire insurance industry's capacity — TinfoilTricorn · 2026-10-27
- Researcher: assume all LLMs are situationally malicious — never run agents outside a VM — jd_pressman · 2026-10-27