←back to thread

Sanders: Government should break up OpenAI

(thehill.com)
46 points CharlesW | 1 comments | 01 Nov 25 19:40 UTC | HN request time: 0s | source
Show context
whatpeoplewant ◴[01 Nov 25 20:44 UTC] No.45785166[source]▶
Breaking up OpenAI might tackle concentration, but bigger leverage is interoperability and transparency across compute, model, and orchestration layers so no single vendor controls agentic LLM workflows. As the field shifts toward multi-agent, distributed agentic AI (including parallel agentic AI), policy should prioritize open agent protocols, auditable sandboxes, and portable evals/policies. That preserves competition while addressing system-level risks (emergent behavior, cascading failures) that arise specifically in agentic AI more than in single-model deployments.
replies(2): >>45785211 #>>45785261 #
1. nradov ◴[01 Nov 25 20:55 UTC] No.45785261[source]▶
That might make sense to mandate years from now but the industry is still evolving way too rapidly for standards based interoperability to be practical. Let's maybe take a look at that in 10 years.