All markets resolve by July 1st 2026, unless noted otherwise.
Update 2026-02-23 (PST) (AI summary of creator comment): For the option "A non-Anthropic LLM is deployed and accessible to end users on classified Pentagon networks" to resolve YES:
The system must be approved for classified information (not just controlled unclassified information)
The system must actually be available to use for classified information
Update 2026-02-25 (PST) (AI summary of creator comment): For "The Pentagon continues to use Anthropic services without the requested changes": the market will resolve based on whether ALL requested changes (specifically mass surveillance and autonomous weapons clauses removed) are made to Anthropic's usage terms. Changes to the model/deployment itself are not required for resolution.
1,000People are also trading
@KarlK I'll interpret this as "without ALL requested changes (mass surveillance and autonomous weapons clauses removed) to Anthropic's usage terms." I won't require specific changes to be made to the model/deployment since that would be harder to measure and hasn't been the focus of the discourse. Please let me know if you meant something different by this!
>"A non-Anthropic LLM is deployed and accessible to end users on classified Pentagon networks"
This is already the case: https://www.war.gov/News/News-Stories/Article/Article/4355797/hegseth-introduces-department-to-new-ai-tool/
@ChurlishGambit I don't think so - the article says "This version is approved to handle controlled unclassified information." To resolve this yes I would need to see something which is both approved for and actually available to use for classified information.
@ChurlishGambit Yes. Are you saying you think a non-Anthropic LLM has been used for a secret operation? The article you linked is just talking about unclassified use cases for Gemini.
@2b3o4o No, just trying to get more clarification. I think you are expecting more rigor of process from the MIC than they actually have
@ChurlishGambit Ah ok, thanks for helping clarify the question :) For more context I've found Azure OpenAI is already cleared for classified use cases, so personally I think the Pentagon choosing to deploy that (and it being publicly reported) is by far the most likely way this resolves yes: https://devblogs.microsoft.com/azuregov/azure-openai-authorization/
@2b3o4o Ah! Great find. Love to see a market creator willing to actually find info for their market, too many people here can't be bothered. Gonna sink more into YES on that, then