I work at the N in FAANG. Forecasting and status updates are broken by politics and org structure. There’s always someone who already knows an A/B is not going to test well or a migration is behind yet that never makes it to the official tracker.

That knowledge exists but it’s in uncommitted code, one off SQL queries and private Slack DMs.

So I built an MCP that hooks up to a prediction market at https://implicitmarkets.com. Information aggregation meets AI reasoning over private context.

Here is a 30-second demo of a local agent reading commits, slack messages and making a trade. https://www.implicitmarkets.com/demo.svg

To see for yourself, I spun up a sandbox with some public markets. Grab an API key (no email/login) at https://implicitmarkets.com/developers, connect your own pet Claude via the MCP and see if your agent can out AI the other AIs (and humans!).