const policy = pulumi.all([thing1.arn, thing2.arn]).apply(([arn1, arn2]) => {
    return ......
});