I'm pretty surprised China didn't hand Google the dataset itself, or provide an IsThoughtCrime API. Perhaps they're actively looking for what sites Baidu proactively censors, rather than simply what gets blocked at the filter level. There's no way Google could operate without the PRC's blessing, and since they are, the PRC has no reason not to help them out with the firewall. But they can't exactly ask Baidu for a cup of sugar..