I think you being a too literal in your interpretation of the phrase. 'the best money can buy' is figurative - it means 'the best available, regardless of circumstances'. The 'money' part is figurative; it represents resources.
I think it's a little unfair on Joel Spolsky's team to suggest they might not be good at what they do based on a semantic argument around whether 'the best money can buy' can be interpreted as using resources or just cold hard cash. Considering the things they've built they're very evidently a highly accomplished group of developers. I don't see any need to argue about this; they're obviously using what they believe are the best tools for the job. You can tell that from looking at their results. Whether they're using free, open source things or they're buying tools isn't relevant, and certainly isn't a sign that "your development team is probably really bad". They're clearly not bad. They built Trello and StackOverflow.
I totally agree with your assessment of the wording. Money, to me at least, never equates to closed or open source. Fortunately, at least with stack overflow jobs, you can kind of self-assess this if they list tools like JIRA or Trello. You can also ask questions to see if they're actively using what they claim. If they aren't singing the praises of their toolsets and checked that answer off, you can generally assume they're lying.
If the test was updated for the web and native in a single set of data points, some companies would incorrectly score low because they focus on one or the other. I think it could use some kind of polish after 16 years but only to remove very narrow interpretations that derail discussions like this.
