Hmmm, let me try to clear some of this up.
"I do not think that your argument is sound. You assume in your first paragraph that agents have 'goals, intentions, and purpose' which, I believe, is the end point of the question as asked."
As I outline in the essay, the question is not whether agents can be described as having goals, intentions, and purpose. You do that every day, I do that every day, every human describes agents this way throughout their lives. Many scientific and quantitative fields also describes agents in this way, and even have associated mathematics and formalisms, as I point out. The real question is not whether some physical systems (agents) can be described as having goals, intentions, and purposes, because clearly they can, but whether those things are actually causally efficacious. That is, are they epiphenomenal, or can they be reduced to the mindless mathematical laws and relations of the microscale?
"I do not think that supervenience is the right concept."
Supervenience means that if you fix the properties x at a lower scale you then fix y at a higher scale. Your example of baseball not being reducible to biology isn't capturing that. Obviously, given the set of fundamental physical properties that underlying a baseball game being played, whether or not someone is on first is fixed by those fundamental properties. Imagine fixing the fundamental makeup of a computer; that fixes the higher scale of the logic gates in the computer. That's supervenience. It sounds to me like you are talking about reduction between laws rather than supervenience between scales: do the laws of organic chemistry reduce to the laws of sub-atomic physics? This is actually a different question than: if you fix the set of fundamental physical properties in a system, do the higher level properties of that system follow? In fact, your very definition, that "lower levels implement the higher levels" implies supervenience (why else would implementing higher-level property y via lower-level properties x bring about y unless fixing x meant fixing y?). When talking about these issues it pays to be precise, and unfortunately special jargon is needed, and thus careful attention must be paid while reading to understand the definitions.
All the best,
EPH