These are not particularly good arguments he puts forth here
Would you care to explain why you think they're not particularly good? You're blaming DHH of being "hand wavy" but I don't see any more depth from your comment.
But to specifically address DHH's arguments from the article:
* Enabling Discovery is a strawman argument; no one expects an API to use / as its only documentation (at least no one expects it of a good API).
* Standardizing API Clients is also a strawman; no one expects to have one generic client magically make sense of any API.
* Comparing HAL to WS-* in order to paint it as committee-driven standards bloat is not fair to HAL, which is an admirably tight and cogent specification.
But to specifically address DHH's arguments from the article:
* Enabling Discovery is a strawman argument; no one expects an API to use / as its only documentation (at least no one expects it of a good API).
* Standardizing API Clients is also a strawman; no one expects to have one generic client magically make sense of any API.
* Comparing HAL to WS-* in order to paint it as committee-driven standards bloat is not fair to HAL, which is an admirably tight and cogent specification.