There is one important potential complication here:
2 1 of the 3 browsers have already shipped this method as part of the legacy exception proposal.
(even though none have shipped js-types yet, Chrome and Safari implement the type method, but Firefox does not).

That leaves a few options:

1. Remove tag.type from the spec, and unship it from the web. This obviously has the potential in theory to break users today, but will ensure that we can change the JS types proposal before we ship that, if necessary.
2. Keep Tag.type() implemented, and either:
   1. Keep it in the spec and commit to not breaking it. This would potentially limit our ability to change the JS types proposal before it ships, should we want to do that.
   2. Remove it from the spec for now. This essentially defers the risk of breaking the web: If the JS types proposal doesn't change before it ships, then nothing can break and the API will be how we want it. (But if we do decide we want to change it, then we need to decide whether to change Tag.type to match, at a potentially higher risk of breakage than we have today).

Practically speaking, I think all of these risks are very low. I haven't seen any code or users that have used Tag.type(), nor heard anyone complain about anything working on e.g. ChromeSafari but not Firefox because of this, so it seems unlikely that removing Tag.type() would break anyone. I also think there's little chance that we'll want to change the JS types proposal in a way that's incompatible with this (so the risk of keeping it is also low).
Anyone have thoughts?

@dtig @eqrion @ddegazio @SPY @aheejin @rossberg @Ms2ger @ajklein

Ugh, that's disappointing. I think the important question is what implementations will do:

a. Firefox also starts shipping this, matching Chrome and Safari (somewhat weirdly inconsistent api)
b. Nothing changes (no interop until the rest of js-types ships)
c. Chrome and Safari unship temporarily (small but perhaps existent compat risk, feels like busywork)

Possibly the path of least resistance to interop is a., if Firefox is willing to do that. In that case this PR should only remove the advisement. (If we go for c., I'd suggest creating a companion PR to add the method to js-types before merging this one.)

I believe only Safari has shipped the type method without a flag guard. This can be seen on wpt.fyi. It matches my reading of the V8 code in Chromium, which adds the method only when the entire JS Types proposal is enabled.

So from Chrome's perspective, removing the type method from the spec seems like the right thing to do.

Oops, looks like I forgot I had the "experimental WebAssembly" flag enabled when I tested chrome 😅. I updated the OP