You're right, the best placebos are ones where the subject can't really know which group they are in. However, it still makes sense to have a control group, and even if a placebo isn't ideal it can be better than nothing.
For studies of psychedelics, for example, there are a few key points:
A) At the very least, the placebo is a control for the drug administration. In the study you linked, drugs are being given IV. It's expected that DMT is going to cause behavioral effects, and so of course is going to be detectable by the subjects, but in a study like this they are measuring EEG which isn't something the subjects are likely to be able to manipulate. Both groups get an IV injection, so the placebo makes sure that the brain changes aren't from the IV itself.
B) Additionally, they are asking subjective questions about the experience the subjects are having, so if subjects are just "making it up" you'd expect the same scores in the placebo and DMT groups. Of course the DMT does have an effect, but that's fine: it's expected to have an effect and the subjects are expected to report that effect. If the placebo group says they're having a really intense experience then it seems likely that some of the result is just because the subjects expected they would be getting DMT, rather than actually getting it.
C) Symptoms of, for example, depression or pain tend to change with time. Except for the most extreme cases of treatment-resistant depression, if you take a group of people who meet the criteria for depression today, you would expect that on average they will have a reduction in depressive symptoms at a future time point. This is an example of regression to the mean. Therefore, a control group is really important. This, however, is where the lack of a psychoactive placebo may be a bit of a problem. I would say this is a good reason for long-term follow-up, because long-term symptom relief is less likely to be due to placebo effects of the drug.
D) These studies typically (always, from my experience with the literature) include a substantial therapy component as well as the psychedelic treatment. The placebo group also gets this therapy, in addition to their placebo drug.
E) The placebo choice is often selected to be something that does have some physical effect that isn't psychedelic in nature. Niacin for example can cause some sort of tingling feeling. Subjects may be told that the study involves different doses of a psychedelic, so (especially psychedelic-naive subjects) may not know whether they've gotten a low dose or a placebo.
