I'm not really sure what the cultural and incentive context for scientists (whether CS, psych, or whatever) studying AI to always include a "to be sure" preface to their actual research. Almost nothing has been more depressing in my descent into pro-am AI Discourse than the _constant_ surrendering in advance