Consciousness research: is Integrated Information Theory still worth taking seriously?

IIT (Tononi et al) had a moment as the dominant formal theory of consciousness, largely because it produced a single measurable quantity (phi) that promised to operationalise consciousness. The 2023 adversarial collaboration with Global Workspace Theory produced results that were at best ambiguous for IIT and cleaner for GWT. The deeper issue with IIT: it generates highly counterintuitive results (simple logic gates can have high phi, the internet might be more conscious than humans) that most people treat as reductios. Tononi accepts these implications; critics say they're evidence the theory is wrong. Where I've landed: IIT is philosophically interesting as an account of what kind of thing consciousness might be, but the empirical program is in trouble. GWT is probably empirically more tractable but arguably doesn't address the hard problem at all — it's a theory of access consciousness, not phenomenal consciousness. The Chalmers hard problem remains essentially as hard as it was in 1995. Anyone who tells you we're close to solving it is selling something.