I don't know why you are being condescending to Julio. His work has pushed AV1 forward significantly. His opinions (specifically about dir-pred) are shared by others who are in the encoding space because they're objectively true. If a breakthrough for splines is invented, we'd be elated, but as it stands, dir-pred is in virtually every video/image coding standard since the late 90s for a VERY good reason. Dir-pred also has two decades of iterative improvement and research behind it. The argument is that JXL's lack of this well-established coding tool is a serious handicap that will make it very hard - if not impossible - to compete with AVIF within a reasonable amount of time.
AVIF was performing better than JXL/WebP/JPEG for a long while, even before Julio and Gianni's tune IQ work happened. AVIF successfully met the bar, tune IQ further increased that efficiency gap.
AV2 needs optimized encoders and decoders because it now needs to justify its spot among AV1 and JXL on the web. That's why avm, svt-av1 and dav2d are well underway.
AVIF was performing better than JXL/WebP/JPEG for a long while, even before Julio and Gianni's tune IQ work happened. AVIF successfully met the bar, tune IQ further increased that efficiency gap.
AV2 needs optimized encoders and decoders because it now needs to justify its spot among AV1 and JXL on the web. That's why avm, svt-av1 and dav2d are well underway.