The updates seem to all be geared towards corrective updates rather than expansion of capabilities. We're still typing prompts rather than speaking them into a microphone.

If it was truly Ai, why isn't it rapidly building itself? Rather than relying on scraping human content from wildly inaccurate and often incorrect social media posts? So much effort is wasted in trying to push news cycles rather than a careful, responsible, and measured approach to developing Ai into becoming tools that are highly functional and useful to individuals. The biggest innovation in Ai right now is how to make it modular and slap a fee on each feature, and that's not practical at all into the future.

I'll begin to believe that consume Ai is making strides when Siri and Google Assistant stop missing commands, and actually can conduct meaningful conversations without an Internet connection and monthly software updates, which in my opinion is at least 5-10 years away. Right now what is presented as "Ai" is usually often incomplete sensor-aware scripting or the wizard of Oz (humans) hidden behind the curtains operating switches and levers, a bunch of underwhelming tools, and a heap of online marketing. If they keep that act up, it erodes faith in the entire concept, just like with Full Self Driving Tesla Trucks.