The developments in AI field are still very interesting. We are so close to the situation where the use cases explode. At the moment much of what we see is still rather niche and on advanced tech demo stage. Once the integrations to different parts of our daily life happens, then it might feel like a Horn of Abundance. Especially so when AI itself helps in building the integrations between the different use cases, agents, systems.
I still remain to be convinced that ANY of these demos will operate realistically in the real world. Connection issues and battery life will still be a problem for many, and the kind of contextual demos we see here are carefully chosen to have well rounded variables, which generally don't exist in real life. Neither Google nor OpenAI have said anything about how they will prevent the AI from hallucinating. Before I believe this can work as shown I will need to see an AI response that says simply "I don't have enough data to give a reliable answer", which is probably a lot harder than it seems.