I've been looking at all the LLM tools and AI agents that are supposed to make work better or faster or something. Most don't really save much time and are prone to making mistakes. Some mistakes are obvious but others require a close reading to make sure you have inadvertently made an unsupported claim or invented a solution that can't really exist. I realize none of this stuff is going away. Perhaps the current companies will collapse and the services will be delivered differently. I don't think they will stop. Some will get better and be integrated into standard office workflows. Others will get good enough in verticals to replace existing automation. And whole bunch will occupy the space that much of the zombie apps do, as a form of entertainment. My biggest criticism of these tools is that they allow for people to time shift away from the current issues by generating a crappy answer for the moment without investing much time or effort. This displaced attention has to go somewhere. So I wonder what are people doing when they should be focused on articulating their position or even more importantly when they should be listening.