The answer to your big question is: It depends.
As an example, it may cost the same to hire either of two consultants, but equal cost does not guarantee equal results.
Whoever is using the agent needs the ability to interact with it in a way to get the most out of it. Professional ability differs whether using an agent or not. This is one of the reasons I suggest the agent be used as the pair partner to the professional. Even if I’m the one making all of the changes, it can often be faster to ask the agent, “Where do I find the code that does X?” than me doing that research myself. This can be especially true if you are not yet familiar with the codebase. I appreciate that time savings. As a further example of that, take a look at deepwiki.com and what it has done documenting some public repos.
How time and/or money are spent is, of course, a personal choice whether for an individual or a business. My usage was two-fold. I wanted to learn more about AI, agents, and the use of them. So far, I’ve gotten a good start on that. The output of that learning was also the time savings of starting my side project. This was not something I would hire a team to create. Perhaps a startup might. But, it did help me learn how to use both AI and agents more effectively. Whether that potential improved efficiency matters is up to the person or company paying for it. My only regret so far is not making that one extra commit before an agent went off the rails. (See the ‘write a minivan’ comment from before.). Overall, I like the results I have achieved so far and will continue to use it.
