I am reading opinions here from agent users, but I haven't adopted the "agentic workflow" myself because I believe I am (for now) now getting a lot of my trouble's worth using Gemini (3 Pro) in the traditional conversational manner. It is adequate at suggesting solutions in the form of code, or reasoning in general. My problems are software engineering but also everything that is not, since I have a subscription it's my go to problem solving partner. I see no reasons to switch to another product for now either, I am constantly in the loop getting samples of chats with Grok and ChatGPT and it seems a very close race. If Claude is that one race horse that's built different -- and I absolutely can believe it is so because they have rightfully tuned it -- I am not convinced I am missing out much. But maybe because I am more traditionalist to most of everyone's having embraced the idea of having an agent run a loop on their workstation(s) and trusting it to deliver. Perhaps if I were in more of a tight time frame, I'd be pressed to do so myself, but for now I am already benefiting from the extra speed "rubberducking" with Gemini all manner of software engineering problems that I need to solve, so I simply have no reasons to abandon it. I think this is also Google's strength -- they have the data, they've already integrated Gemini or a variant of it anyway, into google.com which is one of their prized cash cows, and it's everywhere else too. Like others here have said, Google may not have the absolute best in class at all times, but they're fairly good and they still have the brains that gave us DeepMind and GPT, unless there's some sort of stagnation going on in their ranks, I expect they're not resting on the laurels. With their capital they're still at the head of the race. Anthropic and OpenAI have the benefit of being nimble, though, and it shows too. Anyway, competition is good, the cat's out of the bag and on the greener side of the river :-)
No comments yet.