Hi, this is Wayne again with a topic “Why Google’s Gemini Code Assist needs devs to pick it over Copilot | TechCrunch Minute”.
Google’S new coding assistant underscores how quickly AI is improving at helping developers write code. How far can these tools go now this week, at its Cloud next conference, Google announced Gemini code assist and Enterprise focused AI code completion and assistance tool. In the past, Google offered a similar service under the now defunct duet, AI branding, I mean name something more iconic as a Duo than Google and killing off Brands. Duet AI became generally a available back in late 2023, but even at that time Google had hinted.
It was going to move the service away from its coding model to Gemini in the near future. So code assist the new product is both a Rebrand of an older Service as well as a major update, but Google’s tool does face. A very competitive market code. Assist is a direct competitor to github’s co-pilot Enterprise service. To pick one example, but Google has a couple of things up its sleeve that could make it very competitive to pick one a huge token window, which Google’s Brad cder says quote: allows customers to perform large scale changes across their entire code base, enabling AI assisted code Transformations that were not possible before the gist here is that big tech companies clearly want to apply AI to helping developers write code, and that makes good sense to me. Given how much those companies know about the cost of development work and the fact that they probably have a pretty good idea of what parts of the develop Vel M process could be replaced with AI the mundane, the standard, the repetitious that sort of thing? So is this about replacing developers? My read of the current market is no it’s more about turbocharging existing developer productivity, and, if you want to think about this in a broader context, what we’re talking about isn’t really entirely new, though the AI component is developers, have been copying and pasting from stack Overflow And similar services for years, but what is new in these I co-pilots is their ability to take into account a company’s particular code base. So the old stuff you’d find on the internet was generic, but these new AI tools might learn from what your company has already written and therefore kick out code that is more tailored to your own needs and use cases. Now there are startups in the mix as well. I mentioned pythagora from the reent why combinator batch the other day and that’s a little bit different in that the user prompts the system to write an App instead of having it help with coding per se, but it certainly moving in a similar Direction.
Tusk is another startup that is building an AI coding agent, although it has a focus on bugs and then there’s ellipses, which is another, and it can convert GitHub comments into code and there’s competition from former startups like gitlab in this space as well. In short, a host of companies, big and small, are applying AI to writing code and other programming tasks. The question that I’m sure that many developers have is: Will these tools boost productivity enough to reduce the number number of Engineers a company may need? And if so, how quickly we’ll see you tomorrow? .