The power of AI technology to change code will seriously impact the IT industry — and everyone else.
Weeks ago, tech celebrities called for a pause on developing and deploying AI models more advanced than OpenAI’s GTP-4 — the stunningly higher-level algorithm behind ChatGPT — until risks such as loss of jobs and disinformation are better understood.
Even if OpenAI, Google, Microsoft, and other technology titans stop what they’re doing — and they won’t — the AI models that have already been developed are likely to have profound impacts, mainly in software development.
It might not seem like a normal business agreement, but Alphabet’s agreement to offer AI to Replit, a web-based coding tool with over 20 million users, is a turning point.
Replit will use Google’s AI models, among many others, in Ghostwriter, a tool that offers code and answers code-related queries in the same way as ChatGPT works.
Replit’s CEO, Amjad Masad, tells me that Google has “very exciting technology” that his brand can get into the hands of developers.
Google will also make Replit available to Google Cloud users as part of this collaboration, helping it to reach more business customers.
The move is interesting since Alphabet is fighting with Microsoft and GitHub, who use AI to help developers use Copilot, a Visual Studio add-on.
The same AI that makes ChatGPT seem so smart also works on computer languages.
As you begin entering code, tools such as Copilot will suggest a way to finish it.
Alphabet’s move also suggests what may be the next big battle for big technology companies.
While ChatGPT ring tricks and Midjourey 5 Donald Trump versions are getting a lot of attention, the main story is about whose company can offer developers the finest AI tools — and the new software that developers will build with that AI by their side.
According to Microsoft research, having an AI assistant can help developers complete jobs up to 50% speedier.
Companies that would provide cutting-edge AI can woo developers to their coding tools and keep them hooked on their clouds and other services.
Amazon has developed an AI coding tool called Code Whisperer, and Meta is creating one for internal use as well.
Apple will not want to be left behind.
AI is going to change the way software is generated as well as help developers write it.
This week, OpenAI announced the development of the first ChatGPT plugins.
These will allow the bot to perform tasks such as flight searches, making reservations, and buying groceries.
Implementing Ai into code can also help to speed up software development.
Masad of Replit shared an interesting example this week: an app that turns voice commands into working websites.
“We think that so many software projects will start in this manner in the future,” Masad says.
Because things are changing so quickly, it’s worth considering the risks of rapidly adding AI into software development.
AI tools can mimic problems in code that developers might not even discover or might not spot.
If developers depend too much on AI, they can become lazy or see their skills worsen.
And what type of “technical debt” can develop if programmers have to go back and modify software that has never been thoroughly inspected by humans?
We might not have to wait long to find out.
Top comments (0)