Replit Agents are Here to Replace All Software Engineers and their AI Tools

The Replit AI team has been cooking some agents. While the developer world is obsessing over Cursor and Claude, Replit Agents have taken the lead and replaced the need for even AI to write code. 

“AI is incredible at writing code. But that’s not enough to create software. You need to set up a dev environment, install packages, configure DB, and, if lucky, deploy,” Amjad Masad said, while announcing the launch of Replit Agent in early access, which is going to automate all of these processes.

The Replit Agent is an AI tool that helps users build software projects by interpreting natural language prompts. It simplifies software development, making it more accessible to users of varying skill levels. Currently, the agent is only available in Repls created via the Replit Agent entry and does not support existing Repls or imported repositories.

But apart from that, everyone is impressed. Even Andrej Karpathy, who has been actively building using Cursor, said Replit Agents can be placed under the “feel the AGI” category. 

“As mentioned in the post, making actual apps is a lot more than code, you have to set up the entire environment, deploy it, etc. Automating all of this other infra will allow anyone to quickly build and deploy entire web apps,” said Karpathy.

What All Can the Agents Do?

It was said that Cursor will be able to replace the software engineering team, or at least shrink them, and Claude Artifacts could kill the app stores, but Replit Agents can build almost from landing pages to healthcare apps connected with databases within seconds. This does not even require writing a single line of code.

Masad shared examples of people building a healthcare app in minutes where the agent could fix bugs by itself, a live postgres powered Flask and Vanilla Javascript based website in less than 10 minutes, and even a Wordle Clone in just 2:43 seconds.

Andrew Davison from Zapier was the first one to build a full working browser-based pong game in less than 90 seconds using Replit Agent. “I had to do zero coding myself. Just sat back and watched,” he said.

The most crucial use case of Replit Agent is for building MVPs of products within organisations, which would otherwise require a lot of time. “I like to say that AI makes coding fun again, by getting you past annoying boilerplate and API glues…saves me a day of work slogging through API docs every time I use it,” said Scott Kennedy, VP of engineering at Replit. 

Replit’s move is extremely timely with the conversation around Devin’s release and other agents like Cody.AI and ReactAgent coming up in the market. The only problem it currently has is not having a Python backend.

Agent for All?

The most interesting part about Replit’s approach is that AI coding agents are also available on smartphones, which Replit has always been famous for. It has always been the goal of Replit to make AI accessible for all with an active approach for open source. Sander Saar from RedBull said that he was able to build three functional web apps within 4 minutes on his phone. 

“Replit Agent doesn’t just review and write code, their AI agents plan the functions, create development environments, install dependencies, write the code, configure databases and deploy,” Saar explained, questioning if this is the end for paying for software.

The fear of getting replaced among software engineers remains alive, and keeps getting scarier with every AI release. “do you think it still makes sense for people learning how to code to understand all the low-level details, or are they better off learning how to converse with agents?” asked a developer to Karpathy in his post. 

The truth is that using AI tools helps in learning how to code. The future of coding from ‘tab tab tab’ is now slowly rushing into just single prompts, which might be scary for software engineers. But it is also essential to learn how a code works before rendering all coding jobs obsolete. 

One thing for sure is that the size of software engineering teams would definitely decrease in the future as companies would be able to create prototypes and others in just a matter of seconds. This is also particularly worrying for Indian IT companies and developers who might be on a brink of emergency pretty soon.

Many people, including students in India and elsewhere, lack the financial means to pay for services like Cursor or GitHub Copilot, which typically costs $10 a month or similar amounts. Access to AI-driven workflows is essential because these skills will be expected by future employers or for starting one’s own business.

It seems like what Russel Kaplan, the president of Cognition Labs, predicted that software engineering would be all about managing a team of AI agents, is actively coming true. Since the AI developers talk about how coding is merely a small part of software engineering, their dreams of letting that task be delegated to an AI agent is finally coming true, all thanks to Replit.