The first AI Agent programmer is commercialized and is more than twice as expensive as ChatGPT Pro

At 1 o'clock this morning, Devin, the world's first AIAgent programmer, announced that it would be fully open for use and commercialized. It’s just that the price is not that expensive. The ChatGPTPro released by OpenAI a few days ago can have unlimited use of full health o1, new version of Sora and other functions for only $200 a month. Devin is available for $500 a month in one step, making it one of the most expensive generative AI products currently available.

Netizens said this is a bad business model. Fees are negligible for large groups, prohibitive for small groups or individuals. You should at least come up with a personal subscription model.

$500 is really expensive.

The best month, good products came out one after another.

You guys screwed up on your marketing. The core premise of how it works is trust. How can I trust this thing to work as you say it will without trying it out and building a relationship? With the current state of AI programming tools, I wouldn't pay $500 a month without testing it first.

It’s incredible that Devin went from initial demo to full use in less than 365 days. It has huge potential to make the world a more interesting place, congratulations to the team on today’s launch!

Of course, it’s reasonable to be expensive. Devin’s automated coding capabilities are very strong. For example, developers can assign tasks from the to-do list to Devin, and Devin will generate a first draft PR according to the instructions. For targeted code refactoring, Devin's IDE extension supports VSCode and its branches, which can point to the part of the code you want to edit or upgrade.

Devin is equally great at helping teams with build integrations, migrations, and maintaining documentation. This includes when you assign Devin those low-level, repetitive coding tasks; tell Devin how to test or review the code he has completed; keep sessions under about 3 hours, and break down large tasks;

Share detailed requirements in advance; and train Devin by providing feedback in chat, accepting suggested code, or manually adding your own code.

Devin can also launch sessions in Slack as the primary interface for offloading small tasks and fixing bugs. Once the task is completed, Devin will send a message to notify the user so that they can review the PR submitted by Devin at their convenience. Devin can also automatically respond to PR comments on GitHub to provide ongoing feedback and support.