Hey there! GitHub is about to make a big splash in the AI scene, making it super easy for developers to play around with and build apps using the newest AI models. No more jumping through hoops!
Easier Access to AI Models
GitHub is kicking off a cool new project named GitHub Models. This project wants to be your go-to guide for trying out and crafting apps with the latest generation of AI models.
You might have heard of GitHub’s GitHub Copilot, which is like having a smart buddy that helps you write code faster. Now, GitHub is set to surprise us with even more.
A Playground of AI Models
With GitHub Models, you get a treasure chest of AI models right at your fingertips, all within the GitHub platform using your existing GitHub account.
This new adventure opens the door to models like Meta’s Llama 3.1, OpenAI’s GPT-4o, Mistral Large 2, AI21’s Jamba-Instruct, and Microsoft Phi-3, with more from Cohere.
The Big Picture
The aim is to let developers do way more than just finish their code; they can now tinker with AI models to whip up smarter apps.
Mario Rodriguez, Senior Vice President of Product at GitHub, said that apps in the future will be brains on their own, and GitHub is here to fuel that revolution.
Cutting Down on the Complexity
One of the main goals of GitHub Models is to cut the red tape for developers. Bouncing between different websites and making numerous accounts to try out various AI models was a real pain. GitHub wants to erase that.
GitHub Models is like your one-stop shop for all things AI. You can explore a catalog of models within the GitHub platform and start experimenting without breaking a sweat.
This is a total game-changer because it makes exploring and using AI models to build applications a breeze.
From Experimentation to Enterprise
GitHub isn’t just about playing around; they also want to help developers take their AI-powered apps to the next level. The path leads through Microsoft’s Azure, which makes sense since GitHub is part of the Microsoft family.
Here’s the play-by-play: developers mess around with AI models in the GitHub Models playground to see what they’re capable of. Once they’re ready, they move to a more professional setting like a GitHub Codespace or VS Code, and use Azure’s tools to get the tokens and API keys they need to make their apps work with Azure.
The road to using AI in businesses is about overcoming challenges. Rodriguez pointed out three main hurdles developers face with AI models: latency (how fast they work), quality of responses, and cost. GitHub Models aims to help developers navigate these issues by providing a space to test and compare models.
While there are industry standards for AI models, Rodriguez says you really need to test them out for yourself to make the best choice.
GitHub Models is all about making AI more accessible and less complicated for developers. By providing a centralized place to experiment with a variety of AI models, GitHub is helping to push the boundaries of what’s possible in app development.