Getting the hang of machine learning can be daunting, so Google has decided to provide a few examples that can run on its open-source TensorFlow framework, all bundled into a friendly browser-based package called Seedbank. It consists of a range of different examples, called seeds, that show different use cases and conventions for code related to machine learning. Each tutorial is able to be modified, forked and run directly in your browser, with Google providing the GPU backend you need to run te programs via the cloud. This means that you can start learning about how machine learning works and what it takes to create a compelling machine learning program no matter what sort of computing equipment you have lying around.
There are nine different categories of seeds to choose from, running the gamut from the most basic of machine learning examples to code that can run unsupervised and even more creative code that can generate, examine or modify things like images and music. All of them run in real time on Google's backend, and you see the input code and the results on your own computer. There is one seed, for example, that allows you to check on the state of a neural network from top to bottom. Another one, called NSynth, uses Google's Magenta project to put together brand new music using procedural generation. Yet another lets you train a word examining AI to pick up on bias in pieces of text.
Along with Seedbank, Google has a similar tool called Colaboratory that runs machine learning code in the browser and allows newbies to learn all about how TensorFlow works, among other lessons. These two tools, much like Google's educational efforts, are essentially pushes to get more people into development. Because of just how different coding AI applications is from coding traditional application, those with no coding experience at all, hardened software developers and everybody in between can benefit from these tutorials, if they want to work with TensorFlow. Even those who may have never bothered learning to code otherwise may have enough of a knack for it to make them fit for tomorrow's AI-centric job market, after all.