“For example, you might have a game developer using an LLM to generate game character descriptions or parameters on the fly. For that to work, the LLM simply ‘must’ generate that information reliably in valid JSON format. Right now, developers have to rely on third party tools like Pydantic, Zod, LangChain, etc. to do things like iteratively run the same prompt until the inference returns data that’s usable, which is not a great solution,” Shimmin said.

What developers want is both in-model function calling, where the model’s syntax incorporates this functionality, and a model that itself knows how to work with APIs and formatting languages like JSON properly, the analyst explained, citing the example of Phi-based model NuExtract.

Why Salesforce is ‘open sourcing’ these products?

Analysts believe that Salesforce is releasing these products as “open source” in order to gain market share for capabilities such as Agentic actions.

“By making this tool available to developers and researchers, Salesforce aims to accelerate the development and refinement of function calling models, potentially leading to more robust and reliable AI applications,” Hinchcliffe said, adding that the open-sourcing of APIGen can also lead to more specialized and efficient AI solutions that are better tailored to the unique needs of various industries, enhancing business operations and customer experiences.

AI is the driver

If the concept of abstraction has been around for decades, albeit called different things, why is there renewed interest now? A significant factor driving data hyperaggregation is the growing focus on AI and machine learning (ML). As companies increasingly integrate AI and ML into their workflows, the need for consolidated and high-quality data becomes imperative. Cloud platforms, by virtue of their comprehensive service offerings, provide an ideal environment for AI-driven applications that require large-scale data processing and analysis. With hyperaggregation, AI models can access diverse, accurate data sets and improve the robustness and accuracy of their predictions.

In the context of economic viability, data hyperaggregation has a compelling sales pitch. Migrating to cloud platforms can involve costs, but the benefits derived from enhanced data analytics, reduced operational inefficiencies, and faster time to market often outweigh these expenses. Organizations are empowered to reallocate their financial resources more effectively, directing them toward innovation and strategic initiatives rather than hardware and infrastructure maintenance.

The push toward ubiquitous computing aligns perfectly with the principles of data hyperaggregation. By adopting a model where computing infrastructure spans edge locations, central data centers, and multiple cloud environments, businesses ensure that data is processed and consumed where it is most efficient and valuable. This approach optimizes costs and bolsters performance and resilience against potential disruptions.

GitHub vs. Bitbucket

GitHub isn’t the only hosted enhanced Git service, and GitHub Enterprise isn’t the only on-premises product for companies. Atlassian Bitbucket competes with both of them, with slightly lower pricing and with a free five-member team level that includes unlimited private repos and the use of Bitbucket Pipelines for continuous integration. GitHub is a more popular site for open source projects and it has a much larger pool of open source developers. Bitbucket’s pricing used to be more favorable for small startups. Now that GitHub allows unlimited private repos on free and team accounts, that’s no longer the case.

GitHub vs. GitLab

GitLab competes with both GitHub and Bitbucket, both hosted and on-premises. On the surface, GitLab appears to have more lifecycle functionality than the others, but the difference from Bitbucket mostly disappears if you include Jira when you evaluate Bitbucket. GitLab offers Gold-plan cloud features to open-source projects for free, but that additional functionality doesn’t really compensate for the larger open-source developer community on GitHub.

GitHub Desktop

GitHub Desktop, shown below, makes it easy to manage your GitHub.com and GitHub Enterprise repositories. While it doesn’t implement all the features of the Git command line and the GitHub web GUI, it does implement all the operations you’ll do on a daily basis from your desktop while contributing to projects. Typically, you will clone repos from GitHub to GitHub Desktop, sync them as needed, create branches for your work, commit your work, and occasionally revert one or more commits.

To work with repos for which you lack commit and collaborate privileges, you typically start by forking the repo on GitHub and cloning the fork to your desktop. Then you add any branches you need in GitHub Desktop, commit any changes you wish, test your work, push the commits back to your remote forked repo, and finally generate a pull request to the parent project.

You can see the Pull Request button at the upper right of the GitHub Desktop interface. You can also see many commits in the Neo4j project that were merges of branches or pull requests. That’s typical of open-source projects with few committers and many contributors.

GitHub for open-source projects

Open-source software projects often need ways to enforce quality control while still accepting contributions from outside the core team of committers. The need for contributors is huge, but bringing new contributors into the project while maintaining the integrity of the codebase is a difficult and potentially dangerous undertaking. At the same time, the need for feedback from users of the project is also huge.

GitHub has a number of mechanisms that can help grease the wheels of open source projects. For example, users can add issues to the project on GitHub to report bugs or request features. Some other systems call these tickets. Project managers working with issues can generate task lists, assign issues to specific contributors, mention other interested contributors so that they are notified of changes, add labels, and add milestones.

To contribute to a project, you basically start from a topic head branch that contains the committed changes that you want added to the project base branch and initialize a pull request from the head branch, as shown below. Then you push your commits and add them to the project branch. Other contributors can review your proposed changes, add review comments, contribute to the pull request discussion, and add their own commits to the pull request.

Once everyone involved is happy with the proposed changes, a committer can merge the pull request. The merge can preserve all the commits, squash all changes into a single commit, or rebase the commits from the head branch into the base branch. If the merge generates conflicts, you can resolve them on GitHub or using the command line.

Code reviews on GitHub allow a distributed team to collaborate asynchronously. Useful GitHub tools for reviewers include diffs (the lower half of the screenshot below), history (the upper half), and blame view (a way to view the evolution of a file commit by commit). Code discussions on GitHub go into comments that are presented in line with your code changes. If the built-in tools don’t suffice for your project, you can add code review and continuous integration tools from the GitHub marketplace. Marketplace add-ons are often free for open source projects.

GitHub gists

Gists are special GitHub repositories for sharing your work (public) or for saving work for later reuse (secret). They can contain single files, parts of files, or full applications. You can download gists, clone them, fork them, and embed them.

Public gists can be discovered and found in searches. You can use keywords to narrow down what you find, including prefixes to restrict the results to gists from specific users, gists with at least N stars, gists with specific filenames, and so on.

Secret gists are not searchable, but anyone with the URL can see them. If you really want your code to be protected, use a private repository.

As we’ve seen, GitHub provides Git repositories as a service, along with features for code review, project management, integrations with other developer tools, team management, social coding, and documentation. While GitHub is not the only product in its category, it is the dominant repository for open-source software development.