devopsy.kinsta.cloud

Tag: generative AI

  • Are LLMs Leading DevOps Into a Tech Debt Trap?

    Are LLMs Leading DevOps Into a Tech Debt Trap?

    Everyone thought AI would free us from toil. Unfortunately, that’s not true. What it will do, however, is lead to an explosion of technical debt. Some of the most alluring uses of LLMs have the potential to lead to some of the worst outcomes for enterprises and their DevOps and platform engineering teams.

    SK Ventures and Adrian Cockcroft raised the alarm bells in their warning posts about the tipping points LLM introduced into work and society. And in my own Gluecon 2023 presentation, I challenged just how powerful and disruptive the new generation of AI/ML will be for ITOps and DevOps efforts.

    Working Faster is not Working Smarter

    There is no doubt that LLMs are powerful and will change how we think about expertise and knowledge work.  While they are not truly innovating, their ability to draw on such a vast pool of human expertise makes it difficult to distinguish from true creative work. Since they are unbounded, they will cheerfully use the knowledge to churn out terabytes of functionally correct but bespoke code.

    It’s easy to imagine a future where LLMs crank out DevOps scripts 10x faster. We will be supercharging our ability to produce complex, untested automation at a pace never seen before! On the surface, this seems like a huge productivity boost because we (mistakenly) see our job as focused on producing scripts instead of working systems.

    But we already have an overabundance of duplicated and difficult-to-support automation. This ever-expanding surface of technical debt is one of the major reasons that ITOps teams are mired in complexity and are forever underwater. Generating even more scripts, even if they were better written, does not reverse the trend.

    Can AI/ML Avoid This Technical Debt Trap?

    It depends on how we use the expertise in the models. Instead of asking it to generate new code, we could ask it to interpret and modify existing code. For the first time, we have tools to take down the “not invented here” barriers we’ve created because of the high cognitive load of understanding code. If we can help people work more effectively with existing code, then we can actually converge and reuse our systems. By helping us expand and operate within our working systems base, LLMs could actually help us maintain less code.

    Imagine if the teams in your organization were invested in collaborating around shared systems! We haven’t done this well today because it takes significantly more time and effort. Today, LLMs have thrown out those calculations.

    Taking this just one more step, we can see how improved reuse paves the way for reduction of the number of architectural patterns. If we improve our collaboration and investment in sharing code, then there is increased ROI in making shared patterns and platforms work. I see that as a tremendous opportunity for LLMs to improve operations in a meaningful way.

    It comes down to whether we ignore complexity by replacing it or embrace complexity by understanding it. I see this as a choice between independence or collaboration, as illustrated by the table below.

    LLM Driving Independence

    LLM Driving Collaboration

    Create new code

    Interpret existing code

    Replace existing code

    Maintain existing code

    Design new architectures

    Be a guide and mentor

    Build new stacks

    Evaluate alternatives

    Bypass documentation

    Improve documentation

    Can AI Help us Collaborate?

    We keep taking the same hardware and software components and then building unique systems. That variation makes it impossible to leverage each other’s work. At its heart, the reusability problem lies in the difficulty of understanding systems, and that makes it nearly impossible for us to work together, collaborate, reuse and accelerate.

    Breaking this cycle has been our mission at RackN for nearly a decade. We’ve seen tremendous ROI from availability, reliability and speed when customers are able to leverage standardized infrastructure pipelines. We’ve seen firsthand that it is possible to guide teams into collaboration with sufficient guidance, mentoring and training. Collaboration creates 10x outcomes even without AI, but it takes building expertise, deliberate effort and leadership discipline to implement.

    Why does collaboration create 10x ROI?  It works because true innovation actually comes from understanding and building on top of systems rather than rebuilding them from scratch over and over again. The key is being able to leverage LLMs to better understand existing automation and platforms instead of creating yet another bespoke stack.

    Enterprises Should use LLMs to Drive Collaboration

    I believe that the most effective enterprises will use these tools to elevate their IT leaders and architects to focus on collaboration and reuse. They must actually direct them to produce less bespoke automation; instead, they can use expertise in these tools to find and maintain existing systems. LLM presents options, but collaboration requires new thinking and strong leadership.

  • BigPanda Extends AIOps Platform to Embrace Generative AI

    BigPanda Extends AIOps Platform to Embrace Generative AI

    BigPanda today added generative artificial intelligence (AI) capabilities to analyze incidents surfaced by its AI for IT operations (AIOps) platform.

    Blair Sibille, field CTO for BigPanda, said Generative AI for Automated Incident Analysis leverages generative AI platforms based on the general-purpose large language models (LLMs) created by OpenAI.

    The goal is to make it simpler for IT teams to discern the root cause of an issue by, for example, taking advantage of summaries of alerts enabled by generative AI to make it simpler for IT professionals of any skill level to understand what requires their most immediate attention, he added. Today, insights are lost in the sea of alerts that often inundate IT teams, Sibille noted.

    BigPanda has been at the forefront of applying machine learning algorithms to automate IT operations management. Now, the company is incorporating external general-purpose LLMs to further augment IT teams, said Sibille.

    As IT environments continue to be more complex, it’s becoming essentially impossible for IT teams to manage them at scale without relying more on automation. AI technologies take automation to another level, but they won’t eliminate the need for DevOps teams to manage IT environments, added Sibille.

    BigPanda claimed generative AI capabilities will make it possible to reduce the amount of time required to resolve tickets by seven minutes each, with an accuracy rate of 95% during beta testing. Over time, those capabilities not only reduce mean-time-to-resolution but also reduce the level of toil that often conspires to burn out DevOps teams, said Sibille.

    At the core of the BigPanda AIOps platform is a data lake capable of consuming data generated by multiple applications and systems. That data then provides the foundation on which machine learning algorithms are applied to surface insights.

    It’s not clear how much generative AI will automate DevOps workflows, but in effect, more IT teams will be able to take advantage of having the equivalent of a DevOps engineer that they can query any time an issue arises using natural language, noted Sibille.

    AI doesn’t mean that IT outages will be completely eliminated, but it does mean they will become a lot less stressful, he added.

    Naturally, there’s a lot of speculation concerning the impact AI will have on DevOps. Some contend it’s only a matter of time before the entire software engineering process is automated. As those advances are made, the cost of building applications will effectively drop to zero. Others argue that AI will simply augment DevOps teams in a way that makes it simpler to manage IT at scale using either the same number or possibly even fewer engineers.

    The one thing that is clear is that AI will soon be pervasive. DevOps teams should start evaluating today which tasks will be automated with an eye toward planning how DevOps roles will evolve as AI advances accelerate.

  • Key Insights From the Stack Overflow 2023 Developer Survey

    Key Insights From the Stack Overflow 2023 Developer Survey

    Stack Overflow has released its seminal yearly developer survey. The 2023 report, which surveyed over 90,000 developers, provides a comprehensive view of the current software developer experience.

    Below, I’ll highlight a few important findings, namely the significant programming language and tooling preferences, the use of AI in the development workflow and what these trends could mean for the DevOps field. I also gathered insights from Joy Liuzzo, vice president of product marketing at Stack Overflow, to dive deeper into some of the findings.

    Developer Technology Preferences in 2023

    Of note is that more and more developers are learning to code online. Learning to code using online resources increased from 70% in 2022 to 80% in 2023. Although many developers (47%) still have a Bachelor’s degree in computer science or the equivalent, these trends underscore the movement toward alternative knowledge solutions, especially for younger programmers. The topmost online resources include technical documentation, blogs, forums and how-to videos.

    JavaScript is still king—it’s now in its eleventh year as the most common programming language. But it’s worth highlighting that Python has overtaken SQL as the third top language. “SQL was solidly in the top three (JavaScript, HTML/CSS, SQL) since 2015, so it’s a big deal for it to have fallen below Python,” said Liuzzo. “We’ve seen Python’s popularity increase based on the number of questions asked on our public site, so we’ve been expecting some shifts.”

    TypeScript and Bash/Shell have also consistently grown in usage over the last few years. “These two languages are involved with the functions of other popular programming languages, explaining their prevalence among programmers,” Liuzzo said.

    PostgreSQL also overtook MySQL as the most common database type. In terms of web frameworks, Node.js and React.js are the most dominant. Other frameworks, like jQuery and ASP.NET, are showing signs of going out of style, likely because they are older web frameworks.

    Spotlight on AI

    The explosion of new AI innovations, such as large language models (LLMs) and chat-driven generative AI tools, had a major impact on this year’s technology findings. In fact, 83% of respondents have used ChatGPT over the past year. This is followed by Bing AI (20.6%), WolframAlpha (13.36%) and Google Bard AI (9.86%). GitHub Copilot was ranked as the top most-used AI developer tool.

    There was a significant spike in using AI tools amongst those learning to code—these folks often see benefits like speeding up learning, increasing productivity and greater efficiency. “Pair this with how they are currently using AI tools to debug and get help (68%) and learn about a codebase (50%), and we see that the commonality is that AI tools enable learning,” explained Liuzzo.

    That being said, there’s skepticism around AI’s accuracy, and only 13% see improved accuracy in coding as a benefit of using such tools. These pain points may resolve as LLMs evolve, but for now, there still needs to be human judgment to catch bugs and avoid hallucinations.

    “AI has a complexity cliff. Much like Helmsman’s complexity cliff found in project management, after a certain point, the ability for AI to handle all the nuances and interdependencies of a solution drops off,” said Liuzzo. “That’s when humans, their ability to apply judgment and have original thoughts, saves the day.”

    So what are some takeaways for those working in the DevOps field, namely DevOps tooling providers and platform engineers? Well, new tooling preferences and habits will continue to evolve in the DevOps field in new ways. Here are some potential examples:

    • More Markdown-based collaboration. Markdown is the most admired asynchronous tool and has increased in popularity this year. New programmers are adopting a blend of Markdown, GitHub Discussions and Notion for their asynchronous tooling. Flavors of Markdown are already being used for comments, merge requests, issues and more.
    • Continued Docker reliance. Docker remains the top-used tool, followed by npm and Pip. Despite the rise of managed container services, Docker is still a popular way to package and distribute software.
    • Challengers to the big three loom in the distance. Amazon Web Services (AWS) is still the most-used cloud platform at 48.62%. It’s followed by Microsoft Azure (26.03%) and Google Cloud (23.86%). Yet, newer developers want to work more with Hetzner and Vercel.
    • Online learning is paramount to knowledge sharing. Given the steady increase in new programmers and their reliance on online resources rather than traditional schooling, DevOps and open source communities should seek to enhance their online footprint by making knowledge like tooling documentation and tutorials easily consumable.

    Last but not least, generative AI is set to enhance programming in many areas and is already becoming embedded into many environments. Just as AI can be used for code generation and debugging, LLMs will likely assist the DevOps side of the equation, helping automate the CI/CD pipeline, optimize infrastructure, detect anomalies and provide more chat interfaces to initiate these capabilities.

    “AI tools are empowering and enabling learning,” said Liuzzo. “Generative AI will democratize coding and grow the developer community by several folds—and that growing number of developers will be the ones using the tools and also verifying and validating the outputs as they learn—bringing the power of the developer community and the technological power of AI together.”

  • Tabnine Unveils Beta of Generative AI Automated Testing Tool

    Tabnine Unveils Beta of Generative AI Automated Testing Tool

    Tabnine this week made available in beta a tool that enables developers to use natural language to interact with the artificial intelligence (AI) models embedded within its test automation platform.

    Brandon Jung, vice president of ecosystems for Tabnine, said Tabnine Chat extends generative AI capabilities the company developed to make it simpler for developers to invoke its test automation platform from within their integrated development environment (IDE).

    In addition to making it possible to search a codebase more efficiently, that capability can also be used to explain how a piece of code functions, said Jung.

    Developers can also generate new code via a natural language interface using Tabnine Chat in the same way other generative AI tools do, he added.

    In addition to developing its own LLMs, Tabnine previously announced it will be integrating LLMs provided by Google as part of a strategy to mix and match LLMs regardless of where they are hosted.

    Unlike general-purpose LLMs, the ones used by Tabnine were trained using a narrower base of testing code that the company curated using tests made freely available on the internet. No customer data was used to create any of those tests, so developers do not need to worry their code may inadvertently be accessed by anyone else, noted Jung.

    In general, as AI makes test creation simpler, tests will be run more frequently, said Jung. Over time, those AI capabilities and more frequent tests should significantly improve the overall quality of applications running in production environments, he added.

    It’s now only a matter of time before generative AI is applied across the entire software development life cycle (SDLC). Testing, however, may be one area where it will have the earliest impact.

    In the meantime, the debate over the impact generative AI will have on DevOps continues. Some proponents contend most of the software engineering process is about to become automated, while others suggest there will always be a need for a human to be in the loop of any process automated using AI models. Much of the current level of toil associated with DevOps today will be reduced, but the need for software engineers to manage the overall process will remain, they contend.

    The one thing that is certain is that the pace at which applications are built and deployed is about to be greatly accelerated. Prior to rise of generative AI, testing processes were barely keeping pace, so as more code is created with the help of AI, the need for similar advances in testing will become acute.

    There is as much irrational exuberance about AI as there is fear and loathing, but DevOps teams that committed early on to ruthless automation will naturally be at the forefront of adoption. The immediate challenge is assessing what functions will be automated and the impact those changes will have on the way DevOps teams are currently structured. Ultimately, the goal is to eliminate as many of the bottlenecks that conspire to slow application development and deployment while simultaneously improving the quality of the application experience.

  • OutSystems Embraces Generative AI to Speed App Dev

    OutSystems Embraces Generative AI to Speed App Dev

    OutSystems, via a Project Morpheus initiative, intends to leverage generative artificial intelligence (AI) technologies to make it possible to automatically generate an application that developers will then use low-code tools to customize for specific use cases.

    In the meantime, the company is extending an existing alliance with Microsoft to provide a connector that integrates its namesake low-code platform with instances of ChatGPT. These instances are made available via Microsoft’s Azure OpenAI cloud service to make it possible to build a low-code application in a matter of minutes.

    Rodrigo Coutinho, director of data science for OutSystems, said that as generative AI continues to evolve, it will become possible for developers to use a natural language interface to create an application. However, there will also be plenty of instances where end users still prefer to access applications via a graphical user interface that developers build using low-code platforms, he noted.

    An application editor will also provide ongoing suggestions for everything from how the user interface is designed to the way data is managed.

    Developers will also be able to verify how code functions using the OutSystems visual language. In addition, The OutSystem compiler will be able to deterministically detect threats and code patterns produced by generative AI platforms that will include Google, Amazon and others alongside Microsoft.

    Ultimately, citizen and junior developers will be able to build sophisticated applications that previously would have required the skills of a professional developer, noted Coutinho.

    It’s not clear how much developers will embrace a low-code platform to build software. However, the one thing that is certain is the whole process of defining requirements before an application can be built will soon become obsolete as the building of an application becomes more iterative and involves developers working alongside subject matter experts, added Coutinho.

    It may awhile before that vision of application development is realized, but the pace at which applications are built and deployed is clearly about to accelerate. The impact of all that code will naturally vary, but it’s already apparent that the cost of building many applications may soon drop to almost zero. Developers will soon be spending more time on making sure workflows are logical than writing actual code, noted Coutinho.

    In addition to relying more on AI to build software, organizations should also rely more on AI to deploy it at unprecedented levels of scale. The days when application development projects were measured in months, sometimes years, is coming to an end. AI platforms will enable organizations to experiment at will because the time and effort required to build and deploy a new iteration of an application will be virtually nil.

    Of course, no one can predict what impact all that code will have on society as we know it. But the only limit going forward may soon be what our collective imaginations can conceive than the level of developer expertise that can be marshalled at any given time.

  • Harness Brings Generative AI Capabilities to DevOps Workflows

    Harness Brings Generative AI Capabilities to DevOps Workflows

    Harness today added a raft of generative artificial intelligence (AI) capabilities to its continuous integration/continuous delivery (CI/CD) platform as part of a larger effort to reduce the level of manual toil that today conspires to create DevOps bottlenecks.

    The added generative AI capabilities include an AI Development Assistant (AIDA), which analyzes log files and correlates error messages with known issues. It also includes the ability to use natural language to define policies for governing cloud assets and costs and an ability to automatically identify security vulnerabilities and generate code fixes.

    Other generative AI capabilities that will be added to AIDA over the next three to six months include automated code reviews along with tools for creating pipelines and launching chaos engineering experiments.

    Harness CEO Jyoti Bansal said these are the first of a series of generative AI capabilities that will be applied across the entire software development life cycle (SDLC) using large language models (LLMs) that the company has trained using publicly available sources. AIDA, at its core, provides a tool for creating custom prompts for launching queries against LLMs and a vector database that Harness has incorporated into its platform.

    That approach means no customer data was used to create the model and none of the data used will find its way into a generative AI platform that is available to anyone, said Bansal. As a result, organizations don’t need to be concerned they light run afoul of licensing issues because proprietary code was inadvertently included in an LLM, he added.

    Most of the existing generative AI capabilities that have been brought to market thus far are designed primarily to improve developer productivity, noted Bansal. Harness is now bringing the benefit of generative AI to DevOps teams that today rely on too many manual processes to manage the automations they build on behalf of application development teams, he added.

    Harness is making AIDA available at no extra charge at a time when many DevOps teams will soon be evaluating the AI capabilities provided to them. Most DevOps teams are not going to want to manage workflows without being able to rely on AI to reduce toil and eliminate bottlenecks. Overall, Harness is forecasting generative AI will drive a 30% to 50% increase in DevOps productivity. In addition, Harness is forecasting there will be a 50% to 75% reduction in the time spent remediating vulnerabilities.

    Bansal said that as that process occurs, Harness is betting more organizations will move to replace legacy CI/CD platform with more modern alternatives.

    Previously, Harness has invested in machine learning algorithms to automate a range of DevOps tasks. Generative AI, in combination with machine learning, will transform DevOps workflows in ways that will ultimately improve both the rate at which applications can be built and their overall quality.

    It’s not clear whether organizations might be willing to replace CI/CD platforms that they have been customizing for decades, but at the very least, that option is now more widely considered than ever.

  • BMC Adds Generative AI Capabilities to Helix Management Platform

    BMC Adds Generative AI Capabilities to Helix Management Platform

    BMC today revealed it is embedding generative artificial intelligence (AI) capabilities across its BMC Helix service and operations management platforms.

    BMC CTO Ram Chakravarti said the company is incorporating the open source GPT-J large language models (LLMS) to build generative AI models. Those models are optimized for specific domains using data collected by BMC and imported from other data sources, like the Jira project management software from Atlassian. That persistent layer of storage is designed to aggregate multiple types of data within what amounts to a Helix data lake.

    The generative AI capabilities added to BMC’s portfolio include natural language summaries of issues and events captured by BMC Helix Operations Management, along with an ability to create a historical analysis of a user’s specific environment to identify recurring issues more easily.

    BMC Service Management, meanwhile, can now correlate incidents and provide resolution summaries that can be stored. It’s also now possible to run clustering analysis on chat archives to find new or improved utterances that help virtual agents answer queries better and execute workflows.

    BMC initially opted to build its own LLMs because general-purpose LLMs—such as the one created by OpenAI—are subject to “hallucinations.” Hallucinations are created when incorrect data is included in a model. BMC is training its LLMs on a narrow range of trusted data to ensure the recommendations and subsequent automations it creates are valid.

    Over time, BMC will add additional LLMs; the company hasn’t yet decided if there are use cases where it makes sense to rely on an existing general-purpose LLM to add additional capabilities, noted Chakravarti.

    Organizations can also opt to use a containerized version of the BMC LLM if they prefer, but BMC will encourage most organizations to invoke generative AI via its Helix SaaS platform, he added.

    BMC Helix is an IT services management (ITSM) platform that integrates with a range of DevOps platforms. Helix is part of a larger existing BMC Connected Digital Ops initiative that seeks to unify ServiceOps and DevOps with AIOps, DataOps and AutonomousOps. Generative AI will automate a range of tasks from within Helix that can then be extended to DevOps platforms such as Jenkins.

    It’s still early days as far as applying generative AI to processes, so organizations should focus on use cases where it will add some immediate value, said Chakravarti. Once they develop a level of trust in the automation created and executed, they can then pragmatically expand the number of generative AI use cases where they make sense, he added.

    One way or another, generative AI is going to be applied to the management of IT. Many routine tasks that conspire to limit the productivity of IT teams today will be eliminated, but there will always be a need for a human to remain in the loop to ensure the right outcomes are achieved. However, it will also soon become possible for existing teams to manage IT at levels of scale that were previously unattainable.

    Image source: Photo_Credite_Indira-Tjokorda via Unsplash

  • Supercharging Ansible Automation With AI

    Supercharging Ansible Automation With AI

    Whether the latest incarnations of generative AI will live up to their seemingly great promise remains to be seen. This does not always happen with new technology, but the torrent of AI advancements is promising disruption across nearly every industry. Instead of posing a threat to human roles, AI now aims to make tasks easier, faster and more efficient. “This is precisely what we have been trying to achieve with Ansible and automation through generative AI models,” according to Richard Henshall, senior manager of Ansible Product Management at Red Hat. “At the end of the day, AI is another form of automation, and while we do not have thinking computers yet, we do have highly advanced and powerful models that can automate access to data.”

    Case for Integration

    Picture two scenarios: Consider a user who is already skilled in using Ansible. Their limitation is often the amount of time it takes to find the right way to do something, as Ansible is very broad and encompasses many different technologies. The user may be an expert in one or a few of these technologies, but not all of them. In this case, AI can act as an assistant, helping the user find answers more quickly and efficiently. For example, if they want to automate something in AWS, the AI can help them find the necessary resources, and the user can then tweak and curate the automation to create a final product that can be distributed to a large number of people. “This essentially superpowers the individual user,” said Henshall.

    Another type of user may not know how to automate, but they do know how to use a specific technology. “They may not automate that technology because it’s difficult for them to learn something new,” Henshall added. “AI capabilities can help these users get started with automation more quickly, enabling more people to write automation and manage both new and existing technologies.”

    In both cases, the purpose of automation is to help users solve problems. AI serves as an accelerator for these capabilities. “Once users have created automation, the AI model can be trained on the organization’s DNA, including best practices, security posture, policies and enforcement points,” Henshall asserted. This allows the AI to become a co-worker, empowering users and helping them maintain and manage their technologies more effectively.

    “As the AI model learns from the organization and its users, it can help train junior engineers, who will eventually become senior engineers, continuing the cycle of knowledge transfer and improvement,” Henshall predicted. “In this way, generative AI technologies can be integrated into the Ansible ecosystem, benefiting users and organizations in a wide range of contexts.”

    Solving Problems With AI

    As the DevOps leader of cloud testing organization LambdaTest, Shahid Ali Khan has led the company through open-source engagements and investments in next-generation automation solutions, addressing key challenges with Ansible. “We are dealing with a very large scale inventory in terms of geologically separated data centers and multiple OS. Managing this inventory on the basis of these characteristics becomes a hefty task and requires continuous efforts from the team,” he said. “With AI, Ansible can analyze the inventory and automatically group the hosts based on their characteristics, such as their operating system or their geographical location. This makes it easier to manage the inventory and apply changes across multiple hosts at once.”

    Khan also highlights the issue of auto-healing in the industry. “Auto-healing has always been a problem, and while tools in the market have somewhat addressed it, they still require defining errors and categorizing fixes,” he said. “With current AI technology in place, Ansible can detect and automatically fix issues on the basis of historic data in the inventory without the need for manual intervention, reducing the risk of human error.”

    “Inventory forecasting has always been a pain in order to calculate how much inventory we will need based on usage and on what fronts like location and OS we need to scale,” Khan shared. “AI here with Ansible can help detect the usage and forecast based on exact demand.”

    Benefits and Challenges for the Ansible Community

    In the Ansible community, which consists of people who want to solve problems, the benefits and challenges of AI-driven automation are similar to those faced by individuals in their jobs, just on a different scale.

    Community members want to see how generative AI can help them accomplish more tasks, whether or not they are part of an enterprise. Having access to a community model allows them to contribute to training the AI, ensuring its usefulness, as they may eventually work in an enterprise setting during their careers. “The access to community services enables them to take advantage of the benefits offered by AI-driven automation,” Henshall remarked. “It is a mutually beneficial relationship between the community and AI-driven automation.”

  • PagerDuty Signals Commitment to Adding Generative AI Capabilities

    PagerDuty Signals Commitment to Adding Generative AI Capabilities

    PagerDuty has signaled its intention to add generative artificial intelligence (AI) capabilities to its cloud platform for managing IT operations via integrations with large language model (LLM) providers that it will disclose at a future date.

    These capabilities will make it possible to invoke natural language via PagerDuty Operations Cloud to automatically generate status updates, create drafts of incident postmortem reports and even co-author workflows in any programming language to automate remediation of an issue.

    Damon Edwards, senior director of product for PagerDuty, said as AI continues to evolve, the goal will be to minimize the need for human involvement as much as possible. That doesn’t mean there won’t be a need for humans to be involved, but it does mean that many rote tasks associated with managing an IT incident can be automated, he added.

    For example, the reports that are run to determine the root cause every time there is an incident can be automated, noted Edwards. In addition, summarizations of status reports or postmortems after incidents were resolved can be made readily available to any member of an IT team on demand, he added.

    Those capabilities will also make it easier to onboard new members to an incident management team as required, said Edwards. Today, it’s difficult to add new members to a team after an incident has unfolded because it takes time to bring them up to speed, he noted.

    In addition, generative AI also makes it easier to capture a lot of the tribal knowledge within an IT organization that would otherwise disappear when staff either leave the company or shift roles, said Edwards.

    The generative AI additions to the PagerDuty platform extend an artificial intelligence for IT operations (AIOps) capability that PagerDuty made available this spring. The PagerDuty AIOps platform leverages the data model embedded in its incident management software to reduce the amount of time required for an AI platform to learn how an IT environment operates and begin surfacing valid recommendations to optimize workflows.

    It’s not clear at what pace AI will be adopted within enterprise IT organizations, but it is rapidly becoming pervasive. Each DevOps team will need to decide the degree to which they can rely on AI to manage IT processes, but as IT processes become more complex, there is a clear need for AI to navigate all the dependencies that exist in an IT environment. As more routine tasks are automated, however, there will also inevitably be a realignment of roles and responsibilities within an IT organization. The overall goal is not so much to eliminate IT staff as much as it is to enable them to manage IT environments at a much higher level of scale.

    In the meantime, DevOps teams need to plan for not only what is possible using generative AI today but also tomorrow. After all, there’s not a lot of point in training someone to take on a new task if that task is automated a few months from now.

  • Friend or Foe? ChatGPT’s Impact on Open Source Software

    Friend or Foe? ChatGPT’s Impact on Open Source Software

    OpenAI’s ChatGPT took the world by storm, amassing 100 million users in the first two months after its public launch. The continued interest in the tool has created a buzz among developers, especially those in the open source community. But this begs the question: How will ChatGPT impact open source software?

    Amid the excitement and promise that ChatGPT holds, many wonder whether the open source community, just like any other organization, should fear or embrace the technology. Certain concerns for open source contributors and developers are sparking debate as questions surrounding the origin of generated source code surface, and the possible ethical and legal implications are discussed. While valid, developers should not fear or stray away from using ChatGPT but rather shift their focus to understanding how to embrace it for positive outcomes.

    The Top Three Open Source ChatGPT Concerns

    Collaboration between contributors is the cornerstone of open source projects, which some fear will be disrupted with the increased use of AI-based tools. However, the level of synergy within the community cannot be easily taken away or replaced by tools. Instead, ChatGPT or other AI-based tools, including GitHub Copilot, allow developers to produce code more quickly and more efficiently.

    As developers increasingly use AI tooling to assist with new or enhanced code, project collaboration and oversight will help improve the AI-generated code.

    Debunking and navigating the top concerns will be imperative to take full advantage of the technology’s promise and potential.

    Validity. While the creation of code by ChatGPT has generated excitement among developers, critics argue that—without context—the validity of code can be questioned. Some open source developers worry that teams will begin to rely entirely on ChatGPT to generate code, but this concern fails to consider the humans involved in the process—today and far into the future.

    Developers don’t take ChatGPT’s output as the final word; rather, they use it as a baseline and jumping-off point to streamline their code. In fact, code is rarely created from scratch anymore. Developers rely on others’ source code. For example, code from Stack Overflow, GitHub and thousands of open source libraries that are available on public registries like npm, Maven, Nuget and PyPI. Bringing ChatGPT into the fold will not significantly change how developers source their code, but will improve development velocity, saving valuable time and associated cost.

    Training Data. Machine learning (ML) and deep learning (DL) model training must be fair, robust and explainable to avoid biases. If the data is wrong, the results will be as well. Garbage in, garbage out. Coming from ML/DL models, the code ChatGPT returns in response to prompts could raise concerns surrounding its accuracy. Like any other source (Stack Overflow, GitHub, etc.), ChatGPT’s code outputs are not guaranteed to be perfect, and developers must be mindful of this.

    However, there are more benefits of ChatGPT-trained models concerning code outputs. ChatGPT can also explain new or existing code, it can also effectively provide unit tests for code, which will help with writing better software faster.

    Ownership. Questions of ownership have surfaced regarding the use and distribution of the code that the AI tool generates. While the code that ChatGPT generates is the result of ML/DL inference from many sources, it’s the developer’s responsibility to use that code ethically and safely. It should be treated as any other public data or open source software, as not a final product and must be used in the context of your requirements. It’s also important to carefully review any code generated by ChatGPT and ensure that it doesn’t introduce vulnerabilities.

    Similar to GitHub Copilot, ChatGPT is trained by millions of lines of open source software. Code posted in prompts is likely to make it to models, too. Unlike copyrighted pieces of art and writing, the code that ChatGPT outputs should not be considered in a final state and should not be used in a way that is subject to licensing restrictions or legal implications that have dominated the conversations around its use.

    Ownership closely relates to the ethical dilemma of using AI-generated text, code and more. To address that, there are new tools that help detect whether AI generated the content and how much, which will be useful for educators concerned with students overusing the tool. This emerging area of checks and balances exemplifies the constantly evolving and improving technologies that allow humans, and in the case of open source, contributing developers to improve their craft and produce better open source software.

    ChatGPT’s Impact on Open Source Talent

    Discussions about whether ChatGPT will require new skills for existing positions or new jobs with specialized experts are ongoing. While new and exciting, ChatGPT won’t immediately create new or different jobs. Like any new tool introduced to developers, it takes time for developers to become familiar with the technology and understand how to use it best. ChatGPT is no different. For example, consider the previous push to use low-code/no-code technology. While this great technology is used to speed up the creation of apps and improve the usability for non-developers, it has taken time for organizations to properly use low-code/no-code technology. A similar perspective and trajectory will extend to the use of ChatGPT for all software development, including open source projects.

    In the coming weeks and months, it will be essential to encourage the open source community to embrace ChatGPT and explore its possibilities. Technology has already proven to be an effective educational tool. Consider asking ChatGPT for book recommendations about programming languages and coding; it delivers short descriptions for each book. Or, prompt it for the top takeaways from one specific book. By doing this, individuals can make learning so much easier. By learning from other developers and sharing resources, including results from ChatGPT, the community roots of open source will keep thriving.

    Human Involvement is Always Needed

    The public’s reaction to ChatGPT may be novel, but the idea behind the tool’s relationship to open source is not. Analyzing how ChatGPT works within the open source community reveals that it will be useful because it enhances the capabilities of developers, it doesn’t replace them.

    Whether it’s reviewing code, pair programming or learning from fellow developers, humans will not be replaced by generative AI, only enhanced. Leveraging the tool will decrease the time and effort needed to complete tasks, increasing developer quality and efficiency. The use of ChatGPT does not create new positions or overhaul current skill sets; it makes them better and empowers developers to expand their opportunities and enhance their abilities.

    Soon, ChatGPT and other AI tools will be standard practice within the open source community—or until a new technology is created and shakes up software development once again.