AI Joe Bot, AI in Mobile, Testing AI and More TGNS118

[00:00:00] Joe Colantonio Have you seen the latest solution for AI in mobile app testing? What are some key challenges in testing Generative AI applications? And have you seen the new AI powered Joe Chat bot? Find out in this episode of The Test Guild News Show for the week of April 28th. So grab your favorite cup of coffee or tea and let's do this. Before we get into it. If you haven't already, make sure to subscribe to our Test Guild LinkedIn News Show newsletter that you can find in the link down below.

[00:00:26] Joe Colantonio First up is introducing Mabl's awesome new AI mobile app testing solution, and this goes over how Mabl has launched a new mobile app testing solution designed to help developer engineers, quality engineers, and testers ensure the quality of mobile app testing. And this addresses a huge challenge in mobile app development by enabling high quality app delivery at scale, without the traditional complexities associated with mobile testing. I found out about this on my automation test guild, where I spoke to Dan Belcher, co-founder at Mabl all about it, and he went over a bunch of different ways of how this solution can help you. So the first, it really makes test creation easy. It allows test members to create mobile app tests with low code options, helping reduce dependencies and coding and script maintenance. It also helps with capabilities for software development engineers in test, SDETs, and developers who can now utilize JavaScript and Appium to create reusable libraries and enhancing your test efficiency. And it also tap into the cloud powered scalability feature, which helps facilitate testing across multiple virtual device simulations, significantly helping you reduce test times. And it also integrates with Android, iOS, and APIs into a single suite, streamlining processes and improving overall software quality. So if you're doing anything with mobile testing, this is a solution you have to check out. And the free trial is really easy to get up and running with, and you can find it in the comments down below.

[00:01:55] Joe Colantonio Do you want to tap into the wisdom of over 10 years of automation, performance, and security testing? Check out our new AI chat that pulls from over 10 years of our podcast interviews, 14 years, a blog post, and 14 years of YouTube videos, all with attributions. Check it out. So this AI Joe bot has been trained on all the content I've created, so I need your help us testers to check it out. Give it your best questions, see how it answers and let me know! You can drop an email if for some reason doesn't answer as you expected, but what's great about this one? It actually gives attributions of where it found the information. So for example, if you typed in like what are some good practices for BDD or Playwright or Selenium, anything like that. So not only does it answer you, but it also gives attributions of where it found the information. So you can cross check to see if it's hallucinating or not, because you could actually see where it pulled info from. So if you click on one of the numbers next to where it provided information, for example, if you clicked on one, it'll show you all the resources of read from all the podcasts, all the videos, so you have confidence in knowing where it actually got the information. Not that it's just making things up. Also, it's kind of cool, as you can hear me, I don't know if this is cool or annoying, but if you actually want to hear me read this to you, you can click on this little audio icon. It will read it back. First question, in behavior driven development, you can also ask it to be more concise, to go deeper or to give examples as well. So look forward to see how it handles all your awesome questions. So check it out and let me know your thoughts.

[00:03:22] Joe Colantonio So speaking of generative AI, how do you test it? Well, I found a blog post that goes over how to do this with examples. And this is by John, who has a link to a post that outlines key challenges and changes in testing generative AI applications. And so this blog post goes over some of the reasons why these systems are hard to test. And one significant shift is moving away from traditional Big Bang integration testing to more focused white box unit and component integration tests. And this approach not only is more cost effective, but also essentially due to the computational costs and the need for extensive testing to ensure consistent behavior. This approach is not only more cost effective, but also essential due to the computational costs and the need for extensive testing to ensure consistent behavior. And he goes on how collaboration between testers and developers is becoming more crucial to reduce duplication of efforts and cost. Understanding the nature of the AI model, whether proprietary or off the shelf in its training process is vital for defining the testing scope. Security concerns also take a new dimension in generative AI, requiring proactive threat simulations and robust data protection strategies.

[00:04:37] Joe Colantonio Also on LinkedIn, Selenium has announced a new release. What's it all about? Let's check it out. Probably one of the biggest key takeaways right off the bat is it announced that there's been an uptick in more than 2.3 million monthly users. So don't believe the hype that you hear of the death of Selenium. Selenium is going strong, and they're constantly improving. As you can tell by some of the features in this new release. And if you go to the read me note, you can see how Selenium continues to expand its user base by enhancing new features. Some including the WebDriver BiDi feature for Java and JavaScript, nightly built publications for .Net and C# and JavaScript on GitHub. And the refactor of the Selenium manager code across all supported languages to enhance maintainability. So as always, some awesome work by the Selenium team. Give them some love. Give them some thanks. Check out the new feature and let me know your thoughts in the comments down below.

[00:05:29] Joe Colantonio And for fans of the Robot Framework, I have a resource for you as well. So Roy pointed out to me a new feature his company Testsmith has just released for framework testers. It's a resource that can help you as someone new to Robot Framework, get up and running really quickly with the new project. It's a copy or a template, which makes it easier for you to scaffold a new robot framework project. And as you can see, here's a quick video that goes over how easy it is to get started. And there's also a link to GitHub where you can try the copier robot framework template yourself.

[00:06:00] Joe Colantonio So if you've done anything with infrastructure automation, this next article may affect you. And it is how HashiCorp joins IBM to help its multi-cloud automation. So this article goes over how HashiCorp has officially announced its acquisition by IBM, in this saying that helped significantly step up its mission to accelerate multi-cloud automation. And they also go over how this partnership aims to extend the reach of HashiCorp tools for a wider audience, enhance a company's capabilities in the multi-cloud sector. And it also goes over how this acquisition will help HashiCorp continue to develop its products while operating as a division within IBM. So is this good or bad? Well, experts in this area on LinkedIn have made some noise of some pros and cons of what they feel, how this is going to impact folks, and some of the main concerns is how it's going to impact the OpenTofu project. And if you go into the comments, you could see people talking more about the pros and cons of it as well. Like Paul Bruce is also an expert in this area as well. So if you're using HashiCorp, here's something you need to pay attention to and you can find more about it in the comments down below.

[00:07:09] Joe Colantonio And next up is a Follow the Money segment. And so this blog post goes over how RWX a software development company has announced the launch of Mint, which say new CI/CD, Continuous Delivery platform, and it's supported by $7 million of seed funding that they acquired a few years ago. And how this new solution aims to address the inefficiencies of existing platforms, which primarily rely on outdated methods like running scripts on virtual machines. So how is it different? Well, the article goes on how Mint introduces a task-based direct graph approach, moving away from the traditional script-based operations. And this shift allows for content-based caching, meaning that unchanged tasks do not need to be rerun, thereby reducing unnecessary computational work and speeding up the development process. It also goes over how this platform also enhances the user experience by replacing the raw log outputs with rich, structured semantic results, making it easier for developers to understand the addressing issues directly from within their browsers. And it goes of how Mint local CLI operations enable developers to run workflows directly on their local machines without the need for continuous commits and pushes, which often slow down the development cycle as well. Seems like a cool solution. Check it out. Let me know your thoughts in the comments down below.

[00:08:31] Joe Colantonio And in security news, we have JFrog announcing a new enhancement as well. And this is how JFrog has announced a new machine learning Lifecycle management feature that integrates its Artifactory platform with MLflow, an open source software developed by Databricks. And this integration aims to address common deployment failures in AI application development by ensuring ML models are immutable and traceable. Enhancing the security and the province validation processes. And this initiative hopefully will help organizations streamline your DevSecOps workflows, making it easier to transition ML models from experiments to production with better automation, visibility, control, and traceability. In this article also emphasizes the necessity for a unified system that manages all software artifacts, including ML models, with a straightforward versioning lifecycle and security controls to foster trust and manageability within development teams. So if you're doing anything with ML models, this may be a solution you should check out as well.

[00:09:35] Joe Colantonio All right, for links of everything of value we covered in this news episode. Head on over to the links in the comments down below. So that's it for this episode of the Test Guild News Show. I'm Joe, my mission is to help you succeed in creating end-to-end full-stack pipeline automation awesomeness, as always, test everything and keep the good. Cheers.