New Relic’s Jemiah Sius on How AI Changes Software Development

hi I'm James Maguire and on today's video we're taking a look at our artificial intelligence is changing software development we'll also try to forecast the future of AI and software development to discuss that I'm joined by Jamia cias senior director of developer relations at New Relic jamaah very good to have you with us today hey James thank you for having me I'm super excited to join you today excellent so where are we now with the role of and software development is it still emerging is it honest way to being fully adopted no I think we're still emerging at this state um developers and organizations alike are experimenting with AI um we're going to see a lot of innovation in the next few years uh so for software developers we're still experimenting with uh the usage of AI figuring out how to best utilize it trying to better understand performance cost management um and we're finding that it's great for crunching data right so automating tasks going through data sets um and we're also experimenting a lot with the value of like AI code assistant so we're very much still in that testing and development phase and figuring out how AI really will shape the future you know it may not be possible to to quantify this is there any way to kind of give a rough estimate like oh it's used in 5% of products or it's used in 40% of projects or is that you can't even give that number at this point I I I don't think there's a number we we can give yet I will say I think at least uh 80% of all Enterprise organizations are are testing AI which puts it in the hands of a lot of developers but I think for the developer Market it's really hard to nail down exactly what that count is for uh person to person I would assume that it's pretty high like every software engineer that I'm friends with is testing and experimenting super excited about it we're all going to the AI meetups um but I I don't think that we have the number just yet but I guess the the key phrase is testing and experimenting no one thinks like well I've used it in five you know 10 15 products I guess I'll use it in yet another one it's still kind of it's still new in that sense yep absolutely yeah all right so is AI plays a greater role what what are the key tasks you think it'll be performing more of like what will be the benefits um some of the things I mentioned right so data processing is going to be uh a great benefit of AI it's going to be quicker than a human can get through the large data sets um I think we're going to start seeing um just the pace at how we're able to process data speed up over the next few years as well as the level of insight that we can get out of the data that we currently have um another one is I think our deployment or cicd processes will be changed just allowing AI to understand uh what the deployment process looks like and recommend improvements in your bill paths or pipelines based off of the data that it's tracking right from the output of those systems and another thing I think is um observability will become much more important over the next few years I actually don't see a way uh for AI to be fully adopted without observability uh so many more applications so many more concerns uh you need visibility into the performance of these systems you need visibility into the data that they have access to um and we'll see AI changed the observability industry to allow us to be more um more react I mean more proactive I'm sorry versus being reactive right so using AI to understand Baseline Baseline performance and then make predictions and recommendations of like hey if your system is uh continuing to go in this direction in two weeks you're going to have an outage versus going the two weeks and experiencing the outage having the triage you're everyone's in the War Room trying to figure out so I think those would be some of the shifts that we see from AI in the near future you know I guess I was thinking you were going to say AI writes a lot of the code being not not a developer myself I'm thinking there as a developer can say to you know to the interface you know write this chunk of code and and the and you know the interface will do that is it not actually writing a lot of code for developers oh no we're definitely we're definitely using um assistance uh I think that's one of the parts that um a lot of developers are experimenting with and we'll see a lot of that as well right so using AI tooling to assist with uh writing code uh doing some of the the bits of the work that you wouldn't want to do as an engineer right so like a lot of developers aren't that fond of writing test and now you can focus on some of the code that is really important to you and then just tell your assistant is hey you know write me a test that's going to check for these things so we will see some of that and that's how we're leveraging it now it comes with its hiccups and things that we have to manage right but I do see that as the the path forward with with AI let's talk about uh pros and cons of AI and software development I I know there's many pros and we should talk more about the benefits I want to make sure I I at least touch on the cons because I'm I'm imagining there's a there's many developers out there who going you know what this this system will write code for me it'll do all these things what am I needed for or maybe unless I'm an absolute superstar in the field I'm going to be out of business what what would you say to those folks um so Pros like I mentioned rapidly speeding up work reducing 12 um I think that um AI is going to put developers out of business like it's not going to replace the software engineer uh and it's for some of the cons right like we're still figuring out how to resolve things like hallucinations um and that creates the need for a lot of human oversight making sure that your um especially generative AI products are putting out um trustworthy responses and that's going to require a person to kind of like oversee it make sure that they're troubleshooting and finding these issues within their tools um and I also think that there's going to be a lot of Regulation uh and compliance uh you know needs that come down the pipeline and that will require people as well so more so than uh replacing the workforce I see it Shifting the way that we work and think about uh the software development life cycle I think the the role of a software engineer in some of the skill sets may may uh change a little bit but I don't see it replacing the workforce interesting about the role in the skill sets changing what what do you see in that regard um so I think for the junior level engineer uh where you have uh an AI assistant that can write a lot of the code for you I could see some disruption that space right so the senior level Engineers I don't have too much concerns about but now when you're getting into entry level and Junior Engineers I do think you're going to have some organizations that's looking at how they can leverage an entry level engineer plus AI assistant to have more output right and so uh with that I think the skill set of what we're looking for coming out of University or boot camps may you know shift a little bit into looking for people that have experience in building and working with um AI services or uh have a a background in things like data science or machine learning as you know those very nice to haves now when you're looking for that software engineer to join your team because they'll be able to come in and work with your tools or build some of those new tools understand the way that the product's working so I can see some of those things start to shift right so the expectation of your output will increase rapidly right so the time to deliver to Market shorter you expect it to ship a lot faster but also the skill sets that you need to be able to do that will will shift a little bit let's talk a bit about uh your job at New Relic it sounds really interesting so you do a lot of developer relations what's what sort of issues do you work with in that sense yeah so uh my entire team focuses on the way that our customers use our platform their P pain points and requirements uh globally so we're thinking about our users the best way to provide them with Technical Resources and tooling uh helping them write code building example applications putting uh a lot of technical research out into the market so with AI in particular we're tracking uh open source llms how to build like what are the best practices for working with generative AI what is managing performance and cost look like we're trying to figure out how to find that balance between the two and then putting all that data out into the market for our customers along their Journey as they work to adopt AI into their services um and then a big part of that is just learning what we're doing internally right so eating our own dog food because uh as we launched our AI tool New Relic AI we had to figure out a lot of these challenges right so for us we're learning we're experimenting and then we're you know giving that that uh knowledge back to the community great all right well let's look to the future I think that's the the big question a lot of companies are dealing with it's like where is it all going and how can we prepare for it now so what do you see as the future of AI and software development what will it look like say a couple of years down the road I think uh for me the future of AI is really predictive like right now um Genera of AI is is explod it's exploded over the last few years and everyone's experimenting uh with this one form of AI which is absolutely helpful it reduces the barrier to entry and allows uh a lot of uh even non-technical people to come in and explore with this new technology right so I think that uh generative AI is disrupting a lot of Industries at the moment but as we start to move forward into the future we're going to see that folks uh go from using AI just to create into really exploring data data sets understanding patterns and being able to create predictions uh as the future of where we're going with the technology so as I mentioned before um to give an example if you're running this large scale Enterprise your AI tooling will be able to understand issues that you may have in your underlying infrastructure without you being aware of it and make set recommendations uh either exporting or or creating the code that you need for a fix uh or using what we've learned with gener AI in these few years to provide solutions that are of value uh before issues happen so I do think the future of AI is just um being a predictability it's really interesting about the predictability um it seems like the the system gets more able to make predictions if it itself has more information if it has trained more so so it's like whatever predictions we make we can make now we can make better predictions when the large language models are more trained correct incorrect right absolutely I think data becomes so much more important as we start to build and explore uh using AI because it needs that that data to be trained on and that's also another part of what has to be figured out right so a big part of that is um how do we protect uh and secure the data that we're leveraging how do we make sure that we're training uh our AI tools on our our data versus you know public data sets or data of our customers so there's a lot of uh compliance and regulation details uh dealing with data security data privacy that have to be worked out because the data is going to be so important as we continue to explore and and build with uh AI tools it it is all about the data no doubt um Jamiah I think you said it a lot of interesting stuff I learned a lot thank you so much for sharing your expertise today uh please come back and talk with us again sure I'll be happy to join thank you so much for having me and um thank you for the conversation

This transcript was generated automatically from the video's captions and may contain errors.