Free YouTube Transcribe

Video transcript

Maintaining reliable codebases with GitHub Code Quality

Microsoft Reactor · 5,246 words · 24 min read

Want to search this transcript, jump the video from any line, or download it as TXT, SRT, or VTT?

Open in the transcript tool

Full transcript

0:07Hey everyone, thank you for joining us

0:09for the next session of our spotlight on

0:11GitHub advanced security series. My name

0:14is Anna. I'll be your producer for this

0:16session. I'm an event planner for

0:18Reactor joining you from Redmond,

0:20Washington.

0:22Before we start, I do have some quick

0:24housekeeping.

0:26Please take a moment to read our code of

0:27conduct.

0:29We seek to provide a respectful

0:30environment for both our audience and

0:32presenters.

0:34While we absolutely encourage engagement

0:36in the chat, we ask that you please be

0:38mindful of your commentary. Remain

0:40professional and on topic.

0:42Keep an eye on that chat. We'll be

0:44dropping helpful links and checking for

0:46questions for our presenters to answer

0:48live.

0:49Our session is being recorded. It will

0:51be available to view right here on the

0:53Reactor channel.

0:55With that, I'd love to turn it over to

0:57our speakers for today. Thank you so

0:59much for joining.

1:03Welcome back everyone to another episode

1:06of the GitHub advanced security series.

1:08I'm your host Lupita Caravis and with me

1:10today I have a very special guest

1:12joining me from the product team to talk

1:15about our new code quality offering. But

1:17before we get to introductions, I'd like

1:19to set the stage on what to expect. For

1:22those of you who are new here, Reactor

1:24serves as a platform to provide ungated

1:27content to the community. I highly

1:29recommend folks to opt in to the GitHub

1:32notifications to be the first to find

1:33out about what's happening. This is a

1:36followup from a universe talk, so I

1:38always say this is me casually saying

1:41like and subscribe so you're the first

1:43to know. I actually started the series

1:46with one goal in mind to provide a space

1:48for collaboration especially in this new

1:52era of AI where there's just no room for

1:55compromise and I always say our

1:57competition isn't each other it's the

2:00forces beyond our understanding so

2:03GitHub has always stood to accelerate

2:05the human progress and so our mission is

2:07to help develop the most secure high

2:09quality code possible I'm an enterprise

2:12application security executive here at

2:14GitHub. I've been with GitHub for about

2:163 years and I always like to give a

2:18little bit of insight on my background

2:20so you understand where my perspective's

2:21coming from. But essentially what my

2:24role means is I serve as an ally to the

2:26community to help guide organizations on

2:29their security and code quality journey.

2:32I've been in the space for about 10

2:34years and I actually look back to when I

2:37was a developer like fresh out of

2:38college and I think oh because I had a

2:414.0 know I had it all figured out in

2:44compsai. Um, little did I know I was in

2:47for a rude awakening, right? Classroom

2:49code does not translate to

2:52enterprisegrade scalable programming.

2:54And so I I I remember like looking back

2:57at my code and thinking like who wrote

3:00this? And then I realized it was me. I

3:03did. So I can't even imagine what it's

3:05like nowadays where developers don't

3:08really internalize the outputs of their

3:10inputs. or maybe some do, but like I

3:13think the importance of having automated

3:15code quality scanning is key alongside

3:18with security. And I know you might be

3:20thinking, okay, well, there's like a ton

3:22of solutions already out there that do

3:24this. Um, but I think our goal is to be

3:26like the one-stop shop, right? To be the

3:29single pane of glass that has it all.

3:31Um, I always say like for example, we

3:33have multiple streaming services as

3:35personal users, right? Different

3:37interests, different needs and so on. So

3:41what we aim to do today is to guide you

3:43through our philosophy and approach to

3:45how we're tackling this code quality

3:47problem. As we know, quality can mean

3:50different things to different people. So

3:53I invite you to keep an open mind. Um

3:55with that, I'm going to go off camera so

3:58I can monitor the chat for any

3:59questions, comments. You know, we like

4:00to make this as interactive as possible

4:02and hand it over to Harrow to take the

4:05stage and walk us through what we're

4:07doing at GitHub.

4:12Hi everybody. My name is Carol. I'm a

4:15product manager here at GitHub and I am

4:18leading our new code quality initiative.

4:22I've been with GitHub for about three

4:25years now and I'm really excited for

4:27this new product and I'm going to talk a

4:30lot about it and show you a demo of it

4:31today.

4:33So, let's get into it. These are the

4:36topics that I'm going to hit on

4:38throughout. Uh I'm going to set a little

4:41bit of context about code quality and

4:44what it means in an AIdriven world, how

4:46we're thinking about it at GitHub. I'm

4:48going to give you a demo of the public

4:51preview for our new GitHub code quality

4:53product. Talk a little bit about the

4:55road map. And then at the end, we'll

4:57have some time for uh Lupita and I to

5:00answer some of your questions that come

5:02up in the chat.

5:06So, it's no secret that AI has truly

5:10turbocharged software development.

5:12Suddenly, features that took quarters

5:15can take days or fewer to generate. It's

5:18the age of every engineer becoming a 10x

5:20engineer. The speed of this code

5:23generation is putting strain on our

5:25traditional systems and really showing

5:27where our scale is failing.

5:31There's this saying, slow is smooth and

5:35smooth is fast. When it comes to racing

5:38or software development, the

5:40counterintuitive truth is that speed and

5:43control actually go hand in hand. Simply

5:46pumping the gas or throwing AI at your

5:49code is really great for fast

5:52straightaways, but as soon as you hit a

5:55turn, you're going to crash out. In

5:57development, those turns are where

6:00quality, safety, and control are put to

6:03the test, and that's via things like

6:05code reviews and quality or security

6:08gates.

6:12But it's a trade-off. If you focus too

6:15much on the speed, you get technical

6:17debt, bugs, and ultimately lost trust

6:21and revenue. But if you hit the brakes

6:24too hard, the reviews become slow, your

6:28developers workflows are disrupted, and

6:31you lose the advantage from the speed

6:33that you get with AI.

6:35This is a really easy trap to fall into,

6:38especially when your tools are not ready

6:40to support the new world of AI

6:41development.

6:45And I, as a representative of the

6:47developer tools company GitHub, have a

6:50solution for you. Here at GitHub, we

6:53believe that all code, whether it's

6:55written by humans or with the help of

6:59AI, should be secure and high quality by

7:03default. Not as a trade-off to juggle,

7:07but as a natural outcome of how you

7:09build your software. Quality shouldn't

7:12be something that you bolt in at the end

7:14of your process in a different browser

7:17tab in a different login.

7:20We're building it directly into the

7:22platform at every point in the developer

7:24experience.

7:27Our new product, GitHub code quality,

7:30provides that built-in platform quality

7:33experience so that every repository,

7:36every team, and every line of code meets

7:39the highest standard.

7:43I'm excited to show you a live demo, but

7:46as a former developer myself, first I

7:49have to pontificate a little bit on the

7:52meaning of quality.

7:54When we first wanted to start building

7:57this at GitHub, we posed this question

8:00to all of the internal developers. What

8:03does it mean to have highquality code?

8:06And I've been in this industry for over

8:0910 years now, and even I was surprised

8:11by the intensity of some of the

8:13responses.

8:15The truth is that at the end of the day,

8:18quality is super important to everybody,

8:21but there's no one formula to say, well,

8:23this code is high quality and this code

8:26is not high quality.

8:30It turns out to nobody's surprise that

8:32the meaning of quality is highly

8:35contextual.

8:36A 50-year-old medical devices company

8:39maintaining a legacy C++ codebase is

8:43going to have a very diff different

8:45definition of quality from a 10person AI

8:48developer tools startup.

8:54For the sake of being able to develop a

8:56product and not just a manifesto about

8:59quality, we landed on a couple key

9:01categories that kept coming up in our

9:04conversations internally and with

9:06trusted design partners.

9:08Maintainability.

9:10How easy is it to change and add to this

9:13code over weeks or months or years?

9:17Reliability. How likely is it that this

9:20code is going to crash and page me in

9:22the middle of the night?

9:24test coverage. How confident am I that

9:26this code is doing what I think it's

9:28doing? And there's this kind of new area

9:32that's a little bit more nebulous that

9:34we're still working to figure out, which

9:37is along the lines of how good are my AI

9:41agents at producing what I am asking

9:43them to produce.

9:48Our public preview of GitHub code

9:50quality gives you repository level

9:53summaries of maintainability and

9:56reliability and we have test coverage

9:59and other AI stuff coming later on our

10:01road map. Our quality findings are

10:04calculated using a mix of deterministic

10:07static analysis and LLM power detection

10:11which brings you the best of both worlds

10:13for a more cohesive quality summary.

10:17And I wouldn't be able to call myself a

10:19GitHub product manager if I did not call

10:21out that we have our world-class

10:23co-pilot to generate fixes for anything

10:27that we find in the quality tool. When

10:30you're with GitHub, whether it's for

10:32security or quality, found means fixed.

10:38So now I'm going to switch to my live

10:40demo. I'm done. I'm getting off my soap

10:43box about quality and I'm going to show

10:45you what we actually have live in the

10:46product today.

10:49I just need to make sure that my

10:51monitors are sharing the right thing.

10:55Which one do I want? Not that one. Not

10:58that one. This one. This is where we're

10:59starting.

11:01So suppose that I'm a developer. Going

11:05to dust off my developer hat. I'm going

11:07to jump into some coding

11:09here. I'm pretending that I'm working at

11:12this

11:14uh company that makes an analytics app

11:16for F1 racing. My company's called Rev

11:19Litics. In this app, we're surfacing

11:22real-time stats for drivers so that F1

11:24teams can track pace, plan pit stops,

11:28that kind of thing. Maintaining uptime

11:30and stability for our app is very

11:32important to us. If we have downtime in

11:35the middle of a race, we're going to

11:36lose a lot of customer trust and it is

11:38not going to be good for our business.

11:42This feature that I'm making is a panel

11:44in a web app where we estimate a

11:46selected driver's fuel usage and then

11:49that way we can give an estimate for

11:51when they might need to refuel which is

11:53going to impact their pace.

11:56So if I look down at here, I can see

12:01that the code quality bot has found a

12:04couple problems.

12:06Here is an example where I am trying to

12:10in JavaScript use uh the keyword this

12:14and calling it to something before I've

12:16actually called it super constructor,

12:18which is an error and is probably going

12:20to make my app crash. Clearly, I wrote

12:23this code before I drank my coffee.

12:26In one click, it's already given me the

12:29suggested fix for it right in line. I

12:32can commit the suggestion directly to

12:35my pull request.

12:41From there, it's going to rerun based on

12:45the commit that I just made.

12:49It's going to check if there are any

12:50others that are left. I can see that

12:52there are some other errors that it

12:54found as well. And if I scroll all the

12:56way down, I can see that it's it's

12:59blocked right now because my

13:00administrator has set up a rule set

13:04where I can't merge this pull request uh

13:07unless I have a code quality result set

13:12and that it doesn't contain any errors

13:15or warnings from code quality.

13:19In this way, I can make sure that the

13:22code that is newly being added to my

13:24codebase meets the quality bar that my

13:28company is going for.

13:31Now that we've maintained the bar of

13:32quality for the new code, though, we

13:35want to check out what are all of the

13:37existing quality problems that might be

13:39causing us headaches.

13:41And for this

13:44I can go over to my code quality tab

13:48which is currently under our security

13:50tab with our security tools. We have two

13:53different pages for code quality. The

13:55first one is this AI findings page. Here

14:00the LLM has scanned all of the files

14:04that I have most recently changed within

14:07the repository and it's giving me

14:10recommendations for those entire files.

14:13That way developers can check out all of

14:16the existing quality debt that might

14:19exist within the files that they're

14:22already working on so they can minimize

14:24context switching.

14:26Today, it can scan just the most recent

14:29pushes to the repository. In the future,

14:32we're hoping to expand it to scan entire

14:34repositories. The LLMs can't quite do

14:37that yet.

14:39So, if I look at this example

14:43in this file, I can see that the LLM has

14:46highlighted three potential problems and

14:50suggested some changes.

14:52I love that this first one is about

14:54inconsistent indentation. It turns out

14:56that LLMs care about tabs versus spaces

15:00maybe as much as some developers do.

15:03Once I've reviewed these changes,

15:06if I like them, I can either open a pull

15:09request with the direct diff as it's

15:11presented to me here, or I can assign it

15:14to copilot because I have a copilot

15:15license. in doing so is going to have

15:18copilot coding agent take all of these

15:21changes, apply them within a pull

15:23request and start iterating on them if

15:26it's finding that my build is failing or

15:28that other security problems or things

15:31are being surfaced.

15:35So I'm going to go ahead and assign

15:37those to copilot

15:40unable to create PR. I swear this demo

15:43is doomed.

15:44I tried this demo on the universe stage

15:47and this is exactly what happened as

15:49well. So that's how you know it's real,

15:52right?

15:54Next up, I'm going to go and I'm going

15:56to look at our standard findings page.

15:59Oops, I don't have to do that yet. This

16:01is the standard findings page. I can

16:03just click on it. Here I can see all of

16:06the findings that are coming from our

16:08deterministic code scanner tool. Here I

16:13get a maintainability score and a

16:15reliability score which are based on the

16:17number and severity of each of the

16:19findings that it has surfaced across my

16:23files. Here all of the findings are

16:26grouped by their rule. So each of these

16:29titles is a summary of what the problem

16:31is. All of our code quality related code

16:36rules have been tagged with either

16:38reliability or maintainability. So, we

16:41have them spread over these two

16:43categories and they all have a severity

16:45of note, warning, or error.

16:49When I'm first looking at this page,

16:52it's a little bit disappointing. I see

16:53that m maintainability is like kind of

16:55okay, but my reliability needs

16:58improvement which is a big problem for

17:02my company because as I said reliability

17:04is very important to us in maintaining

17:06that uptime particularly during races or

17:11other important periods.

17:14So since my reliability needs

17:16improvement, I want to look at the

17:18findings that are going to be affecting

17:20this score. With these filters, I can

17:23filter down to reliability and then I

17:26can filter down to the highest severity,

17:29which tells me that I have two instances

17:31of assignment to to constant which are

17:34causing my reliability score to tank.

17:39So when I go here, I can see more

17:41details about this problem and what

17:44might be happening, examples of the

17:46problem and what I should be doing

17:48instead.

17:50And then I have the findings themselves.

17:52In this case, I can see that they are

17:55both in the same file. If I had other

17:58ones across multiple files, I would see

18:00all of the files here. And I can see

18:03that copilot has suggested a change for

18:05me where simply instead of having this

18:08const I have a let which gives me a

18:12actual variable variable. From here I

18:15can open a pull request which will fix

18:18each of these problems.

18:27All right.

18:29So, I'm going to do some demo magic and

18:32I'm going to skip ahead. Uh, cuz each

18:34step of this, you know, requires I need

18:36to open the pull request and then the

18:38pull request needs to be reviewed and I

18:40need the code quality scan and all of

18:42that which takes a couple minutes. Once

18:45I skip ahead and once I fix these

18:48problems,

18:50what I can see here is that my

18:52reliability score has now increased all

18:54the way up to good, which is quite the

18:57improvement.

18:59And if I were to filter back down to my

19:02reliability problems, I can see that I

19:04have 15 or so left that are uh in this

19:08case all under this one rule. So if I

19:11wanted to get my reliability score to be

19:13that excellent like 100%, I could choose

19:16to fix all of these as well.

19:21All right. And now I'm going back to my

19:25presentation. So that was it for the

19:27live demo

19:30section. [snorts]

19:34I said earlier stuff about like when we

19:37were when we were thinking about doing

19:38code quality when we were like asking

19:41people what quality means, but

19:43truthfully it's a little bit silly for

19:45me to pretend that it's this like brand

19:49new idea and we just thought of it. And

19:51have you all ever considered writing

19:54high quality code? Whoa. Yeah, that's

19:58ridiculous.

20:00At GitHub, we've always believed in

20:02giving developers the best tools to

20:04support them throughout the software

20:06development life cycle.

20:08This new code quality tool simply

20:11completes our coverage of the cycle from

20:13creating secure and highquality code

20:15with co-pilot coding agent in the IDE to

20:19reviewing AI and human written code

20:21together in the pull request with

20:23co-pilot code review and now through

20:26maintaining existing code. Together

20:28these tools deliver continuous quality

20:31from the first line of code to the

20:33entire software portfolio.

20:39Coming up on our road map in the first

20:41half of 2026,

20:43we have important features like orgle

20:46level dashboards, which is this. What

20:49we're looking at here is our design for

20:52one of the tabs on the org level

20:55dashboard. We will have org level

20:57enablement and configuration so that you

21:00can turn it on across many repositories

21:03at once.

21:04and everyone's favorite test coverage

21:07metrics. Those are some of the big

21:10features that we're that we have coming

21:11in the first half of 2026.

21:14And as always, you can expect continuous

21:17small improvements.

21:19Things like improving the general

21:22usability and appearance based on the

21:25feedback that we get from all of you.

21:30To wrap things up, let's recap what

21:33makes our code quality approach unique.

21:36First of all, it's all in GitHub, so

21:38developers never have to leave their

21:40flow.

21:41Second, we combine unified intelligence,

21:45LLMs, and deterministic code analysis

21:49working together for unmatched depth,

21:52consistency, and remediation.

21:54Finally, it scales effortlessly from

21:57individual repos to global enterprises.

22:00lowering your management complexity and

22:02reducing your total cost of ownership.

22:05This isn't about stitching tools

22:07together. It's about transforming how

22:10work gets done. Every insight feeds the

22:13next commit, creating a virtuous cycle

22:15where development and quality

22:17continually reinforce each other.

22:20You can enable this tool today in your

22:23repositories for free.

22:25In the future, it's going to be a paid

22:27product, but we're still working on

22:29those details. So, we don't have any

22:30information about billing at this time.

22:33But you can rest assured that it is free

22:37for now, and there will be a lot of

22:39notice before we transition to a paid

22:41product.

22:44So, thank you. That's the end of the

22:46formal presentation and demo. Uh and now

22:49Lupita and I can come on and answer any

22:52questions uh that have come up in the

22:54chat.

22:55>> Yeah. Awesome. Thank you, Carol, for

22:57guiding us through that. I can't stress

23:00enough how authentic these sessions are.

23:02This is live. It's not AI, [laughter]

23:05right? So, we're going to be human. Um

23:07speaking of human, I I could not relate

23:09more when you said

23:11writing code before coffee.

23:14>> Yeah. [laughter]

23:15[gasps]

23:16>> Yeah. So let let's go ahead and open up

23:18for questions. I do see a couple here.

23:20So let's start with the first one that I

23:22see. Will it tell the reason why that

23:25particular code is bad and why the

23:27recommended solution will solve that

23:30problem?

23:32Yes, every every finding and fix that we

23:36have both in the pull request and in the

23:39backlog comes with a little bit of an

23:41explanation of why it is suggesting that

23:45to be changed.

23:49>> All right. How does this new solution

23:52approach critical bugs like CVEes that

23:54are released

23:56>> with GitHub code quality? We don't

23:59address CVEes directly. That is actually

24:02taken care of in uh the rest of our

24:05suite of GitHub advanced security

24:07products uh like uh code security.

24:11>> Yeah. So that would be a part of the gas

24:14pillar.

24:14>> Yes.

24:16>> And then I see you mentioned,

24:18but can an individual developer

24:22[clears throat] do this on their own

24:23open source projects?

24:26Yes, you should be able today to turn it

24:29on for your uh individual repositories

24:33and it will be um at one point it'll be

24:38it'll be paid. Uh typically the approach

24:40that we take with our GitHub advanced

24:42security tooling is that it is free for

24:45public repositories. Uh and that's a

24:47pattern that we're going to be following

24:48as well.

24:51>> Awesome. And then will it be integrated

24:53into the local IDE so the developer can

24:56detect it prior even committing to

24:58GitHub?

25:00>> We don't have any uh work in the IDE

25:03coming up on our road map. But for

25:06general IDE I would recommend leveraging

25:09GitHub copilot uh which is able to do

25:13reviews before something goes into a

25:15pull request. So you would get all of

25:17the AI based ones. Uh but we do not yet

25:20have the ability to run the to run the

25:22CodeQL queries in the ID.

25:28Awesome. We'll see if we have any other

25:31questions. Feel free to chime in. This

25:33will be available on demand. But I think

25:36one of the most exciting pieces to this

25:38new release, uh at least for me, is that

25:41it's free, right? So I say join us on

25:44this journey. uh to help shape the next

25:47generation of baking code quality

25:50natively into the developer experience.

25:53Oh, I do see someone added another

25:55question. So, you mentioned it

25:57transitioning to paid in the future.

26:00Will anyone with it enabled and being

26:02used while under the free plan be

26:04grandfathered in to be able to continue

26:06using it?

26:08That is a great question because I do

26:11not have an answer to it right now,

26:13[laughter] but I would recommend uh

26:16keeping an eye on our releases and

26:17notifications like Lupita mentioned. And

26:19there will be a lot more information

26:21about that transition and uh potential

26:24future costs that come down um well

26:27before we start trying to bill for it.

26:31>> Yeah. So I I will say be sure to opt in

26:34to our notifications, GitHub

26:36specifically. There is Microsoft optin

26:38as well. Uh but when you do that

26:41basically you're tuned in to what's

26:44coming next. And honestly um I I see

26:48another question here. Will we be able

26:50to start running across entire repo

26:51versus new commits? I'll let you answer

26:53that.

26:55>> Yeah. So today the code deterministic

26:58analysis does run across the entire

27:00repository. Um and then the AI scanning

27:05is the one that is limited to new

27:08commits. That's a problem that I mean

27:11generally in the in the industry and

27:13across GitHub we're working on. So I

27:15don't have a timeline for that at this

27:16time but it is something that we are

27:18trying to expand.

27:22>> I love all the questions. I love these

27:24to be interactive. Uh I always say I'm

27:27going to hand out prizes uh for the

27:29folks that raise their hand. So, I just

27:32I need more viewers to get that ability.

27:34So, share this with your friends,

27:36support the channel. Look, I got we got

27:38another question. Uh, will I incur cost

27:41and action minutes for enabling this

27:43even though it's free?

27:44>> Yes, becauseQL today runs using GitHub

27:48actions, it will incur actions minutes

27:52>> and those are usually nominal. I always

27:54get that question too when gas is run.

27:56But think about like if you had to run

27:58your own servers, infrastructure,

28:00configuration, setup, cooling, heating,

28:03electricity, maintenance, folks, you

28:06know, physically securing the place. I

28:09think action minutes aren't that big of

28:11a deal. But that's just my personal

28:12opinion. Um, always happy to chat if you

28:15have an interest in or a need in

28:17building a business case internally for

28:19that usage. That's what I'm here for.

28:23Yeah. Any other questions?

28:31All right.

28:33>> Uh, Lupita, one thing I I want to

28:35mention just at the end here, um, as

28:38people try it out, if you have feedback,

28:40please leverage the GitHub community, me

28:43and my development team are like

28:45watching it like a hawk to pick up on

28:48any bugs that people are experiencing,

28:51um, any weirdness, any feedback that you

28:53have. If you go if you go in there, we

28:55have like this one announcement that has

28:58a bunch of people with feedback on it

29:01and you can see me and my team like

29:04responding very actively within that.

29:05So, we love hearing feedback and yeah,

29:09we want your feedback. Give us your

29:10feedback. Uh it's how it's the only way

29:13that we can really be sure that we're

29:14making a great product. Uh and we want

29:16to make a great product.

29:18>> Absolutely. And that question about how

29:20does this compete with sonar cube, this

29:23whole project actually came about

29:25because of the demand and the requests

29:28from our customers. I personally had

29:30like 31 accounts come to me and say,

29:31"When are you guys going to do code

29:33quality?" And it's interesting because

29:35code, the engine that runs the static

29:37analysis security assessments, actually

29:40started as a code quality solution way

29:42back in the day and then involve evolved

29:45into security because of how powerful it

29:47was. And so that being said, I I can let

29:51um Carol here talk about the the details

29:53on the technical side, but at least from

29:54a a business perspective, consolidation

29:57is key for a lot of organizations. While

29:59I always say there's no silver bullet

30:00and it's good to have multiple kind of

30:02solutions to provide coverage where

30:04necessary, um we still want to be at

30:06least able to provide that option as

30:09part of the overall uh portfolio of the

30:12solutions we provide. So Carl, I'll hand

30:14that to you.

30:16>> Yeah. Uh, so in terms of how we compete

30:19with Sonar Cube, uh, I'm going to give

30:23you the awful businessy response and be

30:25like, you are the only one who can

30:27really decide whether or not we compete

30:30well [laughter] with Sonar Cube. Um, if

30:32you're using Sonar Cube today, uh, I

30:35would pro and you're interested in the

30:37GitHub code quality tool, I would

30:39encourage you to turn it on and just

30:41like check it out. Uh there are a lot of

30:44things Sonar Cube does like they're a

30:46product that has existed for years and

30:47years and years. GitHub code quality has

30:50existed for like a couple months and

30:51like I mentioned like we don't even have

30:53test coverage yet. So it's something

30:55that we're working on. But at the same

30:57time we're not expecting during this

30:59public preview phase that it's going to

31:01be replacing any tool for you. But try

31:04it out side by side. We would love to

31:06hear from you the places where we're

31:09falling over that Sonar Cube is

31:11excelling for you and how we can try to

31:13close those gaps.

31:16>> Yeah, absolutely. I always say um with

31:18with quality, right? There is, as Carol

31:22mentioned, no sort of right answer. It

31:24means different things for different

31:26organizations.

31:27And so what we're aiming to do is show

31:30you what our philosophy and approach to

31:33this problem looks like. But because

31:36we're doing this sort of private

31:38preview, you have an opportunity to

31:41contribute and have a say in what that

31:44looks like. So, do not miss out on that.

31:47As you saw, we do not gatekeep.

31:50[laughter]

31:50We are here to collaborate. There's no

31:52room for compromise.

31:54>> Um, you'll be able to replay this

31:56session or re-watch it on demand on

31:58YouTube. And for the question about

32:02security based cost, um, if you are

32:04working on an open-source project, all

32:07of our security solution is free for

32:10open source for anything that's

32:13obviously private or enterprise or um,

32:16not being open to the public. Uh, there

32:19is a cost associated with that, but if

32:21you want to pay me offline, if you're

32:22just working on a personal project, we

32:24can see what options we have there. That

32:26should help answer that question. And

32:29then the last question I see is I would

32:31love to have those features in Sonar

32:33Cube to exist in [laughter]

32:37love.

32:37>> All right, sounds good. I'm just going

32:38to rebuild all of Sonar Cube inside of

32:42GitHub and then everybody will be

32:44satisfied. [laughter]

32:46>> Love it. We [snorts] have AI. Don't they

32:49say it's easy? [laughter]

32:52>> Yeah, I'll write it up. I'll write that

32:53up in an issue. Please add all of the

32:56Sonar Cube features and then I will

32:57assign it to C-pilot coding agent and

33:00co-pilot coding agent's getting pretty

33:01good. I bet it would give it a good a

33:03pretty good shot. [laughter]

33:06>> Totally. I love that. Well guys, this

33:09has been incredibly awesome. I don't

33:11know if there's anyone else that has

33:13anything in the chat. I absolutely love

33:16the interaction. Thanks y'all.

33:21If not, I think we can wrap it. Um, as

33:24mentioned, this will be able to be

33:26replayed on demand on YouTube. We'll

33:29have another session coming up. Again,

33:32like, subscribe, share, um, and add any

33:35topics that you want to learn more

33:37about. Uh, we're always happy to bring

33:39whatever to the table that makes sense

33:41for the folks watching.

33:45All right, thanks everyone. We'll see

33:47you next time.

33:48>> Thank you so much.

33:50>> Bye.

34:01Thank you all for joining and thanks

34:03again to our speakers.

34:06This session is part of a series. To

34:08register for future shows and watch past

34:11episodes on demand, you can follow the

34:13link on the screen or in the chat.

34:17We're always looking to improve our

34:18sessions and your experience. If you

34:21have any feedback for us, we would love

34:23to hear what you have to say. You can

34:25find that link on the screen or in the

34:27chat. And we'll see you at the next one.

34:45Hey. Hey. Hey.

35:07[music]

35:22I feel [music]

35:24you.

This transcript was generated from the captions YouTube publishes for this video. Get the transcript of any YouTube video atfreeyoutubetranscribe.com: free, unlimited, no sign-up.