Hacker News Reader

Comment by mdaniel | Hacker News Reader

When I read the announcement, I thought they were donating the use of CodeGuru to projects (which I presumed would be limited to open source ones) to (1) improve the code hygiene of the world (2) gather more data for their machine learning models (3) publicity for CodeGuru

But as best I can tell, this is the worst of all worlds: I have to already have an AWS account to both create the competition and also to point CodeGuru at the repos, then I have to pay for running CodeGuru on any potential PRs, but then also separately create a BugSnag account seemingly just for this stunt. And in the end I ... get Internet points on my BugSnag account?

I fail to see how anyone but AWS benefits from this, unless a participant happens to be one of the top competitors who accumulated enough Internet points to get a free trip to re:Invent

View on HN

The companies that put up the contest benefits. For a lot of people at a lot of companies, a trip to re:invent is a huge prize. If your job is to fix bugs anyway, now your company just incentivized you to do it more and in your off time to win that prize.

Also, the bug buster benefits if they get up high on the leaderboard, because job recruiters will most likely use the leaderboard for lead gen. If you're looking to make a job move, moving up on the leaderboard will help.

> Also, the bug buster benefits if they get up high on the leaderboard, because job recruiters will most likely use the leaderboard for lead gen. If you're looking to make a job move, moving up on the leaderboard will help.

I already get plenty of recruiting inbound without having to donate my time to fix bugs for Internet Points.

What you're saying is that this is a worthy use of time for people because they'll maybe move up in the leaderboards which then might lead recruiters to reach out to them.

And after all that, you have to do a phone screen, technical phone interview, a full-day onsite, plus whatever other hoops you'll be made to jump through to change jobs.

Yeah, no. Hard pass. Seems like a huge waste of developers' time.

> I already get plenty of recruiting inbound without having to donate my time to fix bugs for Internet Points.

You're not the target audience.

And also, you wouldn't be donating your time. The whole point is that a company puts up a contest for their own employees, who are presumably getting paid to participate.

So by participating in this, a company is making their own employees more attractive to recruiters?

No the company is getting some extra structure to their bug bash with some prizes provided by Amazon, as well as being able to recruit other people by pointing out how good their current people are (because they are on the leaderboard).

Everything about this seems like a sketch out of The Office (if they were a software company).

the target audience are people stupid enough to work for amazon for free. Amazon, fuck you pay me.

Why bug leaderboard is important? Don't I have few ways already to showcase my profile?

1. Stackoverflow to show my expertise

2. Any Open source contribution that shows up in my github profile

How are you employed if you won't fix bugs. These types of devs are so annoying to be around. Crank out weird esoteric hard to read code using every weird meta feature / reflections / introspections they can. Then "hard pass" on fixing bugs as a waste of time.

You do realize most developers are asked to fix bugs anyways, regardless of any silly competition?

Their pitch to Organizations is that

1) They may be able to help you save money 2) They may make it more fun / easier to solve bugs.

Pitch to developers - they'll give you some perks for doing your job.

Note that I also think it's lame because I don't think code guru is that good and don't care about re"invent

> How are you employed if you won't fix bugs. These types of devs are so annoying to be around. Crank out weird esoteric hard to read code using every weird meta feature / reflections / introspections they can. Then "hard pass" on fixing bugs as a waste of time.

You misread pretty much everything I wrote. My original comment wasn't an explanation for why fixing bugs in general is a "waste of time". My opinion is that uploading your code to this service that uses ML to automatically generate bugs for your team to do a bug bash on is a waste of time.

Our bug backlog has hundreds, if not thousands, of bugs already. We don't need to rely on ML to tell us what to fix.

We have real issues that customers are facing that need immediate attention. This is just a total distraction and a waste of time.

Unless I'm completely misunderstanding how this works, this isn't "give you some perks for doing your job." It's getting perks for doing your job, for someone who isn't paying you. And that's what the problem is.

If you want me to fix your bugs, pay me for it. Don't do this gamification nonsense.

Edit: apparently I did misunderstand how this works. It seems like the "Bugbusters" are just a company's normal developers. It still seems kinda dumb, but not as exploitative as I initially thought.

They're marketing page isn't super clear, but it's adding structure for a company's existing bug bashes. You set up a bug hung for your own developers on your own code.

Then why in the world make Internet Widgets, Pty users create yet another login on some random .aws property that (as very best I can tell) does not support any kind of SSO auth that an outside organization would use: GitHub/GitLab, IAM/Cognito nor OIDC/SAML

Yeah, they are just trying to get devs to try their tooling. If tool was amazing, this would be great. But it's hardly the big exploitation people are describing here though it's a bit true that bug fixing is less fun for some and so is often ignored.

That said, other providers of this type of tooling might take some ideas from something like this.

> We will pay you with exposure...

> Also, the bug buster benefits if they get up high on the leaderboard

There is literally another story on the front page [1] about such "leaderboards" and what happens when people game them to move up.

I would not advise anyone to participate in such a contest, it sounds more like a race to the bottom while providing valuable training data to AWS for free.

-------------------------------------------------

[1] https://news.ycombinator.com/item?id=27629366

That has nothing to do with this though. The global leaderboard is for fixing bugs in your own company.

> The companies that put up the contest benefits

> if you’re looking to make a job move,…

Why would companies allow their best developers to be poached? Unless of course there’s a way for companies to keep their developer scores private…

The risk to the company is pretty low that their employee would show up on a global leaderboard. And even if they do, they can give them a raise or other benefits. It helps them determine the market rate for their employee.

> The risk to the company is pretty low that their employee would show up on a global leaderboard.

This invalidates your other point that it benefits developers. Why would a dev sink countless hours for a “very low” chance at being on the global leaderboard and hence moving jobs.

You can’t simultaneously posit it’s good for both companies and developers.

> This invalidates your other point that it benefits developers.

Not at all. The company undervalues their employee and sees little risk in them showing up on the board, and the employee overvalues their skill, thinking that showing up on the leaderboard is a shoo-in. Maybe the employer is right, maybe the employee is right, or maybe both or neither is right.

Also, the employee does it because they get paid to hunt bugs regardless. It just helps the company by putting a structure around what they are already doing.

> the employee overvalues their skill, thinking that showing up on the leaderboard is a shoo-in. Maybe the employer is right, maybe the employee is right,

Come on, this is just being coy — you and I both know this is a terrible deal for employees — as you write earlier, there’s a very low chance that a random dev will appear on the global leaderboard. What’s more, the best devs won’t even play this game, so it’s left for juniors and people trying to “prove themselves” to trample over each other for peanuts.

Not to single you out, but I wish influential tech folks like you would speak up more about the cynical exploitation that’s going on here in plain sight.

I'm not sure where the exploitation is. The developers are already getting paid to fix bugs. Now they have some structure and added incentives.

The structure already exists. It’s called your bug tracker. By building tools to gamify this process, it’s clearly incentivizing devs to put in more hours. I have no doubt that the global leaderboard will consist of devs that would’ve clocked 100s of hours above their regular hours. For very little benefit to the employee.

Let me put it you this way - would an engineer at Netflix ever “play this game” to win ? Would employees at Netflix even put up with something like this? If not, why would you think employees at random tech co should?

> would an engineer at Netflix ever “play this game” to win ?

I think you're missing the point. The leaderboard is an extra bonus. The main thing is squashing bugs you were already squashing. But yes, there are definitely engineers at Netflix who would play this to "win", just for the love of the game.

But let me relate another similar story -- when Amazon released the Deepracer, they added a leaderboard. All around the world, people got paid by their employers to compete. I think the first winner worked for Samsung and was paid her full time salary just to compete on Deepracer. They saw it as a point of pride that one of their employees was on the leaderboard, with enough value to pay her to work on it full time. I'm pretty sure that is what they are going for here with the global leaderboard. Everyone wins in that situation. The company gets to brag about having the best bug squasher, the employee gets to get paid for squashing bugs, and Amazon gets publicity.

You didn’t address whether engineers at Netflix would put up with this.

> I'm pretty sure that is what they are going for here with the global leaderboard

Sorry, I’m confused. This seems like a change in your position? Earlier you argued the intent is for employees to be able to showcase their skill to recruiters… you also responded to someone else who said they already have recruiter attention with “you are not their intended audience”

> paid her full time salary just to compete on Deepracer

Wait, are you seriously comparing a global competition in cutting edge tech to squashing bugs … in your own codebase? Come on.. there’s no comparison. If anything, a company would be semi embarrassed that they have so many bugs in their pre existing codebase.

I’m asking you in good faith. Why the contortions to defend such a product? I see from your profile that you are an investor and advisor. I would seriously try pushing this product to your companies and see how they react. If they do adopt it, run an anonymous survey of the devs .. then I would love to see you argue with a straight face that it “benefits developers”

> Sorry, I’m confused. This seems like a change in your position?

There is a benefit to both employees and employers. Being on the leaderboard benefits the employee because it raises their profile. Having an employee on the leaderboard, as an employer, also raises the company's profile. It's like how companies send people to conferences to talk. It benefits the employee by raising their profile but also benefits the company.

> I’m asking you in good faith. Why the contortions to defend such a product?

I'm not really making contortions. This product is aimed at large enterprises. It's not aimed at startups (so I would never suggest it for any company I invest in) and it's not aimed at companies that are already well known (FAANG). It's aimed at the 1000s of other large enterprises out there who want to incentivize their employees with some prizes to do a job they're hopefully already doing while also possibly raising the company profile a bit by having people show up on the leaderboard.

But it definitely solves a problem for it's target audience, which you don't seem to be a part of. I am not their target audience either. But I see the value to their target audience.

> I'm not really making contortions

I would politely disagree — in any case the thread is there for all to judge.

If you sincerely believe that this is a product that should exist and will positively impact everyone involved, I can only hope you’re right [0]. In the end it’s all up to your conscience, but I do wish our leaders in tech were a bit more sensitive to seeing the broader picture.

[0]: pour one out for all the naive devs that will no doubt sink 100s of hours to be on the leaderboard for little benefit to them.

> For a lot of people at a lot of companies, a trip to re:invent is a huge prize

Only for cheapo companies. Companies should (and do) _pay_ their devs to attend re:invent. I get your point though. For naive devs, it’s like getting to go to a concert. But it’s quite cynical for AWS to exploit this

Work for us for free! If we like the fix you made MAYBE we will pay you! Yay!

One part I'm still not 100% clear about after looking at the site and even your comment. Would people be running CodeGuru against their own internal AWS code, or against existing public open source Github projects, and sending PRs separately from AWS?

I didn't set it up beyond creating a skeleton "Event", cause I have zero interest in paying real money to kick the tires on this, but as I understand it one would add existing repos to their AWS Code* tooling, at which point machine learning awesomeness takes place and it surfaces findings (inside those AWS CodeWhatever APIs, having nothing to do with BugBust)

If those findings are imported into a BugBust Event [which is defined by the tuple of (start date, end date, scoring, and repos in scope)], then one can (invite|publish) a link to the BugBust Event and at the end date the participant who fixed the in-scope issues with the highest point values claims the cited prizes. Further (again, AIUI), at the end of the *AWS* event, the BugBust account who has the most *overall* points wins a re:Invent trip

I don't believe AWS gets into the business of monitoring the PRs, only observing the CodeWhatever outcomes -- if CodeWhatever found an issue, time elapses, and the CodeWhatever sees the issue is fixed, then BugBust will look up the category of said issue, multiply it by the assigned score in any Events it is associated to, and award the Internet points to any Event participant it knows about

I think you mean BugBust, not BugSnag.

Yes, sorry, I was on my phone :-(

Quite frankly I've never seen people use AWS offerings for CI pipelines, I wonder if other products (Code* family) are any good - can anyone comment on them? I've only used CodeCommit once during a hackathon, and it seemed rather basic.

CodeDeploy fails in horrible ways and the maintainers are somewhat unresponsive. I don't use it anymore but am still subscribed to some issues which apparently weren't fixed for years...

From my experience the products work fine, but they much more complex than they need to be (and than competing products are). They have a lot of weird constraints that make it hard to be really productive.

That said, when you've setup a build/pipeline, it mostly just works as it should.

I have used CodeDeploy/Codepipeline/Codebuild over several years. Overall they work well but there can be edge cases or gotchas that show up on occasion depending what you are trying to do.

I also prefer externalizing the CI/CD process with Jenkins or Spinnaker-based setups.

I have. Their offerings are even less impressive than many of the popular SaaS CI servers, which themselves are usually (and incredibly!) inferior to Jenkins - except they maintain themselves due to all the cloud whatnot

Unfortunately surprisingly poor in quality!

You can get much better results almost anywhere else.

I'm going to chime in and say the only benefit we got from CodeDeploy was its cost was folded into our AWS account. We didn't have to go through the hassle of explaining why we need "yet another <$100/month expense" from a different company.

I appreciate this is a bad reason to recommend a service, and am not defending it. Just explaining why we wound up using it.

The service itself is not horrible, but like every other AWS service, I felt like I was definitely not in the target market: some easily implemented feature was left out and the configuration space easily allowed you to specify things that weren't covered in the (very) thin documentation and seemed to behave unexpectedly.

Many SaaS companies offer the ability to pay for themselves via the AWS marketplace so you can have everything under one EDP / bill (though AWS takes a little cut in this case).

Reply on news.ycombinator.com