leading ai company
-
Why? I genuinely think that daily delivery in my field (b2b specialized software) would be a very good practice. Why in mobile apps it's not the truth?
Because the rate is more a sign of how often problems are found, rather than how many better new things you are applying.
-
Why? I genuinely think that daily delivery in my field (b2b specialized software) would be a very good practice. Why in mobile apps it's not the truth?
It's a bit different with mass market mobile applications because of the supply chain constraints - most notably the Apple reviewing process. Your next app release may for whatever reason they feel like unexpectedly take an additional week, so do ensure that your QA is in order before releasing.
Another significant factor is the lack of control you have over the software once released - any bugs you ship may potentially be out there for a long, long time.
Web applications don't have these constraints and can as such be deployed an infinite amount of times per day. The same goes for backend services, deploy to your hearts content.
This basically means that most larger mobile applications have adopted approximately weekly release cadences, and that we've had to get very good at using feature flagging to control our software in the wild, and avoid large impact of shipped bugs.
-
Add a comment.
Commit.This is why there are a bunch of improvished people adding nonsense updates to git repositories, padding out their resume out of desperation.
It's hard to blame them trying to climb their way out of poverty, but it does harm the people trying to do work.
-
Add a comment.
Commit.wrote last edited by [email protected]I've made over 40 edits to this comment.
Me: 40
You: 1
Pathetic.
-
i wish this was fake https://twitter.com/elonmusk/status/1958854561579638960 oh my fucking god
I thought Musk was in a ketamine hole when announcing lawsuit against Apple and OpenAI. Waking up briefly to add this legal argument is not proof of not still being in hole.
-
It's a bit different with mass market mobile applications because of the supply chain constraints - most notably the Apple reviewing process. Your next app release may for whatever reason they feel like unexpectedly take an additional week, so do ensure that your QA is in order before releasing.
Another significant factor is the lack of control you have over the software once released - any bugs you ship may potentially be out there for a long, long time.
Web applications don't have these constraints and can as such be deployed an infinite amount of times per day. The same goes for backend services, deploy to your hearts content.
This basically means that most larger mobile applications have adopted approximately weekly release cadences, and that we've had to get very good at using feature flagging to control our software in the wild, and avoid large impact of shipped bugs.
Ahhh.. now that makes sense. Thank you, kind stranger!
-
Silicon*
no he was talking about the porn parody of the show
-
i wish this was fake https://twitter.com/elonmusk/status/1958854561579638960 oh my fucking god
This is kind of funny but sad, like when you see an animal accidentally kill itself.
-
pay by the update
Uh, how do you imagine that would work? Are you saying this is contract work?
Nah. It’s most likely not that direct. If the boss makes it clear that they measure productivity by blindly adhering to some dumb metric (number of updates , kloc, number of bugfixes) then the devs will find a way to max out that stat, to get bonuses, promos, etc.
As much as the author has shown his ass, this is the classic comic that comes to mind
https://www.reddit.com/r/ProgrammerHumor/comments/k5hka0/bug_free_programs_a_dilbert_classic/ -
Any update that hits the App Store and increments a version number goes through the Apple review process.
Certain updates can be done on the fly with custom or third party solutions like
https://ionic.io/docs/appflow/deploy/introBut this graph doesn’t make it clear if these updates are new binary app deployments or on the fly updates
Yeah, I haven’t done it in like 8 months so I think I was conflating Code Push with App Store updates. I do think that apps get treated differently based on the priority of the company and there is some judgement used in the scope of changes. Like I wouldn’t be surprised if Grok is never subject to the random review delays just cuz no one wants to deal with Elon throwing a tantrum
-
i wish this was fake https://twitter.com/elonmusk/status/1958854561579638960 oh my fucking god
To me, that just looks like "move fast and break things" taken to an extreme level. Quantity ≠ quality.
Honestly, I think it kinda makes the other companies look better, in terms of software quality.(I'm going solely by the chart though; I have not installed any of these apps)
-
"Look how productive I am!"
"Changes in this release: general improvements"
-
Because the rate is more a sign of how often problems are found, rather than how many better new things you are applying.
And has nothing at all to do with the AI part of the app getting better.
-
Yeah, I haven’t done it in like 8 months so I think I was conflating Code Push with App Store updates. I do think that apps get treated differently based on the priority of the company and there is some judgement used in the scope of changes. Like I wouldn’t be surprised if Grok is never subject to the random review delays just cuz no one wants to deal with Elon throwing a tantrum
Yeah I'm the lead mobile engineer at my company. We release bigger updates once a month and smaller hotfixes generally weekly or biweekly. For smaller updates we get approval in around 6 hours. They also have a way to expedite reviews in which case we've gotten like 30 minute turnaround on reviews (though that’s like boy that cried wolf, only use it if you need to push something really urgent)
-
i wish this was fake https://twitter.com/elonmusk/status/1958854561579638960 oh my fucking god
wow its getting dumber faster than the other companies what an achievement
-
i wish this was fake https://twitter.com/elonmusk/status/1958854561579638960 oh my fucking god
Having to do the most amount of bug fixes for the app (that does not run the AI itself) is not the flex you think it is
-
And has nothing at all to do with the AI part of the app getting better.
Yes, that was my point.
-
Incredibly misleading and/or stupid graphs are so funny to me. Because you ship out the most updates, doesn't mean it's the best, it means youre fixing and/or generating more bugs and issues.
Yeah, I updated my minecraft mod 20 times in a week, it doesn't mean it's a stellar mod, it's less than mediocre at best. It was primarily fixing bugs and a crash. Meanwhile the Create mod updates about once every three weeks or so on average, but that's because they properly playtest and bugfix and patch and do all that before they send out an actual update.
Two weeks is also such a random and unimpressive statistic.
-
Two weeks is also such a random and unimpressive statistic.
It's generally the length of short sprints (blocks of time where some tasks have been estimated/committed for)
If you're deploying a new version more frequently than that it's usually either putting out fires & hot fixing, or someone fucked up the pipeline and now any commit will immediately be deployed straight to prod
