Hacker News new | ask | show | jobs
by Narretz 563 days ago
> GitLab Inc., (NASDAQ: GTLB), the most comprehensive AI-powered DevSecOps platform

Oh, that description explains why the core pipeline authoring and capabilities have made almost no progress in the last few years. I actually thought gitlab still branded itself as a "classic" dev ops tool.

5 comments

I was actually looking at them as a GitHub alternative and their homepage has so much vaporware AI BS on it that’s as far as I got.

Ironically I was very open to paying for a service, but the “AI AI AI!” lost them at least one sale.

I was at a 'techtalk' recently with over 100 attendees, where Gitlab was a sponsor and the second of two talks. Before either talk the GitLab person gave a short pitch on the whole GitLab AI developer productivity vision and how great they are.

There was a short break after the first talk concluded during which about a third of the attendees left, myself included.

I am confused. What is the lesson to be learnt here? Did the last two talks look boring? ... And that is why you left? Were people annoyed with pre-talk pitch by GitLab?
People heard the AI spiel and lost interest.
This press release makes a big deal out of them counting "more than 50% of the Fortune 100" as customers, which goes a long way towards explaining their decline in favor among devs. They're not interested in your sale any more, they're only interested in large enterprises, and are apparently doing very well in that market.
I doubt 50% of Fortune 100 customers are all in on GitLab. My guess is most of their F100 customers are acquisitions that were using Gitlab and continue to use Gitlab.

When my company was acquired by $MegaCorp, I noted one of vendors was like "trusted by $MegaCorp" because yes technically, they got a check from $MegaCorp but $MegaCorp was not interested in becoming further customer.

Or one or two devs in the F100 customers made an account using their work email so they could chuck some OSS prototype code somewhere, or test something out.
Yes I've seen it a place I've worked - GitLab is pushing hard on AI right now, and I don't believe it's cheap either.

Ironically, the JetBrains autocomplete is better than their DUO plugin - JetBrains is faster and the GitLab plugin causes my IDE to completely lock up at least once a day.

Wow. Almost none of those words would have even been among my third-string choices of words to use to describe gitlab in one short sentence. “The” and “platform” might have made it.
"The other code hosting platform"?
“GitHub for if you compete with MS and someone high in your org is concerned about letting MS host your code.”

“GitHub with a worse UI except GitHub’s has been getting worse for years so now they’re both similarly bad so never mind”

“Worse gitea but with more features so sometimes it’s better”

Having used GitLab Enterprise, I'd describe it as having 99% of the features that you could ever want, but those features are generally executed no more than 75% well.
I used to think gitlab was the bees knees, but more recently there's just a lack of user awareness. They've had a open issue for years about failing a job due to not finding artifacts. The logs even say "ERROR" I've concluded they're now a marketing organization.
It's normal to Google "how to do x in gitlab" and then there is a ticker in their issue tracker to add x from 2018. GL employees all agrees that it would be great. Thrn there are 374747 label changes and no resolution until today.
While this is not a completely wrong observation, it's worse for Atlassian BitBucket. With github I have no experience.
This is common for huge open source projects. However, when you send a patch to Gitlab, they will assign someone to guide you through the process, and the patch will be merged eventually unless you bail out or they outright reject it.
That's copy-pasted from https://about.gitlab.com/ .

I love that it implies there is a more comprehensive DevSecOps platform that isn't AI-powered.

Or search. What a useless devops platform if you can't find issue by searching for words that appear in comments. Only words appearing in the issue title/description are found and this is infuriating every day.

Can't believe they'd put "AI-powered" there when it can't even be used to find exact word matches.