Hacker News Reader

Comment by faust201 | Hacker News Reader

Would it be possible to run AI/ML on an existing jira/github large project (k9mail/KDE(gitlab) and show/prove this can be useful/better etc? Thanks

View on HN

If you are suggesting OP do this to demonstrate that their assertion is correct - that AI generated issue titles are objectively better - this is a Good Idea.

I have to guess /hope that they did this already; who would make a time investment like this without first proving that the product has value? Anyway I certainly wouldn’t even pilot it without some proof.

That’s a fun idea. I’ll have to see if I can run it on our Jira and Confluence instances to get better (or even minimal) ticket descriptions from the requirements.

You absolutely can if you add the context from correlations to Product and CX, brand tone and other bits of RAG.

You can study it all by graphing text extracts with NetworkX before you pay for an off-the-shelf LLM to provide a false sense of confidence.

The exercise will help you build the prompt regardless of what tech you use.

We haven't yet modularized the AI features of our product. Currently, it's not possible to use the AI features on top of JIRA/GitLab. Our AI capabilities have achieved better accuracy and output because we have more control over the infrastructure and the product's foundations.

Additionally, incorporating agents and managing their workflows require a distinct set of metadata and separate workflows, which we are still in the process of exploring.

Since one can self-host Tegon, do you have some importers working? That way I could import some Jira issues and see it in action on our own data.

If you have an API to manage issues, that could work as well. Would need that anyway for custom build integration etc.

We are in the process of building those scripts. As we are working with companies for migration we started to improve these scripts to cover more edge cases. These should be out sometime soon.

Otherwise, we have a public API to do CRUD operations for all important entities (issues, labels etc). We are working to get an openAPI spec once that is out we should have all the APIs added to our https://docs.tegon.ai

Sounds great, will keep an eye out.

Can you give me an idea of what you are using AI for currently?

Edit: Never mind, found the list you posted elsewhere.

Current AI functionalities: 1. AI-generated Titles 2. Smart Delegation 3. Duplicate Detection 4. AI Summarization 5. AI filtering 6. Automated Triaging

We are almost done working in 2 other areas 1. AI assistant while create a new issue ensuring the issue has enough information 2. Chat assistant to interact with the tool

Haha, same one. Thanks!

Probably not, from my experience. Certainly not for large installations. There's just too much to glean from a 20 year old Jira (Components, Comments, Issue History, Transitions, Links..) You can build a graph database from all of this (plus the metadata and other contextual stuff, like related code) which is what Atlassian did with their "Teamwork Graph".

Someone else's mystery machine is fine for a start, but if you want to train and test (and validate) your assumptions or do your own experiments, these AI features don't offer much.

This is a bit of an extreme ask, that's basically an entire product in its own right.

All the data is there, you just need to connect to it with a new class of BI tool. It's coming.

That's missing the point. The root comment is asking for a different product (one that integrates with Jira) than the one offered by the op (one that replaces Jira). The manner in which such a product would exist isn't the concern of my argument.

That's not how I read it, but I don't see any case where offloading the entirety of IP in Jira to a third party makes any sense.

I am curious, why would you like to do this?

Reply on news.ycombinator.com