41–50 of 76 posts
I really like the duplicate detection feature; that would save me so much time and prevent pain down the line with product backlog refinement.
If you get a chance to try, we'd love to hear how the duplicate detection stacks up against your expectations
I think project mgmt software suffers from the same problem as project managers - they think they should be in charge, visible and front and centre I disagree (esp about the human ones) I want project management that fills itself in based on work people do, not work people tell the tool they have done. I am dubious that putting ticket numbers in commit messages is needed - let alone IRC, emails, phone calls meetings,…
It's super compelling to hear about different pain points like this. The idea of automating acceptance criteria is indeed tough but absolutely possible to imagine. This is outcome-oriented in the best possible way!
If you are using different versions of GPT, in particular GPT-4, how are you controlling costs once users' projects/backlogs grow in complexity?
I don’t know of too many tools that are offering unlimited free calls to OpenAI, which means that all cool LLM-enabled features are price-gated or premium or otherwise limited. It's a bummer to restrict that value. Our bet is that LLM pricing will follow a Moore’s Law-style pattern, at least for a while, that will mean that we can offer better and cheaper LLM-enabled features over time. So in short, we're subsidizing some of the costs now on a longer-term bet.
That said, we can be smart about how we do things technically. We embed, compress, and omit stuff as much as possible to minimize tokens.
Also, we actually just completely fail to handle some things (something like reprioritizing a backlog of 10k tasks just wouldn't work for us right now) so we do hard cap some actions.
I mean, who the hell wants to do a list of tasks created by an AI?
If you are using different versions of GPT, in particular GPT-4, how are you controlling costs once users' projects/backlogs grow in complexity?
Great question. We're taking an unorthodox approach here. I don’t know of too many tools that are offering unlimited free calls to OpenAI, which means that all cool LLM-enabled features are price-gated or premium or otherwise limited. It's a bummer to restrict that value. Our bet is that LLM pricing will follow a Moore’s Law-style pattern, at least for a while, that will mean that we can offer better and cheaper LLM-…
This is something PMs are supposed to do?
We tried JIRA, Linear, now we are using Github projects. It is too much effort to keep it up to date when things are fast moving. Everything becomes stale.
At first look, it feels like this too will suffer the same.
The best tool in this space should eliminate the need of specialized PM.
PS. Github project has become surprising better in recent times.
Earlier quoted context omitted.
Great question. We're taking an unorthodox approach here. I don’t know of too many tools that are offering unlimited free calls to OpenAI, which means that all cool LLM-enabled features are price-gated or premium or otherwise limited. It's a bummer to restrict that value. Our bet is that LLM pricing will follow a Moore’s Law-style pattern, at least for a while, that will mean that we can offer better and cheaper LLM-…
Regarding embeddings - I am assuming you are using ada-002 or have you moved on to 3-small already? Do you have a particular strategy for migrating embedding models other than re-embed the whole dataset? And lastly, what is/are your vector store(s) of choice? I am not quite sure of your scale but I have found that north of 50 million vectors a lot of the current options get a bit weak in the knees, especially if you…
- Current plan is to re-embed everything but I'm very open to better ideas there haha. Is there a better way?
- I've heard some similar stuff but we haven't run into it yet. What are you working with?