hazzadous

I've tried generating some dingbat puzzles with Gemini, they are comically bad most of the time with no subtlety. Even for existing dingbat phrases it didn't make good ones, which I would have expected it to have been able to just copy prior examples.

Same with cryptic crosswords, not very good. Sure there's some strategy that would work like providing some tactics to use.

built a tool that analyzes who's starring your github repo. paste a repo url and you'll get charts on role and seniority, top languages and frameworks, companies represented, where stargazers are located, and an aggregate feed of their recent blog posts so you can see what your community is thinking about.

it pulls the repo's stargazers, looks at their public profiles, and returns everything as aggregate insights.

curious what would make this more useful, or what other slices of the data you'd want to see.

Yes the coverage is not going to be near 100% so you’d get a lot of anger with this use case. There are some libs that will check via smtp if they are likely reachable, and some services. Is “verify your email” and option ie sending an email on signup as part of the signup flow an option?

let me know what use cases you have. i can update a tweak accordingly if it makes sense. anything goes. almost anything. nothing illegal.

thanks for the feedback

It’s doing real time research so might be a bit tricky with a hash. Can you see a way around that?

Pretty similar except maybe you’ll get lots more nulls judging by the other comments! Cheaper but nulls. Will need to work on the recall a bit. But also potentially based on use case feedback maybe look at other niches and features

Hmm surprising you’re not getting anything for public emails. Hit rates can be low but not expected in those cases. Are you using anything else currently for the same task

Let me know how you get on with the testing. And if you have a use case let me know and i can look at optimising for it

Hah some high value info right there. Thanks for trying it out.

Yes this is open source intel based so does struggle on that. Do you have a specific use case in mind?

Do you have an idea of what would be fast enough and what use case you’d need it for. Definitely not peak recall right now for sure.

Some work to do on precision and recall. Thanks for testing it out. Hah yes 2024 llm sounds right.

I’m interested in what you think you’d use it for if you don’t mind sharing, and what you would benchmark it against.