Ask HN: Who wants to collaborate?

I've been interested in conversational bots and how they might be used for purposes that are more fun than customer service. I'm specifically interested in ideas that are silly, or emotionally salient, or weird, or anything that's not productive and not commercial.

I have been poking around with some stuff, but mostly just ideas in my head. So for "collaboration" I mostly imagine getting ideas from each other, trying each other's stuff out, all very low commitment. Like you know when you find someone who when you talk about ideas together and you both come away more excited than when you started? If there's any community where people do hang out like this that would be cool, though the tedious and tired commercial chatbot discussion seems to dominate.

Some thoughts I've had:

1. How do you make a really creepy bot? (Barring profanity, racism, etc) Sometimes people worry about creepy bots but I honestly am not sure how a bot would be creepy. Making one on purpose might be a way to find out

2. Bots are often very reactive without any will of their own. What kind of personality or (imagined) motivation might make a bot proactive? Would that solve some of the "I don't know what you mean" problem of a chatbot? Like instead of a bland response when the bot doesn't understand, it can instead push back on what it cares about.

3. In this vein I've thought about a depressed bot, and your job is to find a way to cheer it up.

4. User-programmable bots. I've been thinking about this in user-buildable VR worlds. The worlds are all so empty. Can we make it feasible for a person to create their own NPCs?

5. Can a bot be loving and supportive? If it's just a façade, how good of a façade is enough to generate emotional response?

6. How do mediums like text vs voice impact this stuff?

Out of curiosity, why is roblox toxic?

(idk much about it)

do you have anything we can see ?

I want to do something similar, only...for UI frameworks...

So on the sidebar you have components...one is a sidebar right? so you click on it..on the right pane it shows implementations on the top of the pain is a dropdown so you can switch from say. MUI to AntD to Vuetify, etc... some are narrow scoped frameworks (Vuetify for example is only Vue), some are Wide... Tailwind Kit for example or Daisy UI... they target tailwind but are JS Framework agnostic...

So under each implementation of each component might be code for Rails, Laravel, Django, views and maybe even some sort of api integration where you can have a webhook that can either connect to your dev and push the code to the right places, OR it could put it in a queue if your local dev can't receive incoming normal http requests, you could just use a cron job or manually import the code from your localhost backend...

I'm imaging this sort of being like a mixture of low-code meets Storybook with already developed ui components, and you could even import full templates, etc... So for backend devs w/ an idea 90% of the UI stuff is just drag/drop etc.... you could even have a way to combine elements somehow so you can build up a whole page of elements...and some sort of marketplace where people could share designs and maybe charge like $1 for the code snippet MIT license, but access to the code is $1.

Edit: It'd also be cool to extend this to data-layer..so say I am designing my migrations/models... and I want an Employees table....well maybe I type in employees and it organically lists all the columns previous people have used for employees tables and ranks them by how often it was used... so you can basically have most of the data points that are pretty common across similar apps be shared and almost done for you...and then you can export to django, rails, prisma, hasura, laravel, etc...

I'm in a volunteer role as CIO for a nonprofit focused on human trafficking, and we're building a new investigations and case-management platform using Svelte/TypeScript/SvelteStrap, Rust/Rocket, and Postgres (and likely Neo4j at some point) because the options out there are either poor or very expensive. Once we have something working and solid, we're going to open-source it and provide it to other groups in similar spaces.

This project is a great opportunity to learn or expand your skills if any of the stack is interesting to you - a big part of our goal is to help build up the technical skills of anyone who wants to dive in somewhere.

We've got a number of investigators on staff and are starting to get a decent feedback loop with them to iterate on what we're building.

We've also got some unique and useful ideas that haven't been implemented before, and if things go well, we really think this could give a kick-in-the-pants to Maltego and Palantir's offerings, amongst others.

If anyone is interested or wants to know more, please let me know!