jrmyphlmn

when I learnt about namespaces, I included them as part of the system and they worked great! I've documented them here https://whoami.wiki/docs/namespaces

the transclusion pattern for layered visibility is something I haven't implemented yet but stumbled upon when I was evaluating mediawiki

thanks for pointing to pmwiki's approach, I'll look at how said set it up!

fair point, I've added a disclaimer to the landing page + docs to make this clearer

you can also use opencode with local models for a fully offline setup

thank you, I'm sorry for your loss

if you have photos or anything they left behind, I'd love to hear how it goes if you give it a try

sometimes the act of keeping something alive can become its own burden, I'm sorry your mother had to endure that

I'm glad to hear :)

yes! you can use it through opencode and a local model, that also works!

I did use some of my facebook and instagram exports! I was active on those platforms during college so it dug out a lot of fun stories from there haha

timeline view is interesting, I’ll hack around!

yes!

thank you :)

I understand the bittersweet feeling because I did all the editorial work for the wedding page and the first few others and I did feel like a historian trying to connect the dots after stumbling into some primary/secondary materials and spending a couple months doing all the editorial work

after I began experimenting with agents, it sped up my process that otherwise would've taken many more months for every page given that the kinds of data sources also increased over time

I did still spend significant amounts of time like a wikipedia contributor would deciding on what to keep, enhance or delete from the page based on my own personal preferences and what I was comfortable with seeing on the page

the dystopic feeling is also fair and unsettling, I think this ironically also made me realize how important safeguarding my personal data is, we leave digital trails of ourselves everywhere so a powerful agent can string them together to create a story of who you are

if one's childhood had many instances of tv playing in the background, then I suspect it's their nervous system marking that environment as their most safe/comfortable environment and that revisiting it can have a soothing effect on them?

Thank you :)

Thanks!

state machines with llm directed function calling is going to be a huge unlock

This was my intuition as well, glad you're able to resonate with that :)

One thing I’m curious about is narrowing the scope of accessible functions based on a state machine that is designed to match the business domain.

This is an interesting question, I can definitely see how state machines can help with narrowing the scope of accessible functions.

Glad to hear that!

Right, currently Symphony only supports GPT-4 and GPT-3.5-turbo since they're the only ones with native API support.

Thanks, excited to hear what you think after trying it out :)

You're right!

IIRC, only the function signatures (or descriptions) are counted as part of the context window so you could add as many till you exceed that limit. Since the contents of the function itself are not counted, your function can be whatever length.

is there any way to choose only a subject of the functions to share for a given user query?

As of now, no. I can see why this may be a problem soon since right now all functions are available for gpt-4 and each call can become expensive pretty quickly if you send like 50 functions every time.

I'm not sure how to address this yet, but I'd like to think of it as some form of fine-tuning that happens after having a few conversations. Will keep you in the loop!

That sounds great! I'm currently working on adding support for .py files as well – open to PRs :)

Thank you! You're right; the main focus right now is extraction.

One aspect I'm excited about is the possibility of rendering the JSON outputs from these function calls into UI components, as previewed here: https://symphony.run/showcase. Using a function's type definitions is a nice starting point to embed interfaces into the conversation.

Additionally, I hope to make the toolkit language-agnostic. I'd like to incorporate some of my .py and .rs scripts to make them ready for use as well. Not sure if packaging it as an npm package would go against that objective, but will definitely consider :)

a very expensive television