Ameliorate

Ameliorate(ameliorate.app)

85 points by hakkikonu 20 days ago | 25 comments

keyserj 20 days ago |

Cool to see this posted :) I'm the creator.

For those interested, I've slowed work on the app in favor of refining the app's ontology in markdown form[1], building an ontology playground[2], and building a prototype[3] to show off features that the updated ontology enables. I had some major changes I wanted to make and making it work in the real app was going to be too much effort.

I'll likely be making a Show HN for one/some of these soon^TM (in the next month or two?) but they're currently still a bit early.

[1] https://github.com/keyserj/reasoning-tools/blob/main/amelior...

[2] https://keyserj.github.io/reasoning-tools/ontology-playgroun...

[3] https://github.com/keyserj/reasoning-tools/tree/main/amelior... (not yet deployed)

jjtheblunt 18 days ago | |

first, cool idea! i have just a grammar suggestion:

> What if it was trivial ...

what if it _were_ trivial...

( https://en.wikipedia.org/wiki/Subjunctive_mood )

iamwil 18 days ago |

This reminds me a web 2.0 startup in 2006. It was similar in that it wanted people to break down arguments into a flowchart. Had more of a public collaboration flavor to it at the time.

https://web.archive.org/web/20060427152801/http://www.standp...

Though I like this angle a bit better, it'd be interesting to look into why standpoint had a hard time.

keyserj 18 days ago | |

Thanks for sharing, I hadn't heard of that. It saddens me a little when I hear about all the structured argument sites that died. I'm always interested in trying to learn from them, like what worked and what didn't work. I can't seem to find many details about this one.

skybrian 20 days ago |

This might be more interesting if you could read other people's discussions. As it is, it's like visiting an empty forum.

keyserj 20 days ago | |

Thanks for the feedback, I agree. I hadn't gotten around to adding an Explore page. There haven't actually been many users so it'd be mostly my topics and topics from some small group discussions I've hosted, but would be good to have more for people to see.

skybrian 20 days ago | | |

Since you use it yourself, sharing your own topics seems like a good way to start?

rappatic 18 days ago |

In which direction are you supposed to read these charts? The arrows seem to point from conclusion/solution to reasoning, but the reasoning is at the top which makes me feel like I should read it first. It would be less confusing for there to be a single convention (eg “reasoning/evidence at the top with arrows flowing mostly down towards the conclusion”).

Similarly, I was confused by the use of both active and passive: eg sometimes “creates” and sometimes “created by,” with different arrow directions for each.

keyserj 18 days ago | |

Thanks for the feedback. I had intended problems to be seen first to convey motivation, so they're at the top, but I wanted the arrows to point in the direction of causation, so solutions point upward. I agree it's not great, and knowing where to look is a common problem. My latest plan has been to have a more-linear view be primary and the diagram view be secondary (split panes on desktop) in attempts to provide more context for the diagram. Also better diagram views so that it's easier to keep only 5-15 nodes showing at a time, then direction wouldn't matter quite as much.

hankbond 18 days ago |

Seems like an interesting way to both refine and serialize thoughts about something tricky. I think it could be a really useful backbone, but I question if the interface is optimal for creating the graphs vs something more conversational and using the graphs to hold and iterate on the state. Also, I think I would want some kind of essay version as a consumer of someone else's argument rather than reading thru the graph itself.

Really cool space to explore!

keyserj 18 days ago | |

Thanks for the thoughts. I've dreamt of a version where a conversation could be recorded and a map could be generated live to provide both a visual and a structure to further guide the conversation. As for the essay version, one major issue with the app has been that it's just hard to know where you're supposed to look, and I think an essay version would probably help with that. I think LLMs can do a decent job here.

aitchnyu 19 days ago |

I expected scores and justifications to show up all the time, not just hover, no matter the view. Did you consider styling edge types differently to replace labels?

keyserj 19 days ago | |

Appreciate the feedback. I had played around with defaulting scores/justification showing vs not showing mainly because in a big diagram it can add clutter, hadn't been solid on that though.

Regarding styling edge types: Yes and I think it's a good idea. I think there are some things I haven't tried yet, but I added logic gates https://github.com/amelioro/ameliorate/issues/785 (More Actions Menu > Preferences > Edges > Use semantic arrow shapes) and there's an open ticket for using colors to convey good/bad https://github.com/amelioro/ameliorate/issues/836. Thickness is another thing that potentially could be played with, thicker being higher score? Hadn't gotten to that.

vivzkestrel 20 days ago |

- stupid question: i really dont understand what your product does

- mind explaining it like how you would do to a 5yr old kid

keyserj 20 days ago | |

If you have a problem that's hard to think through or agree on with someone else, this app can help you make a (cause & effect) diagram to understand it better. Use numbers (scores) to show others how important you think something is (nodes) or how much you think something causes another thing (edges). You can add reasons (supports/critiques) to justify your numbers, and tradeoffs (criteria) to compare options for dealing with the problem. The app can show where you disagree with people or what people think are most important (generated views).

I might need LLM help to figure out how to explain to a 5 year old haha.

alex7o 20 days ago | | |

Man you could have used this a small piece of marketing with a shareable link to show people how did you reason about the product.

msephton 20 days ago | |

I can only hope this is a joke to ask for an explanation of a product that breaks down complex problems into clear pieces.

esafak 18 days ago |

I think the logical direction for such a product is to make data-backed decisions using causal inference, like so:

https://basisresearch.github.io/chirho/

microflash 18 days ago |

Very nice. I think this can be a good visual way to run LEAN activities like 5 Why and A3 Problem Solving.

0gs 20 days ago |

this is interesting, does it use IBIS on some level? i discovered IBIS and added it and mermaid diagrams to my app but it was a bit of a lark. i think what i pictured was similar to what you have here (and should be possible though i have not tested the IBIS > mermaid pipeline).

in any case: cool

keyserj 20 days ago | |

It's not exactly IBIS, but inspired for sure. IBIS I think is better for modeling a live discussion, it's almost an append-only structure (to model back-and-forth), whereas Ameliorate's structure (the core is concepts with cause & effect relations) is intended to have new information integrated without duplication (which can be more annoying to do live because it can involve modification rather than appending).

Ameliorate models questions and arguments, but they're more auxiliary (modeled as being _about_ a cause or effect) than in IBIS, where they're pretty central. Questions in Ameliorate are more "unknowns" than "core questions we are trying to answer".

If you see my root-level comment in this post, I mention refining the Ameliorate ontology (ameliorate-v2) which elevates questions to first-class (to address the issue where it's hard to know where to look in the diagram, without questions to identify motivation). You might also be interested in my ontology playground in that comment if you want to see an IBIS > mermaid pipeline.

flarg 17 days ago |

Relationship to argument maps?

keyserj 17 days ago | |

This app has you break down information into a cause & effect map, where each node has an implied claim "this node is important" and each edge has an implied claim "the source node <edge name verb, e.g. causes> the target node>. These implied claims each get their own argument map.