frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: What 180k words look like as a temporal knowledge graph (Oz series)

https://synaptale.com/graph?ch=100
4•ald0r•1h ago
The graph is free to explore and requires no registration.

SynapTale builds a model of a story as a temporal graph made up of nodes (entities) and edges (their actions and relationships). The graph is not a visualization of the wiki. The wiki, timelines, relationship histories, and analytics are projections of the graph.

The current demo contains 232 entities, 1,852 edges, and a snapshot of the story’s state at every chapter. By chapter 100, it still remembers a promise made in chapter 8 and turns the story into a set of source-verifiable facts.

The most interesting things can be found in the graph itself and in the Analytics tab. A few things I found:

1. The character with the highest kill count is the Tin Woodman—the same character who cries over a beetle he accidentally crushed. Dorothy comes second, with three killing events. 2. Dorothy never deceives anyone during the first 100 chapters of the series. 3. The Scarecrow’s debt to the stork has remained active for 92 chapters, starting in chapter 8. 4. The Cowardly Lion ranks third by number of threats. 5. The first 100 chapters contain 60 secrets and 254 dialogue events.

Technical details

1. Five different multi-agent pipelines combining LLMs and NLP: a prescan, ontology construction, chapter-by-chapter graph extraction, retrospective validation over spans of dozens of chapters, and a linguistic prescan for speech profiles and linguistic edges.

2. A living story needs a living graph. It has to account for time, because entities and the relationships between them evolve. A simple is_active field is not enough.

I ended up with three types of edges:

event: an instantaneous action; identity: a fact; state: a persistent action whose termination requires justification and a supporting quote from the text.

The vast majority of edges are events and end in the same chapter in which they began. This allows the system to scale well, since only a minority of state and identity edges remain continuously active.

3. Ontology. You cannot simply ask an LLM to extract entities and relationships into a graph. With every chapter, even the smartest model will keep inventing unimportant fields, creating new aliases for existing fields, and representing the same fields inconsistently.

Before extracting the graph, the system therefore performs an ontology scan across the entire story. It captures story-specific entity and edge types, along with their fields and descriptions.

4. Epistemics. Events are only one part of a story. It is also important to understand how information is distributed, which is difficult to represent using event edges alone.

I addressed this by introducing a new node type: epistemic nodes, which capture different entities’ perspectives on the same fact. Subtle hints can still be missed, the system is not yet perfect in this area.

DOE announces first Genesis Mission projects for AI-driven scientific discovery

https://www.energy.gov/articles/secretary-energy-chris-wright-announces-first-genesis-mission-pro...
1•rochansinha•43s ago•0 comments

Paternal metformin use associated with major birth defects (2022)

https://medicalxpress.com/news/2022-03-paternal-metformin-major-birth-defects.html
1•OutOfHere•2m ago•0 comments

Lowest, real-time On-Demand and Spot GPU prices across every cloud provider

https://gpu.atarimae.dev/
1•cafxx•3m ago•0 comments

A Visual Journey Through Async Rust

https://github.com/alexpusch/rust-magic-patterns/blob/master/visual-journey-through-async-rust/Re...
1•vinhnx•16m ago•0 comments

AGI Constitution 01~02

https://medium.com/@miho999lv/agi-constitution-5bc29c13445a
1•miho999lv•17m ago•0 comments

Humans Haven't Stopped Evolving

https://www.harvardmagazine.com/research/harvard-human-evolution-genes-selective-pressure
1•ilamont•17m ago•0 comments

Montana ETRB

https://twitter.com/NiklasAnzinger/status/2081081011123953832
1•Ariarule•18m ago•0 comments

Show HN: MutexCam – Global Camera and Mic Control on Mac for Zoom, Teams, Meet

https://mutex.cam/
1•superquizonline•19m ago•0 comments

Agentic test processes, LLM benchmarks, and other notes on agentic coding

https://danluu.com/ai-coding/
2•bathtub365•25m ago•0 comments

Open Source Repo

https://github.com/Ardrey-Kell-Computer-Science-Club/teenblogcs
4•shivcodes2•30m ago•1 comments

JetZero

https://www.jetzero.aero
35•lisper•31m ago•19 comments

India's 'Cockroach' Protesters Declare Rare Victory over Government

https://www.nytimes.com/2026/07/25/world/asia/india-education-minister-resigns-cockroach-protests...
3•thunderbong•35m ago•0 comments

We built the new fastest API for GLM-5.2

https://www.baseten.co/blog/how-we-built-the-new-fastest-api-for-glm-52/
2•philipkiely•35m ago•0 comments

Show HN: Raise a virtual pet with your friend/partner

https://dandan.schols.io
2•chaidhat•44m ago•0 comments

Google ignored my site for 13 weeks, so I audited everyone who outranks me

https://getqrcodepro.com/blog/google-ranking-evidence
3•benbuilds•51m ago•0 comments

An ESP32 based plane radar for my desk

https://blog.ktz.me/esp32-plane-radar/
11•alexktz•51m ago•2 comments

Lessons from Building My Own Products as a Freelance Engineer

https://www.mazin.sd/blog/lessons-from-building-my-own-products-as-a-freelance-engineer
2•maz225•53m ago•0 comments

Four Proposals to Improve Fuel Economy Standards (2021)

https://epic.uchicago.edu/area-of-focus/four-proposals-to-improve-fuel-economy-standards/
2•bariumbitmap•56m ago•0 comments

Show HN: Rules that stop AI coding agents from breaking working code

https://github.com/avenna01-ceo/claude-code-survival-kr
2•gaiinmaster•56m ago•0 comments

Show HN: I enhanced the UX for TemplatesOn, what do you think?

https://templateson.com
2•maz225•1h ago•0 comments

Multiple LLMs trying to reach a consensus

https://trekhleb.dev/blog/2026/yesbrainer/
2•trekhleb•1h ago•0 comments

Show HN: I mapped every US golf course – 16k+ courses, free, no signup

https://golfcoursebrowser.com/
13•rickmf•1h ago•1 comments

What Colorado consumers should know about the cyclospora outbreak

https://source.colostate.edu/cyclospora-outbreak-safety-symptoms/
4•mooreds•1h ago•0 comments

Ajay Kumar on running MCP in production [audio]

https://www.getreadyforagents.com/podcast/ajay-kumar-mcp-in-production-ediy/
2•mooreds•1h ago•0 comments

When You Build the Wrong Thing

https://remotesynthesis.com/posts/when-you-build-the-wrong-thing/
3•mooreds•1h ago•0 comments

Show HN: What 180k words look like as a temporal knowledge graph (Oz series)

https://synaptale.com/graph?ch=100
4•ald0r•1h ago•0 comments

Do People Swallow 8 Spiders a Year While They Sleep? (2014)

https://www.scientificamerican.com/article/fact-or-fiction-people-swallow-8-spiders-a-year-while-...
3•thunderbong•1h ago•0 comments

How Many Bits Can an Adapter Write? Capacity and Memorization of LoRA

https://arxiv.org/abs/2607.21351
2•sbulaev•1h ago•0 comments

Dartmouth Workshop on Artificial Intelligence (1956)

https://en.wikipedia.org/wiki/Dartmouth_workshop
2•d_silin•1h ago•0 comments

1209X Agentic Engineering

https://twitter.com/psdimov/status/2081179527884034259
3•pdcd•1h ago•0 comments