org for creators reveal config
graph process into linear articles

You focus on writing down the ideas you find relevant to you topic and, as you write, things start getting connected together. You then rewire and reshape the links until you find a solid path connecting all the points.

topic

the main theme

nodes

single entities

items

can be facts, characters, places

ideas

your unique interpretation of the matter

links

the relationship between nodes

There are different methods to manage a knowledge base like: zettelkasten, personal knowledge management, second brain.

Articles are linear sequences of sentences where the author needs to think how to order the arguments and what to stress. This approach is kind of unnatural since one starts thinking about a particular topic in a much more flexible way working in parallel at many different aspects.

More natural is to start writing in graphs:

🆕 topic

draft an idea on the topic you want to cover

🗞️ knowledge

gather the relevant information and collect references

🔵 concepts

how the knowledge is connected to the topic

✏️ tag

tag each fact or idea with a clear connection with the topic

🔛 links

draw strong links between ideas

There are different types of items you should collect:

🫅 characters

what are the main characters to mention

🏔️ places

what places are involved and when

🏹 inventions

what is the particular innovative aspect

🐕‍🦺 myth

what is the open/intriguing point

🧬 facts

what can we cite as scientific result

🛕 cults

what places particularly resonate with that story

🔭 findings

objects/opinions that suggest interesting variants

📖 stories

the narrative around the topic

🏜️ environment

the particular setting of the event

🔛 relationship

how items are related together

As we write we need to think how relevant each argument is in our topic. Initially is only a draft but when we start building and re-wiring connections we might re-think the labels and start assigning more precise labels.

❕ must

some nodes need investigation, like fact checking, but mentioning them can bring off-topic

‼️ important

the item is essential to be mentioned, it’s crucial for the narrative

⁉️ interesting

some interesting topics which might not be relevant

❓ off-topic

this item might not fit in the final story-line

As a creator to keep being productive and focused org-mode can help a lot to run many background activities and help validating your thoughts as you write. During the creative process many sentences require validation and search breaking our thought-flow and creating many branches which can create alternative story lines.

visualize

complete overview of the knowledge

separate

references and main story

organize

all the background information without losing the focus on the main message

This process divides the story line from the linked knowledge. We will first of all define the topic and the scope of the project. Let’s say we want to develop a topic like “why bronze age is relevant nowadays”. We need first to create a knowledge base relative to the topic and to collect information and media around to the subject and try to keep focused without getting lost into details. We need strong support from facts which we need to collect and organize first. We need first to collect background information

overall

information and details about the subject

actors

who where the most relevant actors in that space

highlights

what are the most interesting facts to mention

A big challenge in creating content is to keep the focus despite adding details. It is important to collect the background information which doesn’t mean it needs to be mentioned (eventually added as footnote). Having too many information can lead to unclear messages weakening the main topic in adding too many details but on the other side we need to make sure that our assertions are well grounded. We will treat the topic part then as a result of understanding our graph.

Within the topic we can find some key ideas we want to convey. It is important we answer some key questions:

unique

what makes this topic so unique

audience

who can be interested in thinking about

fascination

what is the inspiring part of the topic

aesthetic

what is the aesthetic aspect

learning

why it is important to learn from it

narrative

what is the style we want to use

irony

do we want a more or less serious style

pop

do we want to include popular culture (memes)

example

what examples we can cite

moral

is there any interesting moral/ethical topic

inspiring

how do we inspire the audience

touch

how do we want to include the creator touch to the topic

Emacs has a clean interface, nothing to distract the eyes. You only see the text which is relevant to you. You focus in provide the content and the rest will be linked and generated together with all the exports. After editing an org-file you get:

presentation

the file is translated into blog post and slide

database

the database gets updated with all the sections and tags

graph

all the nodes and links get parsed by org-roam

visualize

you can visualize how the topic grow and how it is connected

publish

table of content, footnote, bibliography

agenda

you see the status of the work, how much you worked at it and what is left over

agents

agents work directly in the file

commands

placeholders and script will be executed during export

SEO

the output is rich in metadata and perfect for any type of indexing

We use org-mode because is a fully featured file format which allow to integrate with a rich set of tools and metadata:

agenda

schedule, diary

task

todo, current status, alerts

notes

capture, journal, post-it, capture, archive

spreadsheet

tables and computation

links

attachments, internal/external

publish

web page, blog, slides

graph

ui, search, viz

code

snippets, integrated with the code

CLOSED: [2026-09-16 Wed 15:02]

Emacs tends not to have menus which requires more writing. It definitely takes time to find a proper configuration which adapts to your working style, it takes a lot of trials and sometimes you overdo.

theme

(load-theme ’misterioso) (load-theme ’tango-dark) (load-theme ’wheatgrass) (load-theme ’whiteboard) (load-theme ’light-blue) (load-theme ’deeper-blue) (load-theme ’modus-vivendi) (load-theme ’leuven-dark)

center

(olivetti-mode t) (olivetti-mode nil) (olivetti-set-width 120) (olivetti-set-width 80)

modern

(org-modern-mode t)

opacity

(set-frame-parameter (selected-frame) ’alpha 100) (set-frame-parameter (selected-frame) ’alpha 80)

slide

(org-tree-slide-mode) (org-tree-slide-presentation-profile) (org-tree-slide-content) (org-tree-slide-skip-outline-level)

font

(set-face-attribute ’default nil :family “DejaVu Sans Mono” :height 100)

(set-face-attribute ’fixed-pitch nil :family “DejaVu Sans Mono” :height 100) (set-face-attribute ’variable-pitch nil :family “Spectral” :height 130)

variable

(add-hook ’text-mode-hook ’variable-pitch-mode)

CLOSED: [2026-09-16 Wed 15:02]

:EFFORT: 0:15 :RATING: +

Agents can help in different ways:

research

lookup important information and update the knowledge base

update

scan the files and check the consistency of information

display

scan the files and create mind maps of the most relevant concepts

re-write

improve some text

translate

especially if you don’t write in your native language

generate

image and sound

proof reading

a second opinion on paragraphs

exchange

challenge your positions

inspire

find other point of view

summaries

summarize your ideas

index

parse picture and text and create indices

fact checking

risky but worth trying

sense checking

you get your ideas challenged

rephrase

if the model doesn’t understand your sentences

verification

a second opinion on your ideas before saying inaccurate things

metadata

generate metadata for search like description and keywords

I opted to have everything local. I want to control the cost of the project and avoid sharing ideas. I have a basic gamer laptop and I finally found lemonade which serves locally models on CPU, GPU and NPU. Local models are less powerful but that’s intended. I don’t want to follow in temptation of letting the model do the work, I need to keep some authenticity.

Local agents hallucinate a lot, references are mostly completely made up.

I find even ironical how they state it

• References: Need credible academic sources (BibTeX style).

Chatbots are really useful to create summaries

To understand the importance of Bronze Age culture in modern days, you have to view it as the “Operating System” of modern civilization. While we live in a digital, silicon-based world, the fundamental structures of how we live, trade, govern, and create were first prototyped during the Bronze Age (roughly 3300 BCE to 1200 BCE).
...

> list the folder in a directory in python

• 1. Using os.listdir()

import os
folder_path = '/path/to/directory'
items = os.listdir(folder_path)

for item in items:
    full_path = os.path.join(folder_path, item)
    if os.path.isdir(full_path):
        print(f"Directory: {item}")

• 2. Using pathlib (Modern, Recommended)

from pathlib import Path
folder_path = Path('/path/to/directory')
folders = [f.name for f in folder_path.iterdir() if f.is_dir()]
for folder in folders:
    print(f"Directory: {folder}")

• 3. Using os.scandir() (Efficient)

import os
folder_path = '/path/to/directory'
with os.scandir(folder_path) as entries:
    for entry in entries:
        if entry.is_dir():
            print(f"Directory: {entry.name}")

Tip: Use pathlib for modern, readable code.

CLOSED: [2026-09-16 Wed 15:02]

:EFFORT: 0:10 :RATING: ++

Within the same file I can run commands in case I need to run calcuations for resizing an image: return 1920/1080

I can run scripts from this file to resize pictures for web publishing:

cd ~/lav/siti/f/f_storia/
for i in *jpg *JPG *png *PNG
do convert $i -density 150x150 -resize 1000x thumb_$i done

Create gif from video

cd ~/VideoProd/clip/
for i in *mp4; do
    ffmpeg -i $i -y -vf "fps=10,scale=500:-1:flags=lanczos" ${i/.mp4/.gif}
done

grouce_explosion.gif

Remove silence from video

cd ~/VideoProd/clip/
bash ~/lav/src/malastro/media/video_cut_silence.sh obs_try.mp4 obs_voice.mp4

Deshake video

ffmpeg -y -i unstable_$fName -vf vidstabdetect -f null -
ffmpeg -y -i unstable_$fName -vf vidstabtransform,unsharp=5:5:0.8:3:3:0.4 $fName

In org we can keep track of each task:

time spent

(org-clock-in) to start (org-clock-out) to end the subactivity

notes

notes regarding that task at a specific point [C-c C-z]

schedule

when the task is scheduled [C-c C-s]

deadline

when is the deadline [C-c C-d]

effort

specify the estimated effort

status

change the status of the task [S-] [S-]

CLOSED: [2026-09-21 Mon 11:40]

We can use the same information to create Gantt charts

Kanban view of the current status

---
config:
  kanban:
    ticketBaseUrl: 'https://mermaidchart.atlassian.net/browse/#TICKET#'
---
kanban
  todo
    model[compare model serve]
    docs[voice over]
  [in progress]
    id6[image in the knowledge base ]
  id9[review]
    id8[location knowledge]@{ assigned: 'bot1' }
  id10[media]
    id4[record]@{ ticket: MC-2038 }
    id66[volume normalization]@{ priority: 'Very Low', assigned: 'knsv' }
  id11[done]
    id5[text written]
    id2[screen capture]@{ ticket: MC-2036, priority: 'Very High'}
    id3[soundtrack]@{ ticket: MC-2037, assigned: bot1, priority: 'High' }

We can view and print the status of the activities and can be used to prepare invoices.

We need to understand first of all why we want to talk about this topic and what is our involvement. We have different approaches:

teaching

we don’t include our personal experience, we only want to educate

personal

the topic has a personal motivation

experience

the need of sharing how our own experience is relevant to the audience

need

how can we use this information

innovation

what can we build about this topic

Let’s say my answers in this case would be:

teaching

I’m not in the position to teach, I’m not an expert

personal

society

I find multiple stories inspiring and I wish our society could profit from some features of this civilization

experience

building

I tried to replicate some tools

museums

I visited plenty of museums on that topic

need

become more sensitive towards societies based less on rules and more about moral commitment

innovation

building

I want to build an archeo-park to replicate the atmosphere and the challenges of that time

writing

I want to write some fictional stories about some myths

What are the key ideas for this podcast?

borderless

folk was defined by cultures not states

mobility

people were moving a lot

sharing

sharing knowledge was really appreciated

safety

crossing some regions wasn’t as dangerous and expected

infrastructure

how could people move that much without money, streets, maps, common languages…

goods

how big was the interest for additional goods

moral

the big moral questions of the time

society

the first rules where applied, still a lot of moral questions

religion

the role of gods in the life of people

ancestors

the role of ancestors in judging

afterlife

how ancestors stay after death

adult

becoming an adult is a crucial step

korios

rite of passage in the manhood, survival in nature

There is a huge number of facts, places, myths I can mention. Let’s start defining some macro categories. Here we start putting some filters, let’s say we focus on Indoeuropean culture (Yamnaya heritage).

places

Greece, continental Europe, middle east, Egypt, India

We already have defined an important set of knowledge we want to expand on and we want to represent each category +/- equally.

I visited many archaeological museums and collected around 700 pictures I want to use for this research. I watched hundreds of videos regarding interesting stories of that time. I subscribed to multiple channels to get updates on current discoveries.

import os, re
from collections import Counter
dL = os.listdir(os.environ['HOME'] + '/lav/siti/f/f_storia/')
fL = [x.split("_")[0] for x in dL]
fL = [x for x in fL if not re.search("png",x)]
cL = Counter(fL)
cL

I can use the same table in a pandas data-frame.

import pandas as pd
musD = pd.DataFrame(data,columns=['museum','pictures'])
musD

I can eventually collect some pictures from other travels.

import os, re
from collections import Counter
import numpy as np
dL = os.listdir(os.environ['HOME'] + '/Pictures/viaggi/')
#dL = [d for d in dL if os.path.isdir(d)]
fL = []
for d in dL:
    try: 
        dL1 = os.listdir(os.environ['HOME'] + '/Pictures/viaggi/' + d)
        for d1 in dL1:
            fL.append(d)
    except:
        pass
cL = dict(Counter(fL))
np.array(cL).T

We use create_knowledge_base.py to ask a language mode to describe all the pictures we have and create and index we can search.

We can then display the information in a table.

import os, re
import pandas as pd
picD = pd.read_csv(os.environ['HOME'] + '/lav/src/spiega/data/graph/arch_desc_qwen359b.csv')
picD1 = picD[['file','description']].head()
picD1['description'] = picD1['description'].apply(lambda x: x[:50] + "...")
picD1

I list here all the books that inspired me. This can be a tedious task so I seek for agentic help. Once I create a bibliography file I can add a citation to the books in the picture.

this picture contains different archeological books and magazines regarding bronze age. Divide one by one extracting all the text you read from each book and try to guess the title and what this book is about. Output using .bib format
@book{archeology_bronze_age,
title = "important discoveries in bronze age",
note = "from the cover the author seem to be Valerio Manfredi who is a novelist and archeologist writing fictionaly books about bronze age. The publisher is Mondadori. The cover shows an image of an ancient warrior."
}

The result is far from perfect but a good start.

@book{arche_magazine,
  title = "ARCHE",
  note = "An archaeology magazine. Multiple copies are visible in the top row, featuring various archaeological sites and historical artifacts."
}
@book{historie_bronzezeit,
  title = "Die Historie des Bronzezeit",
  note = "A German publication regarding the history of the Bronze Age. It appears to be a historical survey of the period's developments."
}
@book{storica_magazine,
  title = "STORICA",
  note = "An Italian magazine dedicated to history and archaeology. Multiple copies are visible in the top right section, showcasing various historical figures and archaeological finds."
}

The greatest inspirations I got for this topic come from some famous authors. It is important here to get inspired but not plague those authors and still try to find some interesting topics to say.

After mentioning the masters you get a bit depressed but it’s important to think you still have something to say that wasn’t covered that deeply or you find a particular narrative which still brings value. We need as well to organize the big knowledge we collected without losing the focus or the relevance of what we say.

SCHEDULED: <2026-09-17 Thu>

Lists are really useful to keep track of what we need to do:

references

review material [1/3]

[X] videos

note down what you want to mention

[ ] books

write down quotations

[ ] pictures

what pictures we want to show

research

what region we need to cover [1/6]

[ ] Greece

Minoan, Mycenaean

[X] (no term)

Spain ::El Aggar

[ ] West Eu

Bell Beaker

[ ] East Eu

[ ] Central Eu

[ ] Middle East

CLOSED: [2026-09-16 Wed 15:02]

:EFFORT: 0:20 :RATING: +

We load org-roam and define first some variables

(use-package org-roam ;;nodes and links in notes
  :ensure t
  :custom (org-roam-directory "~/lav/src/spiega/")
  :bind (("C-c n l" . org-roam-buffer-toggle)
         ("C-c n f" . org-roam-node-find)
         ("C-c n c" . org-roam-capture)
         ("C-c n i" . org-roam-node-insert))
  :init (setq org-roam-v2-ack t)
  :config
  (setq org-roam-file-extensions '("org" "md"))
  (md-roam-mode 1)                ; must be on before any db sync (md files)
  (org-roam-setup)                ; registers org-roam hooks + protocol
  (org-roam-db-autosync-mode 1) ; this builds the DB (once, now)
  (dolist (cmd '(org-roam-node-find org-roam-node-insert
                                    org-roam-buffer-toggle org-roam-capture
                                    org-roam-db-sync org-roam-db-explore
                                    org-roam-db-complete-at-point))
    )
  )

We need to create nodes and insert them across our documentation. Nodes are created by adding

:PROPERTIES:
:ID: node_insert-org_creator
:ROAM_ALIASES: node_insert-org_creator
:END:

[C-c n i]

insert

[C-c n f]

find

[C-c n l]

buffer toggle

[C-c n I]

insert immediate

[C-c n p]

find project

[C-c n t]

capture task

[C-c n b]

capture inbox

[C-c n d]

dailies map

[C-c C-o]

follow node

CLOSED: [2026-09-16 Wed 15:02]

:EFFORT: 0:10 :RATING: ++

Links are automatically created by org-roam but we need to create cross references between nodes. Every time we see a reference between concepts we need to create the connection [C-c n i]. We can visualize the graph with (org-roam-ui-mode) and

We can access the same database with sqlite

import sqlite3, os
con = sqlite3.connect(os.environ['HOME'] + "/.emacs.d/org-roam.db")
cur = con.cursor()
res = cur.execute("SELECT name FROM sqlite_master")
tabL = res.fetchall()
print(tabL)

[('files',), ('sqlite_autoindex_files_1',), ('nodes',), ('sqlite_autoindex_nodes_1',), ('aliases',), ('citations',), ('refs',), ('tags',), ('links',), ('alias_node_id',), ('refs_node_id',), ('tags_node_id',)]

We create now a personal chatbot to ask our own knowledge base. We will create a collection in a vector-store and use langchain to search within the data.

vecS = m_r.access_vectorstore("storia",os.environ['HOME']+'/lav/src/spiega/data/graph/')
retrV = vecS.as_retriever(search_kwargs={"k": 4})
q = "Where were the vessels found?"
resp = m_r.retrieve_info(retrV,q)

Writing is a difficult process for creators which have a lot of ideas and need to find inspiration to come with a catchy story and coherent message. Org is an incredible tool which doesn’t constrain your way of thinking but rather expand your capabilities to the next level. In this video we show how you can work at your main idea and still proceed with references, background research, notes, sparse ideas, task tracking and graph navigation. The detailed article of the project is https://intertino.it/spiega/a/org_creator.html More on org files: https://intertino.it/spiega/a/org_files.html emacs: https://intertino.it/spiega/a/emacs.html local agents: https://intertino.it/spiega/a/agent_local.html agent graphs: https://intertino.it/spiega/a/agent_interview.html

Moments: 00:00 org-mode 00:30 writing in graphs 02:30 ideation framework 03:30 productivity tools 04:30 UI 06:00 project management 07:30 local agents 09:00 create your story 11:00 connect the dots 12:30 final overview

Here we use hugo, we set few folders and export the whole project ref.

(setq org-hugo-base-dir "~/lav/src/spiega/hugo")
(setq org-hugo-default-section-directory "posts")
(setq blog-dir "~/lav/src/spiega/storia"
      public-blog-dir "~/lav/src/spiega/hugo"
      hugo-process "Hugo Server"
      hugo-server-site "http://localhost:1313/")
(org-hugo-export-wim-to-md)