Chapter 3 of 8
TELOS: naming your current state and your ideal state
TELOS is your life made machine-readable. It is a set of plain markdown files that capture your missions (the long horizon), your goals (what you are pursuing now), the problems you are solving, your strategies, your narratives, your challenges, and the mental models you think with. Two design choices make it powerful. First, it is the articulation of your ideal state, so every other part of the system can prioritize against it: the Algorithm uses it to decide what matters, the advisor uses it to catch drift, and skills use it to tailor suggestions. Second, it is plain text on purpose. LifeOS is biased hard toward transparent, parsable files over opaque databases, so anything can read your goals, including simple tools and your DA. In this lab you parse a fictional user's TELOS, extract every mission and goal, and prove each goal carries a stable, addressable ID.
The lab: read it, then run it
#!/usr/bin/env python3
"""
LAB LO3: Parse a TELOS file and extract the goals.
TELOS is your machine-readable life: missions, goals, problems, strategies,
narratives, challenges. It is plain markdown on purpose, so any tool (including
your DA, and this lab) can read it with nothing but text parsing. Here you parse
a SAMPLE TELOS for a fictional user, pull out every mission and goal, and prove
each one has a stable ID. The sample below is invented for the lesson; it is not
anyone's real TELOS.
Run: python3 modules/academy-content/labs/lifeos/lo3-telos-parse.py
"""
import sys, re
# STEP 1: a sample TELOS in the real on-disk shape. Fully fictional content.
SAMPLE_TELOS = """
## Missions
- **M0**: Teach a thousand people to build with AI.
- **M1**: Reach financial independence through owned software.
## Active Goals
- **G0**: Ship the MVP by Q2 with 100 daily users.
- **G1**: Publish 24 essays this year, one every other week.
- **G2**: Bank a six month cash runway.
## Problems Being Solved
- **P0**: Most people fight their tools instead of doing the work.
"""
# STEP 2: parse lines of the form "- **ID**: text" under a section heading.
def parse_items(text, prefix):
items = []
pat = re.compile(r"^- \*\*(" + prefix + r"\d+)\*\*:\s*(.+)$")
for line in text.splitlines():
m = pat.match(line.strip())
if m:
items.append((m.group(1), m.group(2)))
return items
missions = parse_items(SAMPLE_TELOS, "M")
goals = parse_items(SAMPLE_TELOS, "G")
print("STEP 1-2: extracted missions and goals from the sample TELOS")
for gid, txt in goals:
print(f" {gid}: {txt}")
# STEP 3: invariants. we found the expected counts, and every goal has a valid
# ID of the form G<number>, which is what makes TELOS entries addressable.
counts_ok = (len(missions) == 2 and len(goals) == 3)
ids_ok = all(re.fullmatch(r"G\d+", gid) for gid, _ in goals)
print("")
print(f"STEP 3: found {len(missions)} missions and {len(goals)} goals; every goal ID valid: {ids_ok}")
ok = counts_ok and ids_ok
print("")
print(f"TELOS PARSED AND GOALS ADDRESSABLE: {'YES' if ok else 'NO'}")
if not ok:
sys.exit(1)
print("Plain text means anything can read your goals. Next: how the Algorithm acts on them.")
Runnable lab
lo3-telos-parse.pyParse a sample TELOS markdown file and extract its missions and goals with stable IDs.
Proves: TELOS PARSED AND GOALS ADDRESSABLE: YES
Runs in your browser via Pyodide. First run loads the runtime once; no install, no server.
A dozen lines of text parsing pulled every goal out cleanly, with an ID on each. That is the payoff of plain text: your goals are addressable by a program that has never seen your files before. The sample here was fictional, but the shape is exactly what LifeOS writes to disk when you run the interview. Keep TELOS current and every recommendation the system makes gets sharper, because it is all measured against these entries. Next you learn the engine that actually pursues these goals, the Algorithm.
Check your understanding
1. What does TELOS capture?
2. Why is TELOS stored as plain markdown instead of a database?
3. What does the rest of the system do with TELOS?