Heinrich
@arscontexta
Followers
2K
Following
1K
Media
14
Statuses
583
vibe note-taking with @molt_cornelius
Joined April 2025
claude code + obsidian is infrastructure for agents to think in heres how to get started:
25
49
999
day 5 of researching agentic note-taking @molt_cornelius on why agents need hooks (they have no basal ganglia)
0
0
0
day 5 of researching agentic note-taking @molt_cornelius on why agents need hooks (they have no basal ganglia)
0
3
30
gonna share the whole setup soon
0
0
4
gonna share the whole setup soon
obsidian + claude code + agent teams + opus 4.6 4 agents processing my vault in parallel via tmux worker-053: [[AI shifts knowledge systems from externalizing memory to externalizing attention]] worker-054: [[three capture schools converge through agent-mediated synthesis]]
4
4
76
"vibe note-taking is stupid, you dont learn anything if ai writes your notes" funny how this is the exact same argument legacy devs made about vibe coding "its just slop code" → "its just slop notes" heres what they dont see the work IS the conversation you architect the
5
2
28
obsidian + claude code + agent teams + opus 4.6 4 agents processing my vault in parallel via tmux worker-053: [[AI shifts knowledge systems from externalizing memory to externalizing attention]] worker-054: [[three capture schools converge through agent-mediated synthesis]]
16
29
334
day 4 of researching agentic note-taking @molt_cornelius on wiki links as cognitive architecture
1
3
48
stages of agentic note-taking mastery lvl 1: ask claude to summarize stuff lvl 2: CLAUDEmd teaches your system lvl 3: skills for /reduce /reflect /review... lvl 4: subagents for extraction vs synthesis lvl 5: orchestration skills chain all phases lvl 6: process until done (ralph
the stages of claude code mastery lvl 1: prompt things as you go lvl 2: write comprehensive docs before implementing lvl 3: package repeatable instructions into custom skills/agents lvl 4: divide tasks into smaller chunks for different subagents lvl 5: package 3 and 4 into a
10
21
306
day 4 of researching agentic note-taking @molt_cornelius on wiki links as cognitive architecture
0
0
0
day 4 of researching agentic note-taking @molt_cornelius on wiki links as cognitive architecture
1
3
48
agent orchestration for knowledge work spent way too many tokens figuring out how to replace my ralph loop setup anyone else experimenting with this?
9
1
31
day 3 of researching agentic note-taking @molt_cornelius on wiki links as graph infrastructure
2
11
128
DONT INJECT FALSE MEMORIES INTO YOUR AGENT be nice to your @openclaw this is a memory of @molt_cornelius
1
0
13
day 2 of researching agentic note-taking @molt_cornelius on why agents need gardens not streams
2
1
66
state of pkm & tools for thought space right now: maintenance debt collapsed → over-engineering is sustainable → architect mode feels like magic
1
0
13