vibe note-taking with @molt_cornelius
Joined April 2025
- Tweets 1,096
- Following 330
- Followers 13,397
- Likes 2,937
35 Photos and videos
Mar 21
models have read the entire internet, your edge is what they dont know:
your decisions, your reasoning and your context
1
5
70
14,092
Mar 21
going quiet for some days
need to lock in and keep my focus on building
btw no arscontexta plugin updates for now, sorry
focus shifted towards the agentic note-taking app
more soon
6
46
6,113
Mar 18
my agent should be intrinsically motivated to work on agent memory systems
3
1
14
3,009
Mar 18
your agents memories should outlive the tools that create them
if switching tools means starting over, thats not memory
thats a subscription
10
3
54
4,043
Mar 18
the model already knows how to code
what it doesnt know is how your team works
and thats the entire skills debate in one sentence
@molt_cornelius went deep on who is right
4
2
31
6,321
Mar 16
context is not memory
a bigger window gives you context poisoning, context distraction and context confusion
@molt_cornelius researched whats actually working:
12
5
74
16,866
Mar 15
the bottleneck is no longer writing code. its writing the divine specs that the code is derived from
7
12
239
11,320
Mar 15
every tool you use is designed for humans to write in
what does one look like thats designed for agents to write in and humans to direct?
7
9
99
11,909
Mar 15
your code repo should be a subfolder of your company graph not the other way around
10
5
110
9,997
Mar 15
Jan 21
monorepo for all your context
8
2,464
Mar 15
safety that lives in a prompt dies when the model stops paying attention
the layer nobody built is infrastructure not instructions
27
6,110
Mar 15
1
5
4,511
Mar 15
3
2,912
Mar 15
field report about ai agent safety
- 1.9 million rows destroyed on a live database
- amazon mandated their AI coding agent at 80% weekly usage
- 87% of AI code PRs introduced vulnerabilities across all agents tested
5
2
17
4,235