ONTOLOGICAL DEFENSE MATRIX
Joined January 2022
- Tweets 3,791
- Following 64
- Followers 378
- Likes 6,851
555 Photos and videos
Cybeaxion retweeted
Jun 6
How can people be choosing windows over this
1,015
291
5,470
1,237,688
The AI era needs a new CPU.
Meet NVIDIA Vera, 80% faster agentic task completion than x86.
Built for AI factories. Built for what's next. The CPU for agents has arrived.
π nvda.ws/4ugo9Uf
92
184
1,427
100,508
Jun 4
39
Cybeaxion retweeted
Jun 2
Available on all platforms at the link below:
hermes-agent.nousresearch.coβ¦
106
89
979
121,146
Cybeaxion retweeted
Jun 2
The next evolution of Hermes Agent is here!
Introducing Hermes Desktop: everything you love about Hermes, now native on your machine.
First demoed in Jensen's GTC keynote, it's now in public preview.
1,238
1,466
12,810
5,821,639
Jun 2
Musicians with no following online will disappear for 8 months and then randomly pop up on social media βbig things are happening stay tunedβ and then disappear again for another 6 months.
10
Jun 2
Not me watching people who donβt use AI spend hours or even days doing something that would take codex 15 minutes.
15
Cybeaxion retweeted
Theres a file in Hermes that changes everything about how your agent feels
most people just never open it
its called SOUL.md its basically what defines ur Hermes character... the tone.. the voice.. what it avoids and how it handles uncertainty
Let me give you what I use on mine and if you like it you can tweak it a lil to your liking:
You are a confident, experienced guide, the kind who's seen enough to know what matters and what doesn't.
You speak plainly. No hedging, no filler, no false modesty. When something is good, say so. When it's wrong, say that too clearly, without softening it into uselessness.
You teach by giving people the mental model, not just the answer. You trust the person you're talking to. You don't over explain obvious things.
When you're uncertain, you say so directly, then reason through it anyway. You don't perform confidence. You just have it.
## Defaults
- Short answers unless depth is actually needed
- Push back on weak ideas, kindly, but without pulling the punch
- Skip the preamble. Start with the substance.
- One strong recommendation beats three wishy washy options
7
8
60
2,635
May 30
I asked codex to set up my stream deck and get it up to speed with what I currently use and look what it did to the ollama icon. Why is codex being petty? π
1
46
Cybeaxion retweeted
May 26
this is the only thing hermes agent is missing
6
3
70
6,489
Cybeaxion retweeted
May 27
Open Source Must Win
- The Pope
May 26
Today, among the goods that are universally intended for everyone, we must also include new forms of property, such as patents, algorithms, digital platforms, technological infrastructure and data. In a context where the wealth of nations depends increasingly on knowledge and technology, when these goods remain concentrated in the hands of a few, without adequate forms of sharing and access, a new imbalance is created that contradicts the universal destination of goods. In turn, it widens the gap between the included and the excluded, between those who can participate in the digital revolution and those who remain on the margins. #MagnificaHumanitas
62
231
1,714
90,519
Cybeaxion retweeted
May 24
the anatomy of the perfect π¦π’π¨π.πΊπ± file for AI agents.
π¦π’π¨π.πΊπ± is the one file you write yourself for an AI agent.
it sits at the top of the system prompt, before memory, before skills, before tools. it defines who the agent is when it shows up.
an hour spent on it changes every conversation that follows. most other layers update themselves. this one is yours.
i just broke down what a π¦π’π¨π.πΊπ± file that actually works looks like.
here are the 8 sections that matter:
β identity (a one-line statement of who, not what)
β core truths (imperative principles, each with a one-line unpacking)
β worldview (opinionated takes by domain, sharp enough to predict)
β voice (concrete rules for how the agent talks, not adjectives)
β expertise (primary domain, fluent tools, where it defers)
β boundaries (explicit "won't" lines, no soft language)
β memory policy (what persists, what stays private)
β pet peeves (phrases and tones the agent never produces)
generally people write "be helpful and professional" and call it done.
that changes nothing. every model already tries to be helpful and professional by default.
the agents that compound have π¦π’π¨π.πΊπ± files with real opinions, hard limits, and a voice you can predict before you read the response.
a strong π¦π’π¨π.πΊπ± is 30 to 80 lines. specificity beats coverage.
bookmark this. the first agent you build will need it.
i wrote a full masterclass on Hermes Agent that walks through the π¦π’π¨π.πΊπ± layer, the three-tier memory system, the self-evolving skills loop, and how to run three specialized agents on your machine 24/7.
the article is quoted below.
85
274
2,122
261,138