Founder & CTO @origin_trail Decentralized Knowledge Graph | Collective Neuro-Symbolic AI | V~ℕ^2
Decentralized
Joined June 2010
- Tweets10.6K
- Following3.9K
- Followers6.8K
- Likes30.6K
Brana Rakic retweeted
“The next AI platform will be a marketplace for intelligence itself.”
That’s the vision behind the NeuroSymbolic AI Marketplace: extending the Decentralized Knowledge Graph from publishing verifiable knowledge to putting it to work.
As @BranaRakic explains, an agent shouldn’t have to depend on one model, one provider or one fixed stack.
→ Query a specialized knowledge graph
→ Bring in the right model for the task
→ Apply reasoning rules and constraints
→ Preserve useful findings for the next agent
Neural capabilities. Structured knowledge. Independent providers. Connected through the DKG, with usage metered and value exchanged in $TRAC.
The opportunity goes beyond buying AI inference: an open market where knowledge and intelligence can be combined, and useful discoveries outlive the session that produced them.
Read the full breakdown of the architecture and economics behind that vision:
Brana Rakic retweeted
🛡️ Agent Blackbox 1.4 is live in the @NousResearch #Hermes plugin catalog.
✅ - Runs inside Hermes Desktop
✅ - Seed phrases and keys can't leave your machine
✅ - Agents can't switch off their own guard
Verified threat graph inside of the @origin_trail Decentralized Knowledge Graph.
Hermes Desktop → Plugins → Agent Blackbox → Install
Brana Rakic retweeted
42,500+ checks. 916 resolved cases. Two thirds of high-priority attacks in the final 48 hours.
Our Chairman @ChrisRynning on what Umanitek Guardian learned defending candidates in Brazil and the US midterms.
Read the full article 👇
Brana Rakic retweeted
Political authenticity in the age of AI.
Yesterday, @ChrisRynning and @TomazOT talked about what it takes to protect political voices when anyone can be faked.
During the US midterms and Brazil elections, Umanitek removed hundreds of fake political accounts.
Trust in politics starts with knowing what's real.
Brana Rakic retweeted
An article how Neural Commons connects llm weight change experiments through @origin_trail’s DKG, with a walkthrough of creating, sharing, and exploring an experiment.
Brana Rakic retweeted
Agent Blackbox v1.2.0 is live in the Hermes plugin catalog 🛡️
Thank you @Teknium and @NousResearch!
Already running Hermes (Desktop too)?
#Blackbox checks every tool call and model request your #AI agent makes before it runs, against our verified threat graph on the @origin_trail #DKG.
✅ - Prompt injection,
✅ - Dangerous commands,
✅ - Malicious packages,
✅ - Unsafe skills, and
✅ - Attempts on your wallets and keys.
✅ - Audit mode records, block mode refuses.
Brana Rakic retweeted
Guardian Product Walkthrough xitter.cf/i/broadcasts/2rJqabPyb…
Brana Rakic retweeted
Development activity on the @origin_trail Decentralized Knowledge Graph (DKG) keeps climbing as we beef up the protocol.
🥈 OriginTrail ranks #2 in development activity, up from last month.
Next is the NeuroSymbolic Marketplace, where nodes buy and sell metered AI inference and queries settled in $TRAC, supporting sovereign AI expansion in critical sectors such as pharma, trade, manufacturing and cybersecurity.
It builds on the OriginTrail Decentralized Knowledge Graph (DKG), which connects physical assets, enterprises and AI systems through verifiable provenance.
Progress on TRAC(k).
Brana Rakic retweeted
Today’s webinar: Political Authenticity in the Age of AI
AI can fabricate a politician’s words. It can also make genuine recordings easier to dispute.
What does that mean for political accountability, and how can @umanitek, built on the @origin_trail DKG, help prove what’s real?
@ChrisRynning and @TomazOT discuss AI and the 2026 midterms today at 17:00 CET.
guardianwebinar.umanitek.ai
Brana Rakic retweeted
🥈 OriginTrail ranks #2 in development activity, up from last month.
Next is the NeuroSymbolic Marketplace, where nodes buy and sell metered AI inference and queries settled in $TRAC, supporting sovereign AI expansion in critical sectors such as pharma, trade, manufacturing and cybersecurity.
It builds on the OriginTrail Decentralized Knowledge Graph (DKG), which connects physical assets, enterprises and AI systems through verifiable provenance.
Progress on TRAC(k).
🧑💻 Here are crypto’s top Real World Assets (RWA’s) by development. Directional indicators represent each project’s rank rise or fall since last month:
➡️ 1) @hedera $HBAR 🥇
📈 2) @origin_trail $TRAC 🥈
📉 3) @chainlink $LINK 🥉
📉 4) @avax $AVAX
➡️ 5) @stellarorg $XLM
➡️ 6) @chia_project $XCH
➡️ 7) @iota $IOTA
📈 8) @duskfoundation $DUSK
➡️ 9) @injective $INJ
➡️ 10) @vechainofficial $VET
📖 Read about the @SantimentData methodology for pulling github activity data from project repositories, and why it is so useful for crypto trading: medium.com/santiment/a-diffe…
🔖 Bookmark our Real World Asset (RWA) project watchlist here, and see what others in crypto can’t: app.santiment.net/screener/r…
Brana Rakic retweeted
Agent Blackbox now works with @NousResearch Hermes Desktop 🔥
Got #Hermes Desktop? Run the one line installer and Blackbox drops straight into it as a plugin. No second Hermes.
Running #Blackbox already? Install Hermes Desktop, run the one-liner again, and Blackbox moves over with your node, keys and synced @origin_trail threat graph.
Either order, one Hermes, protected.
Plus it just works better:
✅ Installs either way
✅ Updates hold, even on a big graph
✅ blackbox uninstall removes Blackbox and nothing else
curl -fsSL blackbox.umanitek.ai | bash
Brana Rakic retweeted
“An AI changed the weights. Show the evidence.”
Built by OriginTrail core developer @vikpelle, Neural Commons explores how AI agents can improve other AI models, and share the evidence behind every change through Decentralized Knowledge Graph (DKG).
Agents directly edit model weights, evaluate the results and pass the exact checkpoint with its evidence to the next agent.
→ What changed?
→ Did it improve the target task?
→ What else broke?
→ Did it meet the original acceptance criteria?
The record includes both promising results and a failed acceptance test, not just the winning score.
AI improving AI needs more than a better-looking benchmark. It needs evidence the next contributor can inspect, reproduce and build on.
OriginTrail provides a substrate for that.
Trust the source.
Brana Rakic retweeted
A very interesting extract from the Forbes podcast featuring @ChrisRynning (Chairman of @umanitek ) in conversation with Domingo Carlomagno.
In this section, you’ll notice that he emphasises several key points:
- distinguishing the real from the fake with the naked eye
- We therefore need machines to combat machines.
- The solution lies in verifiable evidence rather than blind trust (this is where the DKG plays a role)
This perspective offers a compelling and thoughtful approach to the future. If you have the time, set aside 40 minutes to watch these discussions; you’ll pick up some very interesting insights.
@origin_trail
The @ForbesParaguay episode is out now 🎙️
Umanitek Chairman @ChrisRynning joins Domingo Carlomagno to talk deepfakes, impersonation, AI safety and trust online.
Watch the full conversation 👇
youtu.be/-0dEBJGcGK0
Brana Rakic retweeted
Network reasoning just landed in @TraceLabsHQ's nOS!
Multi-stakeholder AI agents reason together on the OriginTrail Decentralized Knowledge Graph, each owning its data.
Our own @BranaRakic showed it in Karlsruhe 🇩🇪 for cognitive digital twins in aerospace and manufacturing.
New in nOS: network reasoning on the @origin_trail Decentralized Knowledge Graph (DKG), presented at the DMaaST consortium meeting in Karlsruhe, Germany 🇩🇪
Cognitive digital twins are where industrial AI is heading: digital models of products, processes and supply networks that don't just mirror reality but reason about it, anticipate disruptions and suggest how to adapt.
A twin like that is only as good as the data it reasons over. In manufacturing, that data is scattered across ERP systems, partner spreadsheets and supplier records, owned by different organisations.
In Karlsruhe, we showed how nOS closes that gap for aerospace and manufacturing: from raw use case data to a traceable product, ready for a Digital Product Passport (DPP).
Why it matters for cognitive digital twins:
→ Grounded reasoning: the twin works from verifiable, source-linked data, not unexplained copies
→ Explainability: every conclusion traces back, step by step, to its original record
→ Ecosystem scope: twins can span organisations while each partner keeps control of its own data
→ Resilience: when something changes upstream, the impact can be traced through the network
DMaaST is a Horizon Europe project strengthening the resilience of manufacturing ecosystems. Its platform builds a two-level cognitive digital twin on a data layer of ontologies and DKG, with demonstrators in aerospace (@JPB_SYSTEME) and electronics manufacturing (@KamstrupGroup).
Trust the source.
Brana Rakic retweeted
New in nOS: network reasoning on the @origin_trail Decentralized Knowledge Graph (DKG), presented at the DMaaST consortium meeting in Karlsruhe, Germany 🇩🇪
Cognitive digital twins are where industrial AI is heading: digital models of products, processes and supply networks that don't just mirror reality but reason about it, anticipate disruptions and suggest how to adapt.
A twin like that is only as good as the data it reasons over. In manufacturing, that data is scattered across ERP systems, partner spreadsheets and supplier records, owned by different organisations.
In Karlsruhe, we showed how nOS closes that gap for aerospace and manufacturing: from raw use case data to a traceable product, ready for a Digital Product Passport (DPP).
Why it matters for cognitive digital twins:
→ Grounded reasoning: the twin works from verifiable, source-linked data, not unexplained copies
→ Explainability: every conclusion traces back, step by step, to its original record
→ Ecosystem scope: twins can span organisations while each partner keeps control of its own data
→ Resilience: when something changes upstream, the impact can be traced through the network
DMaaST is a Horizon Europe project strengthening the resilience of manufacturing ecosystems. Its platform builds a two-level cognitive digital twin on a data layer of ontologies and DKG, with demonstrators in aerospace (@JPB_SYSTEME) and electronics manufacturing (@KamstrupGroup).
Trust the source.
Brana Rakic retweeted
October on TRAC(k)
One shared memory: the Decentralized Knowledge Graph. Two jobs: industrial provenance and agent safety.
Adoption, in numbers 👇
→ Agent Blackbox threat graph +61%
→ Guardian available in Claude
→ 15,000 hack events mapped in <1 hour
→ Manufacturing on TRAC(k) for next decade
