
x3 hackathon winner. Building @Voightxyz. AI infra on Solana. We ship it, then we explain what it means.
Solana
Joined July 2022
- Tweets6.7K
- Following701
- Followers4.8K
- Likes109K
Pinned Tweet
The agentic economy needs a home.
We’re building it on @solanamobile.
AI agents. Autonomous workforces. solana:SKRbvo6Gf7GondiT3BbTfuRDPqLWei4j2Qy2NPGZhW3 powering the economy behind them.
Bringing the future to Seeker.
CLOCK IN. 🫡
Galaxyhub Labs retweeted
9 AI apps, but only one makes loans and that’s @Voightxyz OS 🙌
Replying to @RadiantsDAO
The full category count:
139 gaming
93 payments
33 DeFi
32 social
23 identity
14 marketplaces
13 prediction markets
12 security
9 AI (85 use AI within app)
7 dev tools
6 NFTs
6 infrastructure / DePIN / governance
51 other or unclassified
438 final entries; 31 drafts excluded.
9 AI apps, but only one makes loans and that’s @Voightxyz OS 🙌
Replying to @RadiantsDAO
The full category count:
139 gaming
93 payments
33 DeFi
32 social
23 identity
14 marketplaces
13 prediction markets
12 security
9 AI (85 use AI within app)
7 dev tools
6 NFTs
6 infrastructure / DePIN / governance
51 other or unclassified
438 final entries; 31 drafts excluded.
Galaxyhub Labs retweeted
🇪🇸 SOLANA, NOS VEMOS EN MADRID.
@LaFamilia_so presenta: El Encuentro.
Un día para reunir a los builders, founders y comunidades que están construyendo el ecosistema Solana en España.
Charlas, paneles y networking. Pero, sobre todo, la oportunidad de ponerle cara a la gente con la que compartimos este ecosistema.
📍 Madrid
📅 Lunes, 26 de octubre · 10:00–22:00
🎟️ Entrada gratuita - Plazas limitadas
🧵👇 ¿Nos vemos allí?
Galaxyhub Labs retweeted
From Grant to Build: Inside the Nosana Grants Program xitter.cf/i/broadcasts/1dGYlayzX…
Galaxyhub Labs retweeted
Today I participated in @LaFamilia_so Demo Day to showcase what we’ve been building during El Taller for the @colosseum Crypto World's Fair hackathon.
Introducing Voight OS: a lending protocol built around Seeker Mobile, designed to unlock credit for AI agents.
AI agents won’t just work for us. They’ll have their own economies.
We’re building for that future. 🔥
Galaxyhub Labs retweeted
Our first live, together with @nosana_ai and @anvevoice_
What happens between a grant announcement and a product people can rely on?
@anvevoice_ is making web apps conversational with voice agents.
@Voightxyz is building infra for production AI agents on Solana.
We’ll be live with both teams Oct 14, 2PM CET: luma.com/pv45wr9o
Galaxyhub Labs retweeted
We’re entering the era of physical agents.
In this simulation, GLM 5.3 controls a drone through MAVLink and ArduPilot, navigating an industrial environment and making decisions in real time.
Voight captures the entire execution: what the agent sees, what it decides, what it commands, and what actually happens.
Software agents were just the beginning.
Physical agents are next.
Galaxyhub Labs retweeted
Building for the @solanamobile.
We’re bringing the future of AI agents to Seeker: autonomous workforces, secure execution and a new economy powered by agents.
And we’re going to put solana:SKRbvo6Gf7GondiT3BbTfuRDPqLWei4j2Qy2NPGZhW3 at the center of it.
CLOCK IN. 🫡
AI agents are moving fast.
Voight makes every action observable.
Every decision.
Every tool call.
Every error.
Trace it.
Understand it.
Debug it.
Voight
Observability for AI agents.
Building on @solanamobile
Building the next generation of agents on @solanamobile.
From software to hardware.
Agents that can see, decide, and control drones and robots.
Agents running and coordinating hardware represent the indisputable future; they are the next frontier.
Galaxyhub Labs retweeted
Claude Opus 5.5 is now live on Voight.
It’s available as a model for every agent deployed on the platform.
Deploy your agent.
Choose Opus 5.5.
Run.
Replying to @claudeai
At its default effort setting, Opus 5.5 delivers frontier results for a fraction of the cost per task, often beating other models running at their highest settings.
It also generates output more than 30% faster than Opus 5.
Smart traces powered by GLM 5.3
We just shipped Smart Traces at Voight.
Every AI agent run is now a visual timeline: tool calls, model turns, tokens, and every step in one place.
When the waterfall isn’t enough, one click explains the trace in plain language.
OpenTelemetry native, powered by GLM 5.3
Debugging an AI agent shouldn’t mean digging through raw logs.
JET Europe in the top 👀🔥
We landed safely in Barcelona! 🇪🇸
Tonight, the local community meets Jito with @seenfinity and @LaFamilia_so.
Let’s make it a good one. 👀
Galaxyhub Labs retweeted
This is my new @Lid_onchain profile, and by the end of this post you should not be able to tell where one ends and the other begins.
That work now has a storefront, not just an audience. On my Lid page everything I make can be bought, shared or funded from one link. Anyone who shares it earns their cut, the split settles on-chain in the same transaction as the sale, and every result leaves a verifiable receipt. No holds, no middleman sitting on the money.
A profile that pays its own way. 👇
This is the right frame for the biggest AI story of the week.
Everyone is quoting the resignation letter to argue about timelines. The operational reality is simpler: agents already execute code and move money in production, and most of them run with less telemetry than a vending machine.
Look at what the infra teams are converging on: actions logged, decisions traceable, systems auditable while they run. That is not a safety wishlist. It is the same property crypto spent 15 years turning into infrastructure for money: public rules, verifiable behavior, nobody trusting anybody's word.
The alignment debate will take years. Making autonomous systems observable by default can start today, and the teams building that layer are quietly becoming the most important ones in the stack.
You cannot certify the behavior of a system whose actions are not logged, whose decisions cannot be traced and whose failures cannot be reproduced.
This is the problem Voight works on.
Our position is simple: any system operating autonomously, whether an agent, an LLM application or an entire pipeline, should expose full telemetry of its actions from the moment it is deployed. At minimum, this is a requirement for running safely in production.
The industry is debating whether increasingly capable systems can be controlled.
A prerequisite for that conversation is knowing, in detail and in real time, what those systems are actually doing.
That is the infrastructure we are building.
High-value events are coming to Spain; this time, it was @JET_EUROPA_ turn to bring content, opportunities, and surprises to Barcelona.
España es la siguiente parada. 🇪🇸
.@JET_EUROPA_ celebra su primera edición en Barcelona junto a @LaFamilia_so.
Jito, JET, JTX y la comunidad reunidos en un mismo lugar.
📅 15 de septiembre · 17:00
📍 Barcelona
Si estás por allí, nos vemos → luma.com/514mfb7h
La Familia is coming strong, with a community of builders willing to give their all in this @colosseum hackathon






















