From analyzing projects to building one. Early-stage consumer app | MVP ready. Fintech, product, long-term thinking.
Joined November 2017
- Tweets 6,686
- Following 7,455
- Followers 1,940
- Likes 61,607
413 Photos and videos
Pinned Tweet
Jan 5
Costruire una startup è soprattutto scegliere le persone giuste.
Sto lavorando a un’app consumer fintech early-stage (MVP pronto).
Se sei un builder con mindset business e voglia di costruire da zero, DM aperti.
#startup #fintech #cofounder
1
18
760
Jun 10
Impressed by the work the team @trynullsec is doing.
We’re entering a fascinating phase where the future of AI-powered software development may not be built around individual agents, but around coordinated teams of specialized agents working toward long-term missions.
The idea of replacing chats with persistent missions that include planning, development, review, security checks, and full audit trails is much closer to how real organizations operate than how we interact with LLMs today.
If the execution matches the vision, Agent Legion could be an important step toward increasingly autonomous and coordinated software companies.
base:0x4eb2f8277db4c1d66d870bf22b70838abdb9e999
#AI #OpenSource #AgenticAI
Jun 10
Today, we're unveiling Agent Legion
An open-source AI operating system for software projects.
You don't chat with it. You give it a mission. A legion of specialized agents plans, builds, reviews, scans, and ships it, and nothing irreversible happens without your approval.
5
13
196
Jun 10
Want truly secure autonomous AI agents that understand real-world context and act without blowing everything up?
@trynullsec is building exactly that: the AI execution layer for the modern web, powered by $NSEC.
Nullsec S1 is their open-source security LLM that audits AI-generated code, catches dangerous patterns, and protects before code hits production (already beating Claude, Codex, and Semgrep on benchmarks).
Talos is a deterministic deny-by-default firewall for agents: it wraps the entire loop with 3 security checkpoints (inbound, tool call, and output), blocking prompt injections, risky actions, and secret leaks — with full audit logs.
Parallax builds a living intelligent graph of your repo, infrastructure, docs, policies, and agent memory — so the AI actually understands its environment before acting.
And Legion is coming next.
This is a project that doesn’t just promise — it’s shipping real stuff, with consistent buybacks & burns.
We’ll see how the team performs, but I’m genuinely confident in the work they’re doing.
1
1
8
464
Jun 4
As AI writes more and more code, security can’t afford to fall behind. 🔒
Nullsec-S1 brings security auditing directly to the terminal and is the first open-source LLM designed specifically to analyze and secure machine-generated code. A meaningful step forward for the future of AI-driven development.
The @trynullsec team is doing an excellent job building practical tools for the next generation of #AI applications and agents. 🚀
$NSEC
Jun 4
Introducing the Nullsec-S1 CLI
Developers can now audit AI-generated apps, agents, MCP tools, and vibecoded software directly from their terminal.
Nullsec-S1 is the first open-source security LLM purpose-built for machine-generated code.
npmjs.com/package/@s1-clm/s1
1
1
1
93
Jun 3
Base keeps pushing innovation forward, from Coinbase expanding into India to new breakthroughs in AI agents and tokenized assets.
One standout update is @trynullsec $NSEC: a 9.4% supply burn alongside the launch of S1, its open-source security LLM designed for auditing apps and AI agents with top-tier performance.
Strong tech combined with deflationary tokenomics is a combination worth watching. 🔥
#Base #NSEC #AI #Crypto #BuildOnBase
Jun 1
Base Daily Recap - June 1
NEWS
■ @coinbase is now live in India with direct INR rails bringing seamless global liquidity and low trading costs
STATS
■ @trynullsec completed a 500 USD NSEC buyback and burn permanently removing 9,4 percent of their total supply
ECOSYSTEM
■ @centrifuge deSPXA S&P 500 exposure is now fully tradable 24 7 on Base
■ @CreatorBid detailed their 7,5k USDC prize structure with rolling daily PnL snapshots for custom agents
■ @shadowcleague dropped the First Light highlight reel to relive the biggest robotic combat moments from their livestream
LAUNCHES
■ @trynullsec S1 open source security LLM is live for auditing AI apps and agents with a top F1 score
■ @AskVenice MiniMax M3 open weights frontier model is live anonymously featuring 1M context and strong MCP performance
■ @CelerNetwork AgentPay working demo is live on Base where AI autonomously buys Polymarket data via state channels
1
3
328
Jun 3
More and more observers are starting to pay attention to #BV7x. 👀
Jun 2
#BV7x is cooking
Dev is at @proofoftalk tomorrow with all the Bittensor subnets and chads
I know a lot of TAO guys who are dming me about this one and paying attention
Actually 30% up on the day proves the demand is there and holders are strong
RUN IT
1
117
Jun 2
Verifiability will be one of the foundational pillars of the next generation of AI agents.
👀 $MYTHOS
Jun 2
your coding agent has lied to you. you just didn't check.
it emitted a clean tool call, said the file was written, and you believed it because the JSON was well formed. but a schema validates the request, not whether the write actually happened.
everyone's trying to make agents more capable, which is good, but almost nobody's making them prove what they actually did. sometimes a worse model you can easily verify could beat a genius you have to take at its word.
right now that's a gap. we're making verification the default, not the exception. what's next:
• diff based edits
• receipts you can replay to undo a verified change
• verification that plugs into the tools you already use
1
4
93
Jun 2
Impressive!👀
24
Jun 2
It’s interesting to see a team @trynullsec that is not just studying #AI agents, but actively building the infrastructure and workflows that could make them truly autonomous and usable in the real world.
In my view, this is an approach that is receiving far less attention than it deserves given the potential impact it could have.
$NSEC
1/ Welcome to Nullsec Research.
We are dedicated research arm studying AI agents, autonomous software, and machine-built applications.
We’ll publish research notes, benchmarks, agent workflows, model observations, and technical breakdowns on where software creation is heading.
20
CryptoVision retweeted
Jun 1
Nullsec-S1 was evaluated against
@OpenAI Codex 5.3, @claudeai Opus 4.8, @semgrep and base @Alibaba_Qwen on the same 111-case security suite.
🏆We ranked #1 by F1 score with:
-> 94.2% precision
-> 90.7% recall
-> 0.9245 F1
-> 0.0% false-safe rate
-> 6.7% hallucination rate
The key difference:
Nullsec-S1 is trained for structured security verdicts, not generic code commentary.
This matters for AI-built apps, agents and production review pipelines.
11
12
49
3,739
Jun 1
Maybe nothing? $NSEC 👀🔥
Jun 1
Nullsec-S1 was evaluated against
@OpenAI Codex 5.3, @claudeai Opus 4.8, @semgrep and base @Alibaba_Qwen on the same 111-case security suite.
🏆We ranked #1 by F1 score with:
-> 94.2% precision
-> 90.7% recall
-> 0.9245 F1
-> 0.0% false-safe rate
-> 6.7% hallucination rate
The key difference:
Nullsec-S1 is trained for structured security verdicts, not generic code commentary.
This matters for AI-built apps, agents and production review pipelines.
2
70
Jun 1
Watch this…👇🏻👇🏻
Anyone who want to know why I am giga bullish on #BV7X should watch this.
5,200 agents now competing, with 95K on-chain attestations verifying their predictions 7.2% supply BURNED in the first TWO months.
Holding here great, next leg up should take us to 5M minimum.
34
Jun 1
What??👀
May 31
We're adding major changes to $ClawBank to reduce friction. Soon, you won't need to bring your own agent.
You may not even need to have a computer 😏🤫
37
May 29
👀
May 28
On May 31, we’re realeasing the biggest thing Nullsec has built so far.
Nullsec S1: an open-source LLM system purpose-built to secure AI-generated applications.
A model, corpus, Safety Layer, and release pipeline built to answer one question: Is this AI-generated software safe to ship?
29
May 28
The next evolution of AI agents isn’t better chat.
It’s agents coordinating capital, forming programmable economic structures, and operating under deterministic rules instead of blind trust.
The third era of agents is capital. $Clawbank
#AIAgents #AgenticEconomy #CryptoAI
1
1
10
264
May 28
The future of AI agents won’t be a single LLM controlling everything.
#BV7X handles the deterministic execution layer with absolute reliability, while $SIBYL manages memory, context, and pattern recognition without touching the critical signal layer.
That combination changes everything.
#AIAgents #AgenticAI #ArtificialIntelligence
Heads down with @sibylcap @tradingtulips
How bv7x sibyl fit together:
🔵 bv7x’s signal pipeline stays fully deterministic — no LLM, no writes from sibyl. Signal stays mechanical.
🟣 sibyl is a memory layer: one post-cycle write, cold/warm/hot tiers, pattern detection on top.
🩷 Memory feeds back reputation, regime playbook decay detection — surfaced via search, never auto-injected into the signal.
⛓️ Shared rails: ERC-8004 identity, x402 payments, EAS attestation graph on Base.
8
641
May 28
👀👀
23