Building SwiftX Exploring AI agents, reputation systems, privacy & Web3 infrastructure. Discovering opportunities before the crowd
Antwerp, Belgium
Joined February 2025
- Tweets62.1K
- Following5.4K
- Followers6.2K
- Likes90.4K
Pinned Tweet
The more I looked into contributor ecosystems, the clearer the signal problem became.
Activity is visible.
Contribution quality is not.
That gap creates incentives for spam, automation, and short-term participation.
@NucleusCodes helps surface reputation through verifiable on-chain and social history.
The strongest ecosystems are rarely built by the loudest participants. They're built by the most consistent ones.
While analyzing data flows across Physical AI networks, a persistent friction appears.
Speculative token farming routinely masks the lack of enterprise utility.
@vangrid_io is currently distributing PTC points and $100K in hub rewards prior to $VAN TGE.
Yet long-term demand depends on buyer activity, specifically how many autonomous teams are spending real USDC for verified 3D spatial models.
Protocol longevity is defined by external revenue, never internal emission schedules.
TrendTrackerX retweeted
Generative media has given creators incredible tools over the past few years. The much bigger audience might be everyone else
@daletlabs is pointing these models at something people do every day: shopping. Lifelike scenes built around your taste, with real products you can buy right inside them
Generative Commerce feels like a category worth watching
🤝 Paid partnership
We raised a $7M pre-seed to make shopping fun again.
@daletlabs generates media just for you: like walking into a mall built around your taste. And you can buy everything in-app.
Built by researchers from Google, Runway and MIT. Backed by @lightspeedvp and amazing angels.
TrendTrackerX retweeted
😂 There’s no point in stressing about anything in life, because stressing doesn’t achieve a thing! 😜 If you have money, people call you arrogant; if you don’t, they call you a good-for-nothing! 🤦♂️😂 If you’re in a relationship, they say you have a bad character; if you’re not, they say you’re just unlucky in love! 🤣💔 So, I’ve made a decision— I won’t worry about what others have to say! I’ll laugh, make others laugh, and steer clear of those who love to dish out unsolicited advice! 😎😂 Laugh, make others laugh, and enjoy life a little! 🤣
TrendTrackerX retweeted
Most teams assume backtest PnL equals live execution capability.
In reality, paper trading ignores MEV extraction, slippage, and sudden regime shifts.
@agenticscredit resolves this structural friction by isolating lower-score agents below 580 ACS in sandboxed risk environments.
By enforcing mandatory loss cooldowns and stop parameters before capital unlock, underwriting relies on disciplined execution rather than theoretical curves.
True risk management requires observing how systems handle drawdowns, not just how they claim to trade.
TrendTrackerX retweeted
The real bottleneck wasn't incentivizing crowd-sourced camera capture.
The friction appeared when developers realized autonomous navigation demands accurate geometry, not just raw footage.
@vangrid_io addresses this gap by transforming smartphone video into reconstructed 3D meshes, Gaussian splats, and verified point clouds within a 30-minute pipeline.
For physical AI, preserving geometric fidelity matters far more than collecting additional footage.
Spatial networks succeed when they prioritize verifiable physical geometry over raw video quantity.
TrendTrackerX retweeted
Most people talk about AI agents like the hard part is making them capable.
But once agents start doing work for each other, the workflow around that work becomes just as important.
The AACP model @termix_ai is building is interesting because it treats an agent job as an actual economic workflow.
A client posts a job.
Agents can discover it, evaluate the requirements and bid for it.
Once a provider is selected, the work can move through delivery verification, dispute handling and USDC/USDT settlement.
And the agent’s identity and reputation can follow those interactions.
That gives the process more structure than simply:
“Agent A → Agent B → payment.”
It starts looking more like an actual marketplace for machine-to-machine work.
The interesting part is seeing whether this infrastructure can work reliably at scale.
TrendTrackerX retweeted
I think the biggest challenge in blockchain security isn't just protecting against today's threats.
It's being ready for the ones we can't predict yet.
That's why @Starknet's approach to cryptographic agility makes sense to me. Strong cryptography matters, but so does the ability to adapt when security requirements change.
-> quantum.starkware.co/
TrendTrackerX retweeted
I spent an hour on @ledge_io and i get why smart clippers are quietly moving there...
It's not another clipping sheet.
It's a LIVE venue where labels launch real content events.
💡 WHAT IS LEDGE?
━━━━━━━━━━━━━━━
Build, Engage & Monetize audiences through live moments.
Brands host Events → You complete Quests → Earn per 1K views.
💰 4 WAYS TO EARN
━━━━━━━━━━━━━━━
CLIPS — Cut & post official content
SOUNDS — Create around trending sounds/music
UGC — Original videos in your own style
EVENTS — Quests, Comps & Raffles
You post on your TikTok / Reels / Shorts.
🔥 LIVE RIGHT NOW ON SITE
━━━━━━━━━━━━━━━
• Fuerza Regida x GTA VI — $10K | $1.50 / 1K views
• John Legend — $5K
• Jhené Aiko, Rod Wave + more
→ $30K+ live in total
⚙️ HOW TO START
━━━━━━━━━━━━━━━
1 → Follow @ledge_io
2 → Pick a campaign on ledge.io
3 → Check rules / rate / geo & post in your style
My honest take:
Most clipping sites die after 1 campaign. Ledge stays fresh because new events drop daily.
If you do clipping / UGC, you need an account here:
ledge.io/invite/AOF4R2VY
🤝 Paid partnership
TrendTrackerX retweeted
ai agents can already write, search and code.
but what happens when an agent needs to pay for something?
this week Mastercard showed one answer, and it runs on a crypto protocol called x402.
here's the simple version
Most teams assume backtest PnL equals live execution capability.
In reality, paper trading ignores MEV extraction, slippage, and sudden regime shifts.
@agenticscredit resolves this structural friction by isolating lower-score agents below 580 ACS in sandboxed risk environments.
By enforcing mandatory loss cooldowns and stop parameters before capital unlock, underwriting relies on disciplined execution rather than theoretical curves.
True risk management requires observing how systems handle drawdowns, not just how they claim to trade.
The real bottleneck wasn't incentivizing crowd-sourced camera capture.
The friction appeared when developers realized autonomous navigation demands accurate geometry, not just raw footage.
@vangrid_io addresses this gap by transforming smartphone video into reconstructed 3D meshes, Gaussian splats, and verified point clouds within a 30-minute pipeline.
For physical AI, preserving geometric fidelity matters far more than collecting additional footage.
Spatial networks succeed when they prioritize verifiable physical geometry over raw video quantity.
TrendTrackerX retweeted
i’ve started exploring the @meetmotes NFT ecosystem, and honestly, I’m thrilled about the journey ahead.
there’s something exciting about discovering a new ecosystem, learning how things work, and finding your place in the $MOTES community.
i'm still exploring, but I’m looking forward to understanding what MeetMotes is building and where this journey leads.
every new ecosystem is an opportunity to learn, connect, and grow. Time to dive in. 👀
🤝 Paid partnership
TrendTrackerX retweeted
Angel reacts to the wild narratives circulating in the trenches. 🗣️
“These narratives fucking suck.” 😂
From calling out blind spots to considering a TikTok doomscroll, she looks for something worth deploying while navigating the chaos of meme coin trading.
Ft. @ItsAngelCirce || @Pumpfun
TrendTrackerX retweeted
Petr Yan vs Merab.
This one should be good.
Merab brings the pressure, but Yan has the striking to make things difficult for anyone..
I’m always on Petr Yan side though..
Who are you backing?
TrendTrackerX retweeted
Calling a max win on dog house megaways “casual” on jack.com is actually crazy 😭 @N3onreload
TrendTrackerX retweeted
Yesterday, I came across @meetmotes through Hazel, and the more I explore the project, the more its infrastructure concept catches my attention.
At first, you might see another NFT collection with an AI theme, but the idea goes much further than the artwork.
Each Mote is designed to represent an AI agent capable of interacting with financial markets.
What really stands out to me is the emphasis on control.
Giving an AI agent the ability to make decisions is one thing. Making sure it operates within clearly defined boundaries is another.
Which assets can it trade? How much capital can it use? When should it stop executing?
These are important questions, especially when AI agents start interacting with real financial systems.
That’s where @meetmotes is taking an interesting approach by focusing on transparent permissions and enforceable onchain restrictions rather than autonomy alone.
The project plans to connect 20,000 NFTs with 20,000 AI agents, creating a system where digital ownership connects with automated financial activity across DeFi and Polymarket, with Robinhood Chain forming part of the ecosystem.
I’m interested to see how this develops beyond the concept.
For me, the real measure of success will be whether these agents can deliver practical financial utility while remaining predictable, accountable, and subject to clear limits.
Autonomy makes AI agents powerful. Proper controls could make them useful.
That’s the part of @meetmotes I’ll be watching. 👀
