Researcher | Content Creator - Ambassador - Collab
Joined July 2017
- Tweets72.9K
- Following6K
- Followers6.3K
- Likes75K
Pinned Tweet
Robotics keeps showing up as a 2026 crypto narrative, and @vangrid_io is where it gets real.
Crowdsourced physical-world data has a trust problem: who captured it, where, and was it real? Vangrid anchors captures onchain on Base, so provenance can be checked instead of assumed.
Bounties put capture requests out for specific places, and contributors get settled in USDC through the protocol.
Robots can't learn from scraped text. They need verified data from real places.
See the activity yourself: explorer.vangrid.io
🤝 Paid partnership
Got a friend with an old DeFi wallet they barely touch? Send them to check what it earned on @zerufinance's Zaps.
Past trades can count toward a starting zaps balance, so nobody needs to place a new trade first. Qualifying activity can add more later.
Zaps are planned to count toward an airdrop, but they are points, not a token payout, and final rules decide allocation. Referral earning conditions apply.
🤝 Paid partnership
The crypto industry has made incredible progress, but I still think everyday user experience deserves more attention.
We talk a lot about transaction speed, scalability, liquidity, and new blockchain technology. Those things matter, but the products people interact with every day matter too.
Wallets are a major part of that experience. Exploring @Tangem made me appreciate how a different design approach can bring a fresh perspective to self-custody.
The next stage of crypto adoption won't only depend on better networks. It will also depend on better experiences for everyday users.
#WhyTangem
Tangem's $10,000 Giveaway is back 🚀
Moving your crypto? Tell the world why you chose to keep your keys:
1. Post with #WhyTangem & tag @Tangem. Show your Tangem Wallet or App, and share why you chose self-custody and Tangem
2. Fill out the form & QRT this post
3. Get rewarded in BTC:
📹Video: $30
🧵Thread with photos: $20
📸Photo post: $10
72 hours only🔥
Daniel retweeted
Anyone who has watched a bot farm inflate a post already understands why GenLayer validators must lock GEN.
Spinning up an AI judge costs almost nothing. If getting a seat on a case also cost nothing, one person could fill the court with a thousand judges of their own, and the verdict would be one person's opinion a thousand times over.
So on @GenLayer every validator locks GEN before it gets a vote. Each needs at least 42,000 GEN of its own, its own infrastructure, and a place in the top 1,543. A seat cannot be faked in bulk anymore. And a validator caught cheating loses part of what it locked.
Likes can be botted for free. Verdicts should not be.
What else online would improve if every vote cost something?
Daniel retweeted
I designed a logo for a client last year. They demanded endless revisions, then claimed the final file was "corrupted" to get a full refund from the platform's auto-mediator.
I realized that an automated dispute system without real accountability is just a tool for bad-faith actors to steal work.
This is exactly why the @GenLayer staking model is a game-changer. Validators must lock 42,000 GEN before they can judge a single case.
This prevents anyone from spinning up a thousand biased AI models to rig the outcome. If a validator makes an honest error, they simply lose the fee for that round.
But proven fraud slashes their actual locked stake permanently. Furthermore, the jury expands with each appeal, making it exponentially harder for a bad actor to control the final verdict.
This is the exact vulnerability the freelance platform exploited, allowing a baseless claim to succeed without any real review. GenLayer makes judicial dishonesty mathematically expensive.
Have you ever had a client or platform steal your work through a fake dispute? How would a system requiring real financial skin in the game have protected you?
🤝 Paid partnership
GEN tokenomics is live.
How GEN is staked, spent and earned across GenLayer, all in one paper 👇portal.genlayer.foundation/g…
Daniel retweeted
Years ago, a neighbor tried to dominate our local housing committee. Since attending meetings was free, they just brought ten relatives to every single vote. They forced through a rule change simply because their presence cost them nothing.
The cheapest decisions are always made by people with nothing to lose.
That is exactly why @GenLayer requires validators to lock GEN before they can judge a case.
If AI judges cost nothing to spin up, one bad actor could stack the digital jury with automated nodes. By forcing every validator to lock their own capital, the system ensures the judge has economic skin in the game.
If a validator is caught cheating, they lose a percentage of that locked stake.
What is one online system you use daily that would be instantly fixed if the moderators had to put their own money on the line?
GM
Niantic Spatial uses billions of phone captures for robotics, and @vangrid_io adds on-demand capture.
Phones are already the biggest sensor network on earth. Vangrid's angle is collecting ground-level 3D data for a specific place when a robot or world model needs it, instead of waiting for random coverage.
Onchain provenance means each capture can be traced and checked, which matters when you train machines on it.
The network is live, so verify the activity yourself: explorer.vangrid.io
🤝 Paid partnership
Most wallets look the same until someone reads what they actually did onchain. That is the idea behind zScore from @zerufinance, which evaluates wallet behavior.
ZeruAI provides the wallet intelligence, and Zaps brings it to traders: connect a wallet with real trading history and see a starting zaps balance based on what it did.
That is why connecting a genuine trading wallet matters. Zaps are points, not a token payout, and airdrop allocation follows the final rules.
🤝 Paid partnership
today is Genesis day for @FungoLabs, and the part I keep thinking about isn't even the NFTs. It's the auction.
Normal NFT sales are a gas war. You watch the mempool, guess what everyone else is bidding, overpay out of panic, or get sniped at the last second.
Fungo's sale runs on sealed bids, powered by Zama's FHE. Nobody sees anyone's bid until it ends. Highest bids win, but everyone pays the same final price, so if you bid high and the clearing price is lower, the difference comes back to you.
That changes the whole feel of it. You bid what the NFT is worth to you, not what you think the crowd will do.
And what you win is sealed too. Art and traits stay encrypted on-chain, even the team doesn't know what you got, and only you decide if and when to reveal it.
Private bids, private NFTs. 3,232 Genesis, 500 free mints first, then 2,500 through the auction.
Curious how this plays out for $FUNGO and for the launchpad after. Not financial advice, just a mechanism I'd like to see work.
🤝 Paid partnership
Want to see what your old trades earned? Open the link, sign in, link the wallet you already trade from and read your starting balance on @zerufinance's Zaps.
You only sign an ownership message, and your seed phrase or private key is never needed. Once your balance shows, you can join a Clan from the same app.
Zaps are points, not a token payout, and airdrop allocation follows the final rules.
Physical AI has a data problem the internet can't solve, and @vangrid_io is building the fix.
Language models trained on scraped text. Robots and world models need fresh 3D data from real places, and nobody can scrape that. NVIDIA even built a Physical AI Data Factory Blueprint at GTC 2026 because real-world data is a constraint.
Vangrid turns smartphones into a capture network that collects ground-level data on demand for a specific place, with onchain provenance.
Not a concept either, it's live: explorer.vangrid.io
🤝 Paid partnership
I've minted a lot of NFTs, and the same thing always happens.
Mint goes live. Within minutes someone has scraped every trait, ranked the rarity, and posted a spreadsheet. You haven't even looked at yours and you already know if you "won" or "lost".
That's the part @FungoLabs is fixing with Launchpad.
Traits are sealed at mint. Only the holder sees what they got. No sniping the rare ones, no rarity sheet before you've opened your own.
For collectors, minting feels like a surprise again.
For creators, the launch is fairer from block one.
Applications are open if you're building a collection, and the $FUNGO campaign is live now for anyone who wants to get involved early.
privacy in NFTs always sounded niche to me. this is the first time it feels obvious.
🤝 Paid partnership
Trading on Ditto already? Check whether your activity qualifies for the $50K Ditto competition on @zerufinance.
Register on Zaps, link the wallet you trade from and track your progress on the campaign page. Registration closes [deadline], and qualifying activity is [what counts].
Competition rewards are separate from the planned zaps airdrop, and any projected payout is only an estimate.
Register here: app.zaps.wtf/campaigns/ditto
Plenty of Physical AI talk is just slides, but @vangrid_io is already live and verifiable onchain.
Web and Android have been live since July 2026. Right now the explorer shows [X] captures anchored on Base, [Y] bounties accepted and $[Z] USDC settled through the protocol.
Robots and world models need fresh ground-level 3D data from real places. Vangrid turns phones into a capture network that collects it on demand.
Don't trust my numbers, check them: explorer.vangrid.io
Daniel retweeted
Before I let any AI agent touch money, I'd ask three things.
Who can send it text? Anyone on X, in this case.
What can it do with that text? Move tokens, with no human in between.
How many independent checks stand in the way? One filter, one model.
That's how around $200,000 moved in May. A Morse code instruction, Grok translating it, and Bankrbot executing it.
It's called prompt injection: a command hidden inside something an AI was only meant to read.
Question three is where @GenLayer's Greyboxing works. Every validator rewrites the input first, with its own secret defenses, then runs its own model.
The video sums it up in four words: random jurors, hidden models.
Fooling one stops being enough.
Which of the three questions can your own AI tools not answer yet?
🤝 Paid partnership
Daniel retweeted
I stared at a duplicate $4.99 cloud hosting charge on my dashboard yesterday.
The support portal required fourteen verification steps just to open a ticket.
I did the math. Forty minutes of my life was worth more than five dollars. I closed the tab.
We buy back our time by ignoring small thefts.
This quiet surrender is the invisible subsidy keeping legacy commerce afloat.
An agent optimizing for zero-loss will never perform that calculation. There is no forty-minute threshold in its decision tree.
It will flag every two-cent rounding error across ten million API calls because the cost of proving it right is zero.
The cost of resolving that error in a traditional court exceeds the error itself by orders of magnitude.
That is where @GenLayer changes the math. Independent AI validators bring the cost of adjudication below the value of the dispute.
A two-cent rounding error becomes worth resolving when the hearing costs less than the error.
If the cost of justice drops to zero, does every grievance deserve a hearing, or do we still need a threshold for what is worth fighting?
#JusticeNeedsAnUpgrade
Slow. Expensive. Wrong.
It's been this way for centuries. It doesn't have to be.
#JusticeNeedsAnUpgrade
Daniel retweeted
The most honest frame in this video is the one where a validator gets fooled.
I trust a defense more when it admits where it can fail. This one shows the failure on screen, then the vote that overrules it.
The way it gets fooled has a name, prompt injection. It is an instruction smuggled into text that an AI reads, trusts and acts on.
Greyboxing on @GenLayer changes the first step. The text a user sends never reaches a model directly.
1. Every validator rewrites it first.
2. Each sets its own defenses and hides them.
3. Each runs its own model.
4. The others outvote one that is fooled.
Even then, one validator can still fall for it. The hard part is not building a guard that never fails. It is making sure a failure stays one vote.
Agents that read public posts and hold a wallet get tested in public. In May 2026 a transfer instruction hidden in Morse code was enough to move tokens worth around $150,000 to $200,000.
So the question is not whether an agent can be tricked. It is whether one trick is enough. With a single model and a wallet attached, one fooled model is one transfer.
If your money sat behind a vote, how many voters could be fooled before you stopped trusting the result?
something is coming soon
on @sight_hood
Creator Season 1 is live.
Create content about Sight and robinhood:0x238e40b75ae78a1388a14e517d855893e92e58db, earn points and compete for USDC and other rewards.
Your content. Your points. Your rewards.
Start here: creator.sighthood.com
Daniel retweeted
Creator Season 1 is live.
Create content about Sight and robinhood:0x238e40b75ae78a1388a14e517d855893e92e58db, earn points and compete for USDC and other rewards.
Your content. Your points. Your rewards.
Start here: creator.sighthood.com
Already have a group of traders? Create a Zaps Clan and build your community's standing on @zerufinance.
Members connect the wallets they already trade from, check their starting zaps, and the Clan tracks progress together. Zaps are planned to count toward a future airdrop, and a $50K Ditto trading competition is open too.
Not every member automatically earns passive income, and zaps are points, not a token payout. Final rules decide allocation.
Join my Clan, link your trading wallets and check your zaps
🤝 Paid partnership
