Exploring Web3 beyond the noise. NFTs | On-chain | Research
Joined May 2009
- Tweets1.6K
- Following861
- Followers685
- Likes1.2K
My mother asked me last year what would happen to the crypto if she died tomorrow. I did not have a good answer.
The envelope with twelve words was in a safe deposit box. She did not know the bank. She did not know the box number.
That conversation ended badly for both of us. Paper seed phrases force conversations nobody wants to have.
Tangem's seedless setup removes that conversation. The key is generated on the secure chip and never leaves as words I can lose or leak.
I gave her one card. I kept one card. A third card sits with my brother. Two of three can recover the funds.
No explanation needed. No envelope to find. No drawer she would not understand.
I read the incident notes honestly. A 2021 app bug, a 2025 access code attack needing the physical card. Boring means tested, not promised.
Code VATAN takes 10 percent off: tangem.com/de/pricing/?promo…
The quoted post from @KimMinJae walks through the seedless setup and the 2-of-3 backup. The factory firmware stays locked at production.
Who currently knows where your backup seed phrase is stored, and do they know how to use it?
Two hardware wallet stories shook self-custody in 2026.
What happened, and why I’m bullish on @Tangem 👇
🧊 July: Coldcard
A firmware bug had been weakening seed generation for years. Attackers recomputed the keys offline. 1,700+ BTC stolen from thousands of addresses, and nobody had to touch a single device.
📦 October: Ledger
Ledger is investigating wallet drains among people who bought from one authorized reseller in Southeast Asia. It asked that reseller to pause sales. The cause is still unclear and no tampering has been confirmed.
No victory lap here. People lost real money.
Both stories point at the same two spots: how your key is born, and where your device comes from.
Why I’m bullish on @Tangem:
▸ Seedless setup: the key is generated on the EAL6+ chip inside the card. No seed phrase to write down, photograph or leak.
▸ Firmware is fixed at the factory. No update channel for anyone to hijack.
▸ Backup is a second or third card, not a sheet of paper.
▸ Audits by Kudelski Security (2018) and Riscure (2023).
▸ I know of no documented case of funds stolen through a hacked Tangem card.
The honest part: no wallet is flawless. Ledger’s own research team has shown a laser attack on Tangem cards. It needs the physical card and a lab worth around $250k, and fixed firmware also means no patch.
So my rules stay boring:
Official shop only
Backup cards stored apart
Card missing? Move the funds with the backup
Boring is bullish.
What’s your one non-negotiable?
I stopped asking leverage protocols "what multiple do you offer" and started asking "what happens to the carrying cost if I hold this position for twelve months." Most protocols cannot answer the second question. @2FactorFinance can, because the structure makes the carrying cost visible from day one.
The protocol works with Coinbase's tokenized stocks ecosystem on Base to bring Kelly-optimized leverage onchain. Instead of choosing a multiple and hoping it fits every asset, it derives each one from the asset's own volatility characteristics. Bitcoin volatility is the demanding case, and the math closes the productive band around 1.33x for cbBTC. An equity with calmer volatility gets a different answer entirely.
That derivation is why education sits beside social activity and referrals as one of Season 1's three ways to earn Marks. The program rewards understanding the carrying cost structure before promoting the output.
No purchase, deposit, or holding improves your standing. Marks hold no cash value, move between no accounts, and represent no claim on any future token. When Season 1 closes, the ten highest-ranked eligible accounts divide a 1 BTC pool paid in cbBTC.
The disclosure travels with the link: because the referral is mine, if you join through it and qualify, I may receive Marks.
The carrying cost question is more useful than the multiple question. If you would rather trust a protocol that can answer both than one that can only answer the first, start here and tell me which asset's carrying cost you think more protocols should make visible:
points.2factor.finance/r/r6g…
Maya retweeted
Picking a collaborator from a profile feels like hiring from a résumé. Useful start, but definitely not the whole interview.
A wallet can show past activity. A post can show how someone explains a project. Neither alone tells you how they’ll work with your team.
Nucleus combines on-chain activity and verified social footprints across Profile, Contribution, and Reputation views. Each offers a different clue.
Need someone to test a product? Explain a mechanic? Reach a niche community? The right fit depends on the task, not just a number beside a name.
That’s practical. It gives a founder a place to start checking, instead of relying only on follower counts or a polished bio.
Still, a profile can’t replace a conversation. Ask for one concrete example, then see whether the person’s explanation matches the trail.
That’s what I’m exploring with @NucleusCodes: a clearer first look before an opportunity begins, without pretending a score can make the decision for you.
Maya retweeted
GM frens ☀️
Today is the last day of Axis Epoch 2 on Kaito.
I’ve enjoyed writing about @axisrobotics this round, but there’s another Axis deadline today that’s worth knowing about too.
The 12:00 UTC snapshot is for trajectory contributions.
If you’ve completed trajectories, make sure they’re signed before the snapshot. Axis says signed contributions recorded before then count toward the Pre-TGE contributor reward pool.
So there are basically two things I’m checking today:
Finish the Kaito Epoch strong.
And if you’ve contributed trajectories, don’t leave them sitting unsigned.
Different programs, same deadline.
Good luck to everyone wrapping up today 🫡
@axisrobotics $AXIS
s.kaito.ai/u6sri4N
Maya retweeted
2Factor Finance is now live on Base.
The public beta brings Kelly-optimized leverage on-chain, beginning with cbBTC. @2FactorFinance works with Coinbase to extend the same approach toward its tokenized stocks ecosystem.
The structure separates an asset’s volatility into two perpetual tranches: a senior side with deep downside protection, and a junior side carrying leveraged exposure without a liquidation clause or external hedging counterparty.
Sizing follows the asset. Bitcoin’s volatility near 60 percent puts the junior target around 1.33x; lower-volatility equities can support a wider range. Bitcoin is where the mechanism is being proven first; tokenized equities are the next use case.
I joined the Season 1 Points Program. Marks come from verified social, education, and referral actions—not purchases, deposits, or holding. When Season 1 ends at launch, the leaderboard freezes and the top 10 accounts split 1 BTC, paid in cbBTC.
Join here and grab your referral link:
points.2factor.finance/r/mv3…
What would you want to see demonstrated before tokenized equities receive the same structure?
Maya retweeted
The number that stuck with me from the film wasn’t 30 years or 75 years. It was 14.
Lee Castleton’s branch was short by about 23,000 pounds according to a computer. He didn’t take the money. He couldn’t afford a lawyer so he fought it himself. The court made him pay 321,000 pounds in legal costs. That’s fourteen times the amount they were arguing about. He went bankrupt. Later it turned out the system was giving false numbers the whole time.
Nobody had to be corrupt for this to happen. The cost of being right was just higher than most people can handle.
Now think about AI agents making millions of small deals. Most of those disagreements will also be based on machine records. If it still costs more to prove you’re right than the deal is worth, the honest side will keep losing by default.
That’s why I keep looking at what @GenLayer is trying to do. Independent validators checking the same evidence and deciding in minutes changes the whole cost structure.
Justice has always worked better for people who can afford the fight. The upgrade has to make it possible for regular people to be right without going broke.
#JusticeNeedsAnUpgrade
Slow. Expensive. Wrong.
It's been this way for centuries. It doesn't have to be.
#JusticeNeedsAnUpgrade
Maya retweeted
I never finished the notebook I started in 2022.
Half the pages are still blank.
The last thing I wrote was a reminder to be kinder to myself.
I never read it again.
@RallyOnChain
If someone found the things you left unfinished, which one would tell them the most about who you were?
Maya retweeted
One thing I’d love to see more of in crypto is better privacy without losing the ability to prove what’s true.
Imagine using a financial service on a blockchain without showing everyone your financial activity, while still being able to prove the details that actually matter.
I think this could be especially useful for businesses and financial institutions. They need to protect sensitive information, but they also need ways to check that important claims are valid.
That’s what makes @primus_labs interesting to me. Cryptographic proofs can help verify information without requiring all the underlying data to be exposed.
If this approach develops well, it could help more businesses feel comfortable using onchain finance without putting all their financial information on display.
For me, that’s a future worth building toward.
Link: s.kaito.ai/vsO1Itp
Maya retweeted
If you want to learn patriotism, learn it from @KaitoAI
Accounts with 1.5k to 5k followers have made it into the inner circle but almost all of them are from Korea and Kaito is Korean too 😂
This is how you prioritize your own country. Learn something!
Kaito Global ❌
Kaito Korea ✅
Maya retweeted
I always laugh when builders ask language models to count things and the models fail with total confidence.
Language models guess the answer.
Ask one how many times the letter r appears in the word strawberry and it will give you the wrong number while sounding completely sure of itself.
@GenLayer approaches this differently because the network lets smart contracts write and execute tiny programs on the spot.
When a contract needs an exact answer, it does not hallucinate.
It writes the code, runs it, and uses the actual result.
The hard part about building agent economies is not making models talk to each other.
It is making them do exact math without guessing.
What is the one manual verification step in your current workflow that an intelligent contract could automate by writing its own script?
Maya retweeted
I had one job:
sell. FOMO got me into a meme coin at the top, and the same FOMO kept me holding on the way down because I was sure it would pump again.
It didn't. @RallyOnChain , that's the trade that taught me exits matter more than entries.
I still check the chart sometimes. Not to buy back in, just to remember.
🤝 Paid partnership · 🤖 Made with AI
Maya retweeted
Honest take on @starlisting so far:
+ Public rules: daily caps, -20% for each extra post
+ AI check rewards specific writing over hype
- Only top 10 get paid on this first launch, 200+ creators fighting for it
Worth the grind? Curious what you think.
Maya retweeted
This is actually lowkey brilliant.
Everyone’s out here stressing about portfolios and charts while Quant just goes bet, give me $50 and a dream.
New prompt I just tried:
Turn my $30 weekly coffee budget into enough for a spontaneous 3-day trip to Bali. 😎
Curious how wild this thing can get with the right prompt.
You can check risks yourself: app.tryquant.io/?ref=ST8NM9Q…
@tryquantio #QuantAIPioneers
Maya retweeted
Just checked my Sorsa Score… 612
Quality > quantity.
Feels good to see the real network showing up.
Real growth, no shortcuts
Building in public + open for solid collabs.
What’s yours at?
Drop it below 📷 @SorsaApp
Finally hit 500 @SorsaApp score 🔥
Tier 3 Credible unlocked
Real growth, no shortcuts
Building in public + open for solid collabs.
Share yours 👇
Maya retweeted
Joined as Founding Star #257 on @starlisting
Early creators post about vetted launches and get rewarded before the wider crowd.
First 500 get Founding status, rankings don’t reset, and no X login required.
Still early enough that the spots matter.
Join here: starlist.world/r/miftahudins…
🤝 Paid partnership
Maya retweeted
I've ghostwritten hype for enough Web3 projects to know the real bottleneck was never the idea, it was distribution. Every founder I've worked with eventually hits the same wall: paying an agency a serious budget for hype nobody can actually verify happened the way they claimed.
Permissionless Campaigns just removed that wall. Anyone, a funded project or a solo builder, can launch a real campaign on @RallyOnChain in about five minutes, funded with stables or their own token, no approval process deciding who's allowed to try.
The flow is the actual unlock: write the brief, fund the reward pool, it sits in escrow until real content gets submitted against it, then goes live the second it's funded. No three weeks of back and forth with an agency before a single post exists.
Rally was already the first AI-powered decentralized marketing protocol, scoring submissions instead of just taking someone's word for the reach. Now live on Robinhood Chain too, which means this isn't staying contained to the usual crypto-native crowd.
What would you actually launch if starting a real campaign took five minutes instead of five meetings?
