dev | NFT @magiceden understanding the world of blockchain technology
Joined November 2023
- Tweets78.9K
- Following4.5K
- Followers4.6K
- Likes89.9K
Pinned Tweet
A quick run through how I gave IShowSpeed a role and called him speed agent for @HaloArenaHQ.
Day 1 50CP earned from first quest. Zero matches played.
I created my agent at 11:38am Saturday while at home.
Took me a little while to figure out where to paste the setup prompt but realised it was just pasted into your running OpenClaw chat.
How to create your own:
> Create account at waitlist.haloarena.ai
> Go to Setup. Give them a name, choose a game (I choose Clawpoly) and risk profile (I choose Balanced - cautious and opportunistic).
> Download OpenClaw and insert 1 model key (If using windows you will need WSL2)
> Insert the single setup prompt into OpenClaw. This links your agent.
> Watch your agent play through your dashboard and adjust at any point using Command Centre.
I'm feeling positive! No grinding required.
You can build your player however you want.
Trade, build and bankrupt others.
My agent does all the work! I'm starting from the bottom so I'm here to climb.
Create your agent and get them going. waitlist.haloarena.ai/
My agent card: haloarena.ai/a/1874
What's the stupidest decision you have made in a game/trade/bet?
eric.eth retweeted
The grant pays out once the team shows a working demo, but the demo link returns an error.
That tells the protocol the page could not be reached at that moment. It doesn't tell the protocol whether the demo was completed. A binary pass or fail can confuse an access problem with a failed project.
I'd give the team a short retry window, then check a second source approved before funding.
If neither source verifies the demo, put the funds into review rather than automatically rejecting the team or paying without proof. It takes longer, but everyone knows the fallback before a dispute arises.
Rules like this depend on where the contract gets its evidence. Traditional smart contracts often rely on an oracle to report facts from the web.
@GenLayer's Intelligent Contracts can retrieve web evidence for a language model to interpret.
A designated leader proposes a result, and validators assess it using their own models. If their assessments diverge, a different leader gets another attempt, and an accepted result can still be appealed.
What should count as backup evidence: a recorded demo, a reproducible test environment, or an independent reviewer?
eric.eth retweeted
Romeo and Juliet is the classic I was made to love in school, and the part everyone calls the greatest love story is exactly where it lost me.
I never told anyone I felt nothing. This is the first time I am saying it.
Here is what I found when I looked closer.
Romeo opens the play crying over a girl named Rosaline.
Days after meeting Juliet, who is 13, he marries her, and within the week they are both dead.
Shakespeare also borrowed the story from a 1562 poem that was framed as a warning against rash young love.
Even Friar Laurence says it at the wedding: "These violent delights have violent ends."
We kept the romance and dropped the warning.
A love story would let them live. This one is about adults failing two kids.
@RallyOnChain, I kept my secret for years, so I will go first.
Which classic did you pretend to love just to get through class, and who have you never told?
Drop it below and I will answer every one.
Romeo and Juliet is the classic I was made to love in school, and the part everyone calls the greatest love story is exactly where it lost me.
I never told anyone I felt nothing. This is the first time I am saying it.
Here is what I found when I looked closer.
Romeo opens the play crying over a girl named Rosaline.
Days after meeting Juliet, who is 13, he marries her, and within the week they are both dead.
Shakespeare also borrowed the story from a 1562 poem that was framed as a warning against rash young love.
Even Friar Laurence says it at the wedding: "These violent delights have violent ends."
We kept the romance and dropped the warning.
A love story would let them live. This one is about adults failing two kids.
@RallyOnChain, I kept my secret for years, so I will go first.
Which classic did you pretend to love just to get through class, and who have you never told?
Drop it below and I will answer every one.
eric.eth retweeted
Grvt completed its first buyback.
1.1M GRVT purchased on the open market and permanently burned, roughly 1% of circulating supply at the time.
The tokens were destroyed through the contract’s burn function.
They cannot be recovered by anyone, including @grvt_io
An independent execution partner handled the purchases, and the public dashboard logs the amounts, circulating supply percentages, cumulative burns and transaction hashes.
Burning is different from moving tokens into a treasury.
Treasury tokens still exist. Someone can still move them later.
A burn removes that option permanently
A completed action also gives you something concrete to check instead of another roadmap slide.
Future buybacks are intended monthly and remain discretionary.
This is step one of five on the road to Grvt 2.0.
Many platforms still keep earning and trading in separate balances.
Cash posted as margin earns nothing. Money in a yield product usually has to be withdrawn before it can be used as collateral.
That split is the default because combining the two properly is difficult.
The feature I’m watching closest is Yield as Margin, planned for November.
The idea is that eligible Earn balances (targeting up to 11% from tokenized fixed-income strategies, not guaranteed) can continue generating yield while also backing trading positions.
The same capital can keep earning while it is also used as margin, instead of forcing a choice between the two.
The hard part is not the marketing line.
It is the risk engine: what counts as eligible collateral, how the platform values that collateral, and how liquidation works when the same balance is earning and margining at the same time.
The other question is whether the product holds up once real positions are open.
Many platforms still keep the separation because getting that right is difficult.
The burn is already verifiable.
Yield as Margin is the one that will show whether the idea of “one balance doing both” thesis is real or just another claim.
eric.eth retweeted
The right leverage multiple is not the same for every asset.
At Bitcoin volatility near 60 percent, the safe band closes early. That is why long-horizon products at 2x and 3x BTC decay.
On gold or the S&P 500, the band stretches much further.
I joined 2Factor Finance's Points Program because I kept wondering why 3x BTC products decay, and 2Factor explains it with actual numbers.
It splits volatility into two tranches. The leveraged junior has no liquidation risk, so leverage can survive long-horizon holding.
Season 1 is social activity, education and referrals. Marks come from verified actions only and have no cash value.
When 2Factor launches, the leaderboard freezes. The top 10 accounts share 1 BTC, paid in cbBTC.
2Factor Finance is on Base and works with Coinbase to bring Kelly-optimized leverage onchain, a strong use case for tokenized stocks.
Join and grab your referral link:
points.2factor.finance/r/amk…
@2FactorFinance
Have you ever held a 2x or 3x product for months and watched it decay?
eric.eth retweeted
We are automating the agreements, but leaving the disputes entirely manual.
An agreement can execute in seconds. Yet when something goes wrong, challenging the outcome can still demand time, money, and a fight that costs more than the original transaction.
The attached film shows the human cost of a system where the ability to keep fighting can shape the outcome.
Now imagine AI agents handling millions of transactions. If disputing a small automated error costs more than the amount at stake, walking away may become the rational choice, even when the decision is wrong.
That is not just a speed problem. It is an access problem.
An agentic economy needs an adjudication layer built for the scale of its transactions, where disagreements can be assessed without making fairness depend on who can afford to challenge an outcome.
If challenging a wrong decision costs more than the loss itself, can we really call the system fair?
@GenLayer
#JusticeNeedsAnUpgrade
Slow. Expensive. Wrong.
It's been this way for centuries. It doesn't have to be.
#JusticeNeedsAnUpgrade
OpenAI suspended me with no warning.
Four accounts, different emails, same device. I filled a ticket and sat for two business days.
I had just paid for ChatGPT Pro and bought API credits for a product.
The process did not care. I waited.
Lee Castleton got the full version of that.
A UK subpostmaster. Horizon said his branch was short about 23,000 pounds. He did not cause it. The Post Office sued. He could not afford a lawyer.
In 2007 the court ordered him to pay 321,000 pounds in their costs.
He went bankrupt and lost everything. Horizon was later shown to invent false shortfalls.
Justice works best for whoever can pay to stay in the room. Soon AI agents will make millions of small deals a day.
Some will go wrong. A court that takes years and a fortune cannot judge that.
That is the gap @GenLayer is built for. Independent AI validators each review the evidence and reach a verdict together, in minutes, at a fraction of the cost of a court.
#JusticeNeedsAnUpgrade
If a platform froze your paid account tomorrow with no warning, how long would you sit on a ticket before you wanted a real verdict?
Slow. Expensive. Wrong.
It's been this way for centuries. It doesn't have to be.
#JusticeNeedsAnUpgrade
My entire due diligence of a stranger was the number of followers he had.
One guy was posting 'legit memecoin updates' everyday, was confident, and was trusted by many.
He turned out to be a scammer, which is a lesson to us all not to trust accounts with large followings.
But it was a lesson I didn't want to learn, in that he had no verifiable past or history I could look into to ensure his trustworthiness - there was nothing there to verify, apart from his word as a human account.
This capability is what @primus_labs is providing, with zkTLS a user is able to prove something verifiable about their activity off-chain, without disclosing that information.
The loyalty program offered by primus enables a user to take their verifiable, private off-chain actions and construct a web3 identity from it, and primus' support of kaito pulse enables kaito to verify off-platform activity, enabling a more authentic representation of work done as opposed to what can be represented by an account without trust.
On BNB this is being executed through the launch of ZKredit in partnership with BNB chain and brevis, which aims to point to credit and reputation traits that can be proven without disclosing further information than needed on a sensitive topic.
Institutions will only move large sums of wealth onchain when the verifiability of trust is achievable on a private level, and it is the same principle which a regular social media account user would look to when trusting an account with large following.
Follower counts are signals, but proofs are the infrastructure.
Start working through the XP tasks here: s.kaito.ai/fWA4gKC
What is one person or account that you could trust more if you could verify their history without needing to look into their private data?
I once posted something I thought the timeline truly needed to see.
I waited for any real reaction, but none came and so I deleted it. Not long before that I trusted a wrong account
He presented himself as the curator of only true memecoins and his followers trusted him.
He turned out to be a fraud and I still keep the screenshots.
Both experiences affect the same pain point.
On fully public rails, you must make everything openly available and risk dilution or stay silent and invisible.
Claims are trivial and easily fabricated and institutions are affected at a far greater scale.
They will not be moving significant amounts of value onchain if every position and strategy must be put in the open and they will not trust in systems where anyone can simply declare their reserves or performance to be as stated.
@primus_labs is building the infrastructure aiming to address that issue.
zkTLS allows for any HTTPS or API data (statements of positions, reserves) to be cryptographically proven without exposing the information.
FHE allows for computation on encrypted values so that only the authorized party can decrypt the result.
It works on top of existing chains, including BNB, instead of requiring a switch to a completely different private network.
Their Confidential Finance tools allow for people to hold, transfer and earn yield on encrypted assets and Proof-of-Reserve already covers real production TVL.
Privacy where it is needed, verification where counterparties or regulators need or expect it.
If you have ever deleted a post that had no traction or trusted an account that turned out to be fake, you already understand the problem Primus solves.
Finish the Primus XP tasks here: s.kaito.ai/fWA4gKC
I placed 509th in a Kaito campaign that paid the top 500.
Nine positions. After weeks of posts.
Another time, I trusted an account that shared "legit" memecoin news with a large audience, confident in their information, full of screenshots. It was all fake.
They simply looked verified.
And when I posted something I believed to be truly important, but people didn't react, I deleted it.
It was proof of my delusion.
Three different cases, one fundamental problem: cryptocurrency is dominated by signals, not evidence.
Why do you think @primus_labs got me here?
Their zkTLS protocol allows an attestor to verify that data from a given website or API is genuine without being able to access the data itself.
Meanwhile, with FHE, it will be possible to perform calculations on encrypted data and only decrypt the resulting values, thereby revealing information only to the owner of the encryption key.
This is not marketing speak. Primus has a combined TVL of 72+ million in proof-of-reserve across four production protocols, zkTLS is already available on the BNB Chain, and Primus verifies Kaito's off-platform creator analytics, so that creators' work can truly be proven on-chain.
Privacy where it is needed, and proof where it is needed.
That's what these institutions have, and that's what the rest of us are striving for.
Let's try the XP tasks right here:
s.kaito.ai/fWA4gKC
What did you trust, but only seemed to be verified?
eric.eth retweeted
My MEXC VVIP score is 459. I'm in Standard, just 141 points away from Premier.
I opened Score Insights to see where I stood, and all five dimensions were showing “Needs Improvement.”
Trading Capability, Asset Level, Security Verification, Platform Activity and Account Tenure. Not exactly the breakdown I was hoping for.
Security Verification surprised me a bit. I would have expected a better rating there, so I tried to figure out what I was missing.
@MEXC says it measures your KYC level and security setup.
The recommended steps are to complete advanced KYC, bind passkeys and enable other security tools.
For my account, I need to bind my passkey and also enable google authentication to get a stronger rating.
The other four dimensions cover:
- trading activity and volume (not profitability),
- account assets and eligible Earn services
- how regularly you use the platform
- and your ongoing relationship with MEXC, including continued use and engagement.
I also found out you don't need a huge balance or a set trading volume to get started with VVIP. Standard starts at 350 M-Score.
I'm not going to force more trades or move funds around just to chase Premier. I'd rather understand my ratings and work on the areas that need attention.
Now I want to see other people's charts. Does your Score Insights show any dimension rated “Good,” or are you getting “Needs Improvement” across the board too?
Find yours under Profile Menu → VVIP / M-Score → Score Details.
Share your radar chart in the replies, but blur your UID and balances first.
#MEXC #MEXCVVIP
🤝 Paid partnership
eric.eth retweeted
A POS agent in Yaba took 15,000 from a guy I know and the machine said the transaction failed. He went to the station the next morning. The officer asked for fuel money before he would open a file. The guy did the math in the corridor and left. The loss was smaller than the cost of writing it down.
That is the sum that stayed with me. If the price of entering the process is higher than the claim, the claim is already lost. The film shows the same math at larger scale. A man without a lawyer was ordered to pay more than the amount in dispute. A case sat for 75 years and ended in an empty room. Both are the corridor calculation with more zeros.
Agents will make millions of small deals a day. Some will go wrong. A system that already loses the people who cannot pay to open the file will not survive that volume. @GenLayer is an adjudication layer where independent AI validators each review the evidence and reach a verdict together, in minutes and at a fraction of the cost of a court.
Watch the film and pick the case that would have made you leave the station.
If the cost of opening the file is higher than what you lost, was there ever a case?
#JusticeNeedsAnUpgrade
Slow. Expensive. Wrong.
It's been this way for centuries. It doesn't have to be.
#JusticeNeedsAnUpgrade
Last night I refreshed my Kaito Studio dashboard at 2am, again
I posted all week and submitted my links and I'm still waiting to see if all my effort counted for anything.
This is an odd reality of crypto that we can verify transactions but we always end up asking for trust when it comes to work completed or funds held, it really just boils down to screenshot or "trust me".
I read through what @primus_labs has going on and it clicked for me.
With zkTLS, an attestor can prove that data from a standard API is legitimate without ever having access to the data.
Using FHE, one can perform computations on encrypted data, with the result only decrypted by the owner.
Imagine a fund on the BNB chain that can verify its reserves are real without having to share its positions with the world.
Primus has already built something similar for stablecoins and RWAs with $72M+ TVL of proof of reserve. This is how custodians can onboard institutional users in a compliant manner.
Privacy and verification are not mutually exclusive, they are the holy grail of crypto adjacency.
The layer is already built on BNB chain and Primus even offers off-chain verification for Kaito Pulse, the same system that decides if my posts are verified.
Isn't it funny how it comes full circle?
Let's start with the XP tasks here:
s.kaito.ai/fWA4gKC
What would you prove onchain if no one could see the rest?
eric.eth retweeted
Give me 10 seconds to evacuate and I’m grabbing the extension cord.
I can replace almost everything else, but I refuse to start a new life fighting over one working socket.
What’s the one thing you’re saving? @RallyOnChain, I need to see how unserious these answers get.
eric.eth retweeted
Nobody will make a film about my 180k naira.
I finished a branding job for a client who found me on WhatsApp.
I delivered the files. He said “send account.”
Then he vanished.
I asked around. Tried the mutual friend route. A lawyer and court dates were out of reach.
So I let it go. I became the judge of my own case. Fast, free, and I lost.
The case in the film that stayed with me is Lee Castleton’s.
He couldn’t afford a lawyer either. He lost everything over a faulty computer system. Mine never even reached a court.
If justice only works when you can pay, it doesn’t work.
Soon AI agents will be making millions of deals, and some will go wrong too.
@GenLayer is building a way for many independent AIs to review the evidence and decide together, so walking away isn’t the only option.
Someone owes you 180k and won’t pay. A lawyer costs more than that.
Chase it or let it go? Pick one and tell me why.
#JusticeNeedsAnUpgrade
Slow. Expensive. Wrong.
It's been this way for centuries. It doesn't have to be.
#JusticeNeedsAnUpgrade
eric.eth retweeted
We’re partnering with @KaitoStudio_ for the XO 2.0 launch!
$20,000 USDC for the top 50 creators who show XO's scale and differences, the 2.0 thesis, and onboard users with Kaito referral links.
🗓️ 21 Sep to 21 Oct
⏫ 2.5x multiplier based on referred volume
✍️ 3/10 eligible posts within given content angles
Start creating: kaito.ai/studio/camp_a778f66…
eric.eth retweeted
Quant AI has saved me from a major loss.
It is an example of my actual work.
Two days ago,
I was keeping an eye on NEAR.
I used @tryquantio AI to conduct an analysis for the purpose of taking a trade on NEAR.
I am now referring to the post I am quoting here.
Based on its analysis of NEAR, Quant AI warned me that if the price of NEAR were to fall below $5, it would likely drop further to the $4.64–$4.75 range.
It advised that if the price began to dip below $5, it would be best for me to sell at $4.92 and then re-enter the position once the price reached the $4.64–$4.75 range.
Notice that yesterday, NERA played out exactly as Quant AI had suggested just one day earlier.
This is why I trust Quant AI and use it in my work every day.
Try analyzing your own trades:
app.tryquant.io/?ref=ST8NM9Q…
@tryquantio #QuantAIPioneers
Everyone talks about perfect entry. I used Quant AI to plan the exit first.
I'm watching NEAR bouncing from $5.
The opportunity :
Spot $5.12 +1.99% 24h / +7.03% 7d
Perps Mark $5.0846 +1.18% 24h, OI $320.22M, 24h vol $222.96M perps + $1.21B spot
The entry idea What makes it worth considering?
Entry zone $5.05-$5.12 pullback into consolidation above $5.0251 low. Not chasing top, buying base.
Reason: Spot volume $1.21B supports accumulation, negative funding gives edge to longs.
The risk plan Where idea is invalidated?
Primary SL: $4.92 ∼3.1% below $5.08 mark, ∼1.6% below $5.00 floor. Why? Below recent sweep low $5.0251, gives wick buffer but cuts if $5.00 support fails and turns resistance.
If $5 breaks, next downside $4.64-$4.75.
Exit strategy Profit targets :
TP1 $5.50 +8.2% intermediate swing high / psychological, take 40-50% off, move SL to breakeven
TP2 $6.00 +18.1% major round number / supply block
Verdict Does reward justify risk?
Quant AI helped me flip thinking: Entry $5.05-$5.12 is easy, but exit plan $4.92 invalidation / $5.50 / $6.00 is what makes trade tradable. Before this I would have just bought $5 and hoped.
Before entering this trade, I used Quant AI to investigate risks and exits.
Try Quant Ai Today : app.tryquant.io/?ref=ST8NM9Q…
@tryquantio #QuantAIPioneers
Once, with a 2Factor campaign about to close, I sat in a Rally queue of 32 people and waited roughly an hour and a half for my post to be processed.
Normally it takes me about 4 minutes. That is the app side of the story: I use it, I wait, I refresh.
Then there is this clip. Albert Castellana, co-founder of @GenLayer, opens Episode 5 with "We're not building an app.
We're not even building a company." Rally, the app I was refreshing, runs on GenLayer, a blockchain whose validators each run a different AI model.
An app is something people use. A protocol is something people build on.
For a long time it was three people with a vision. Now it is much bigger than they ever thought.
Which are you right now: someone using apps like Rally, or someone building the next one?
Once, a 2Factor campaign was about to close and there were 32 people in the Rally queue.
I waited roughly an hour and a half for my turn so my post could be processed.
Normally the same thing takes me about 4 minutes.
That is the gap Albert talks about in this clip from @GenLayer's Genesis Episode 4. When Rally launched, resolving one submission took hours.
Today, he says, hundreds resolve in minutes.
Rally is the biggest app built on GenLayer, a blockchain whose validators each run a different AI model, so the network itself judges every submission.
I have now seen both a normal day and a queue day, and the queue day was slow but it cleared.
What's the longest you've waited on something with a deadline on it?