Founder of @harecrypta | Researcher with Alpha | Trader @lighter_xyz @variational_io @polymarket | Builder | @zscdao member
Joined March 2021
- Tweets 35,150
- Following 5,289
- Followers 12,521
- Likes 113,274
3,875 Photos and videos
Jun 13
Good Evening X!
Chilling… We play poker and watch the match between Qatar and Switzerland
Jun 12
GM X!
I sleep on the terrace and get up with the first rays of the sun!
The Aztecs won 2-0 yesterday, they would have been playing football for 1000 years, even before the English they played with balls-heads of killed enemies.
6
21
458
Jun 12
My Variational points this week: 0.90. My base points dropped 98%.
Last week18.
The week before 12.
I went heads-down vibecoding for a week and stopped trading.
The dashboard told the story instantly. But the drop taught me the thing most farmers miss.
✅ Miss a week and it's gone, permanently. No catch-up. That week's pie got divided without you. Consistency beats bursts, my 98% drop is the proof.
The OLP, the house market maker you trade against, has historically run up to 369% annualized.
The community vault isn't public yet, but it's next on the roadmap, 90% of earnings to depositors.
When it opens, you stop farming points for a maybe-token and start earning the actual spread the house collects, as real yield.
$VAR
5
32
1,362
Jun 12
GM X!
I sleep on the terrace and get up with the first rays of the sun!
The Aztecs won 2-0 yesterday, they would have been playing football for 1000 years, even before the English they played with balls-heads of killed enemies.
Jun 11
Good evening X!
Today I’m swimming without panties and watching the ships.
I did a lot of work on vibecoding, I downloaded all profitable wallets on all Polymarket markets,
I know the signs of a market maker, win rate, stability, ROI,
I understand how each wallet generates its profit.
9
23
834
Jun 11
Forward Deployed Engineer (FDE) - the hottest job in AI
I did it for 16 years before the name existed.
The demand is vertical:
224 open FDE roles across 39 companies as of late May, Palantir, OpenAI, Anthropic, Mistral, Cohere, vacancies up 800% year over year.
The reason for the demand is the most honest stat in AI:
MIT study reviewed 300 corporate AI projects and found 95% produced no measurable P&L impact.
The models work.
The demos impress.
❌ Production is where it all dies.
I ran an implementation consulting firm for 16 years, deploying ERP at PhosAgro, MMK, Rostelecom, General Electric.
The dirty secret of enterprise software was always the same:
✔️ Technology is 20% of the project.
✔️ The other 80% is people, processes, and the politics of change.
AI didn't change that equation. It made it worse, because now the demo takes a day and everyone believes the hard part is done.
What an FDE actually does, understand the client's process, design the solution, drag it through resistance into production, is what implementation methodology has done for decades.
Same discipline, new product. ERP then, AI systems now.
The scarcest skill in AI (Claude, ChatGPT) isn't model knowledge.
It's methodological thinking:
taking a system from a working demo to a process people actually run.
Researchers built the models. Now the market needs people who can make them land.
Jun 3
Anyone can get Claude to write a trading bot in an afternoon.
Surviving one is the skill nobody teaches.
I built delta-neutral funding bots with Claude and lost real money to bugs that never threw an error, they reported success in green text while the position bled.
Just published the field manual. 9 lessons, each with the exact pattern to steal:
✅ reduce_only or the bot flips you long
✅ order_status lies, verify the position with retries
✅ your PnL log counts deposits as profit
✅ isolate the wallet like you expect the key to leak
The failure mode is silent. 👇
1
13
417
Jun 11
Good evening X!
Today I’m swimming without panties and watching the ships.
I did a lot of work on vibecoding, I downloaded all profitable wallets on all Polymarket markets,
I know the signs of a market maker, win rate, stability, ROI,
I understand how each wallet generates its profit.
Jun 10
GM X!
The underpant season is opening!
I follow the market less and less, I hit vibecoding more and more, coding inspires!
4
20
781
Jun 10
The World Cup kicks off tomorrow.
The price discovery already started months ago.
Polymarket's World Cup Winner market:
$1.8b in volume before a single ball is kicked.
The largest single prediction market ever.
Liquidity pool: $352m.
This is the first World Cup of the prediction-market era.
Qatar 2022 happened before Polymarket went mainstream.
Now the tournament gets priced onchain, live, every match.
What the market says:
✅ France 16.2% vs Spain 16.0%, and they meet in the groupe stage
One group game will reprice a $1.8b market
✅ USA: $51m traded, 1% implied odds. Patriotism is a position
✅ The daily volume leaders aren't France or Spain. They're $0.001 longshots trading millions, lottery tickets with no path to the trophy
That last one is the actual lesson. Longshot demand inflates no-hope outcomes above fair value, the classic favorite-longshot bias.
The crowd pays for dreams, which means the edge sits on the other side:
fading the romance, not buying it.
7
3
27
396
Jun 10
GM X!
The underpant season is opening!
I follow the market less and less, I hit vibecoding more and more, coding inspires!
2
14
707
Jun 5
GM X!
Green days!
How we all miss them, a stable green background!
Jun 1
GM X!
June 1. The Law of Samsara.
The important ones are boring to watch.
2026 is a quiet crypto year. Flat price, no catalyst, retail asleep. By that metric, you’re right.
But quiet is where the real work happens.
While the price chart did nothing, the settlement layer of finance got rewritten underneath it, prediction markets passed sportsbooks, FX perps went live, commodities and equities moved on-chain. All while nobody was watching, because everyone was watching the price.
Loud markets make you money.
Quiet markets make you early.
GM to everyone who knows the difference and $HYPE goes to $100.
12
1
26
753
Jun 4
Price discovery left the IPO. Onchain price is important!
SpaceX IPO big, pre-IPO perps exist.
The real story that this is the first $1T company that got price-discovered onchain before a single real share trades.
1. The bankers' number: $135/share, $1.77T,
listing June 12 on Nasdaq as $SPCX.
2. The onchain number:
synthetic SPCX perps across a dozen venues (Hyperliquid, Variational, Lighter launched) trade near $200, around $2.4T.
Polymarket market leans to the $2.2T (63% possibility)
3. The crowd is pricing SpaceX 35% above the bankers.
Watch what just happened to the IPO:
✅ Price discovery used to BE the IPO, the roadshow set the price, the first trade revealed demand
✅ Now it's already done, onchain markets have been discovering SpaceX 24/7 for weeks while the bankers were still printing the prospectus
✅ June 12 isn't the discovery anymore. It's the settlement, the moment the onchain price and the order book finally reconcile
That's TradFi's crown jewel, and onchain just took the first swing at it.
The crowd can be wrong, leverage makes perps loud, and a $2.4T print leaves longs underwater if the IPO opens at target. The synthetic market is loud. The order book speaks last.
But whatever prints June 12, the IPO stopped being where the price gets found.
That already happened somewhere else.
Jun 2
Who is Brevan Howard for Variational?
The most important name in Variational's $50m round isn't Dragonfly. It's Brevan Howard.
Most coverage buried it under and other investors.
But a $35b macro hedge fund doesn't write a perp DEX Series A check for the 10x. Crypto VCs do that. Macro funds invest to get positioned.
This summer Variational routes 100 TradFi markets (equities, forex, rates) directly from institutional dealers via RFQ.
Quotes come from registered, whitelisted LPs, not an anonymous pool.
Someone has to BE those dealers.
Brevan Howard trades FX, rates, and commodities at institutional scale.
They invested to become a liquidity provider on the rails they now partly own.
✅ own the equity
✅ quote the Phase 2 flow
✅ capture the spread
Three revenue lines from one check.
The founders' moat was always relationships.
This is the first dealer relationship, already signed, already funded into the cap table. $VAR
4
26
425
Jun 4
SpaceX IPO closing market cap above ___ on Polymarket? $SPCX
polymarket.com/event/spacex-…
1
9
125
Jun 3
Some Forex observations after a week of trading.
I've been trading $EUR/USD perps on both Hibachi and Lighter
Same pair, two completely different markets.
✅ Lighter: 0.00% funding → $6.00m OI parked, barely turns over (0.11x/day)
✅ Hibachi: 0.0096%/8h funding → thin $286k OI but churns 11.8x/daily volume
Lighter holds 21x the open interest but turns it over 100x slower.
Why? Funding design.
1. Zero funding on Lighter means it's free to hold, so positions get parked and OI piles up. Positive funding on 2. Hibachi means carry costs money, so nobody parks, the book churns intraday instead.
Same asset. Opposite microstructure. Entirely a product of how each venue prices carry.
The FX-perp map isn't one winner.
It's a market splitting by carry design before most people noticed it started.
May 29
Everyone's reading this as Hibachi added FX.
That's the surface.
Here's what nobody is pricing:
this is the first FX venue where you can write a smart contract that auto-hedges treasury.
Traditional FX in 2026:
✅ T 2 settlement
✅ Bank intermediation
✅ Prime broker relationships required
✅ $5-15/lot rollover fees
Stablecoin-native FX on @hibachi_xyz:
✅ Instant settlement (Circle Arc)
✅ Self-custody ZK privacy
✅ Composable via smart contracts
✅ USDC-native, no fiat rails needed
The trade isn't faster FX. The trade is FX as code.
A neobank paying out in 12 currencies could auto-convert daily revenue without ever touching a bank.
A treasury could hedge programmatically by event trigger.
Three venues launched FX perps in 6 weeks (BitMEX, Kraken, Hibachi).
Only Hibachi is institutional programmable.
Some advice for farmers:
✅ FX perps earn more points.
✅ Diversity bonus applies.
4
3
22
1,131
Jun 3
Anyone can get Claude to write a trading bot in an afternoon.
Surviving one is the skill nobody teaches.
I built delta-neutral funding bots with Claude and lost real money to bugs that never threw an error, they reported success in green text while the position bled.
Just published the field manual. 9 lessons, each with the exact pattern to steal:
✅ reduce_only or the bot flips you long
✅ order_status lies, verify the position with retries
✅ your PnL log counts deposits as profit
✅ isolate the wallet like you expect the key to leak
The failure mode is silent. 👇
4
16
772
Jun 2
The truth is out there. X Files...
Last week I wrote prediction markets passed US sportsbooks on volume.
1. Here's the shadow of that growth, live today.
Polymarket market is MicroStrategy sells any Bitcoin by May 31?
Strategy did sell.
32 $BTC, May 26-31, dated "as of May 31, 4:00pm ET", inside the window.
But the filing disclosing it landed June 1. One day after the deadline.
2. $155m in volume on this one contract on Polymarket.
Now in final review, and the UMA oracle vote has split into three camps:
✅ P2, resolve YES, the sale happened in the window, the filing says so
✅ P1, resolve NO, nothing confirmed it before the deadline, disclosure is what counts
✅ P4, too early, the rules were too vague to resolve at all
3. Three camps. Same facts. No agreement on what happened even means.
Polymarket even floated clearing the orderbook entirely, voiding every position, if no clean resolution lands.
This is the actual hard part of eating finance. Not liquidity. Not settlement. TRUTH.
Where's the strength, bro? In the Truth? But which truth exactly?
When does an event happen, when it occurs, or when the world can verify it?
4. TradFi spent a century building courts and disclosure law to answer that.
Prediction markets are speedrunning it with a token vote.
Resolution risk scales with every dollar of new volume. The bigger they get, the more $155m what is truth fights we'll see.
May 30
Prediction markets just went vertical!
$5b per month in Sep 2025 → $24b per month in April 2026 🚀
5x in 8 months.
Total US legal sportsbook volume averaged $14b per month in 2025.
Prediction markets just got bigger than US sports betting.
And the mix flipped:
✅ Late 2024 (election peak): 65% of Polymarket volume was political
✅ Now: sports culture crypto dominate. Politics is a minority category. 😱
This is the same arc as perp DEX niche → infra → category.
Once liquidity, settlement, and resolution all work at scale, the surface area expands faster than anyone expects.
Polymarket on this curve has serious IPO optionality. But the deeper play might be the picks-and-shovels: data providers, oracle networks, regulatory rails.
6
21
1,000
Jun 2
Who is Brevan Howard for Variational?
The most important name in Variational's $50m round isn't Dragonfly. It's Brevan Howard.
Most coverage buried it under and other investors.
But a $35b macro hedge fund doesn't write a perp DEX Series A check for the 10x. Crypto VCs do that. Macro funds invest to get positioned.
This summer Variational routes 100 TradFi markets (equities, forex, rates) directly from institutional dealers via RFQ.
Quotes come from registered, whitelisted LPs, not an anonymous pool.
Someone has to BE those dealers.
Brevan Howard trades FX, rates, and commodities at institutional scale.
They invested to become a liquidity provider on the rails they now partly own.
✅ own the equity
✅ quote the Phase 2 flow
✅ capture the spread
Three revenue lines from one check.
The founders' moat was always relationships.
This is the first dealer relationship, already signed, already funded into the cap table. $VAR
May 23
People kept asking how Variational actually gets institutional VIP rates on Binance, Bybit, HL.
1. Schuermann and Yu founded Qu Capital in 2017.
Barry Silbert's DCG acquired it in 2019.
2. The two then ran engineering and quant at Genesis Trading, one of the largest crypto desks ever.
Hundreds of billions in volume passed through their team.
They lived through 3AC FTX Genesis collapse from the inside.
3. Left in 2021. Spent the next 2 years running Variational as a market-making firm before Omni shipped a single line of code.
4. By the time Omni went live
- they already had taker volume history on every major CEX
- counterparty trust from every major desk
- a working risk hedging stack from years of prop trading.
The moat isn't the RFQ architecture.
It's everything the founders built before Omni existed.
You cannot fork this! You can only buy it!
▪️ structural moat
▪️ priced in only after TGE @variational_io
5
24
1,852
Jun 2
if you are still not trading on Variational, then it's mistake OMNIREN
omni.variational.io/?ref=OMN…
6
177
Jun 1
My Proof of Variational! Rank 202.
Not bad, I write about what I like.
Thanks @muarmemuar
May 31
Built an activity checker that tracks how much someone has helped @variational_io grow on X.
By views, I'm ranked #5
By likes, I'm ranked #2
By number of posts, I'm ranked #492
Hopefully @variational_lvs rewards the top 10 users in each category 😁
So far, 28,739 users who mentioned Variational have been indexed.
You can check your stats in the link below 👇
2
1
25
1,823