Joy (turn back the clock to 2021 & defi summer) retweeted
Someone is trying to blackmail the U.S. Government
These attacks aren’t random
Joy (turn back the clock to 2021 & defi summer) retweeted
Emerging markets tech hedge fund manager Sinan Xin (@SinanXin) on why AI won't be the durable edge:
"AI, like alternative data, is just another tool for the fundamental investor."
"If you're first to that kind of tool, it is helpful and you do become differentiated."
"But if it's something that you've built that's obviously based on public information, it will diminish over time."
"If you were reading, you're going to the library and you were borrowing the 10-K, reading it, you had an edge."
"The tools will always evolve."
"What is actually harder to build but is more durable is behavior."
Joy (turn back the clock to 2021 & defi summer) retweeted
Databento CEO Christina Qi (@christinaqi), who co-founded HFT fund Domeyard out of her MIT dorm room, on why the easy alpha is gone:
"You can literally just use AI and just vibe code your way into doing a quant strategy."
"10 plus years ago, getting access to high quality order book data was really, really difficult to do unless you were like a large institution."
"Today, you can literally just go on your laptop in your bedroom and access data."
"The barrier to entry has lowered, and as a result of that, the alpha, the edge, of course it's going to be harder. The lower hanging fruit has all been kind of picked out."
"All the strategies that you think are really novel and bleeding edge has already been picked out and is already dead. Sorry. That's just kind of the reality of it."
Joy (turn back the clock to 2021 & defi summer) retweeted
99% of options traders are gambling because they never learned the fundamentals.
This 1-hour Yale lecture changes everything.
In just 60 minutes, you’ll learn more about options trading than most overpriced trading courses ever teach.
No hype. No fake gurus. Just real knowledge.
Save this and watch it without distractions. 📌
Joy (turn back the clock to 2021 & defi summer) retweeted
做了个追踪 Hyperliquid 头部交易员的网站
▌已追踪 118 个钱包
▌其中 33 人已关联 X 账号
功能如下👇
总览:组合合计 + 净敞口 + 多空分歧 + 持仓变化
交易员:支持按账户价值 / 持仓筛选,展示主要持仓,点击可查看仓位明细
关注标的:收录约 60 个币种,左栏选币,右栏即时呈现谁在持有、多空分布、杠杆与名义金额
观点:聚合关联账号的近帖
🔗 传送门:hyperliquid-515d81.surf.comp…
只是初版,后续会不断迭代。
made by @SurfAIHQ
Joy (turn back the clock to 2021 & defi summer) retweeted
As someone who builds quant trading systems, I wrote the paper I wish existed when I started on Hyperliquid.
6 pages on how to build a market making bot for a fully onchain order book. Bookmark this before you write a single line of code.
Joy (turn back the clock to 2021 & defi summer) retweeted
Ex-Tudor PM Tom Costello (@tcoste110) on what he looked for as head of equity trading at a $1B quant fund:
"At one point I was head of equity trading for about a billion dollar quant fund. I had guys coming to me with their strategies multiple times per day."
"I would interview about 10 of them a week, for a couple of years."
"People would come to me with that overfitted backtest. Well, this is just a backtest. Come back to me when you have real results."
"If they did come back with real results, very often the results would include a component of beta. I would say, well, I don't need you to trade beta. I can get my own beta. You've got to give me what's different from beta."
"What a hedge fund was always looking for was an uncorrelated return. If I have the S&P 500 in one hand and another asset that goes up as well but is completely uncorrelated, I can combine them and lever it and get an even higher return than either of them."
Retail investors vs. hedge fund managers: who wins?
Ex-Tudor quant PM Tom Costello (20%/yr, 1.4% max drawdown) DEBATES Market Wizard Chris Camillo ($20K → $80M, tells you to expect 70% drawdowns).
Chris Camillo (@ChrisCamillo) turned $20K into ~$80M over 18 years, audited by Jack Schwager for Unknown Market Wizards at 77% annualized. Founded TickerTags, sold to Jefferies. Co-founder of Dumb Money.
Tom Costello (@tcoste110) ran money at Tudor, Moore Capital, and Caxton. Started as a quant on JPM's exotic swaps desk. Now CIO at Bedrock Digital Assets.
We cover:
- Why a 50% drawdown is "a career-ending event" for a PM and "completely rational" for a retail investor with a 20-year horizon
- Chris's case that everyone should have a levered, concentrated risk bucket, and expect a 70% drawdown in it
- Tom's pushback: "At least half of the aspiring engineering students who wanna be day traders are gonna go broke"
- Why hedge funds don't beat the S&P, and why that's a category error
- 3x leveraged S&P ETFs held for decades: ~1.85x market returns, and the futures roll cost buried inside
- The e.l.f. trade: one YouTube video, a full day sitting in a Walgreens, and every cosmetics analyst on Wall Street asleep
- Chris spent 5 years selling social data to the biggest funds. They couldn't institutionalize it because "the words keep changing"
- Vita Coco: a viral TikTok sound → the biggest quarter in company history → +34% on earnings
- Tom's bet: LLMs hand social arb to the quants. Chris: "there's still plenty of time"
- "I know very few day traders who ever made anything like real money"
- Why Chris gives the strategy away after an 18-year run
Highlights:
00:00 Intro
00:50 Who's better positioned to take risk: retail or hedge funds?
05:58 A message from ONYX
06:30 Why a 50% drawdown is rational for concentrated retail capital
08:57 Beta vs alpha: what two-and-twenty actually pays for
18:56 Should retail embrace 70% drawdowns? Bucketing risk capital
23:59 Can ordinary investors really compound 2–3x market returns?
30:08 Fair value, mistakes, and why 99% of day traders lose
35:32 Are 3x leveraged S&P ETFs a long-term buy?
43:31 Too big to fail: AI concentration and the case for leverage
46:06 Situational Awareness, crypto correlation, and overfit backtests
53:44 Social arb explained: Beacon Roofing, Hunger Games, Stranger Things
01:01:07 Is social arb crowded? Tiger Cubs, Ticker Tags, e.l.f.
01:07:42 Will Citadel and LLMs arb away TikTok alpha?
01:17:09 Incentives: why "hedge funds don't beat the S&P" is a category error
01:19:35 Why Chris shares social arb: $20K to $80M and the wealth gap
01:25:45 Closing: retail freedom vs institutional resources
Joy (turn back the clock to 2021 & defi summer) retweeted
My friend is an Anthropic engineer making $1.2M/year.
I asked him one question: “How did you get so good at prompting?”
He sent me a 30-minute video.
I watched it last night, and halfway through, I realized something:
I’ve been using Claude completely wrong for the past 2 years.
This wasn’t another “10 prompts you need to know” video.
It completely changed how I think about prompting - the way Anthropic’s best users structure problems, give context, set constraints, and get AI to do serious work is on another level.
If you use Claude for coding, research, writing, business, or anything serious, watch this before you write your next prompt.
Then read the article below.
Because once you see how the best AI users actually work, you’ll never prompt the same way again.
Joy (turn back the clock to 2021 & defi summer) retweeted
4月,Jane Street裁掉了这名量化交易员。
他在那里带了11年pod,年薪41.2万美元。只因一个季度没达标,一周内打包走人。
昨晚,他用刚开放权重的Kimi K3,在Polymarket上跑出91.7万美元。
对冲基金每年砸数亿美元养私有模型,他只租了一台GPU服务器,丢进去一个agent循环。
Kimi K3是2.8万亿参数的MoE模型,百万token上下文。Kimi Delta Attention据称把百万token推理提速约6.3倍,循环从一分钟压到几秒。
他从春天开始搭这个bot,缺的正是能一次性塞进全部交易日志、一整年5分钟BTC K线、全部历史回测,且不用裁剪的模型。
K3上线后,它自己交易、自己看结果、自己改规则、接着跑。第一个晚上就完成了917个循环。
链上页面:
polymarket.com/@0x75973c667c…
讽刺的是,他在Jane Street那11年学到的最值钱的东西,不是更贵的模型,而是怎么让系统跑得更快、更直接。
基金那边新模型还要过风控委员会,他这边公告一出,已经开始迭代。
他不是在智商上碾压老东家,是在速度上把对方甩开了。
🤝 Paid partnership
this guy is the only substack im subscribed to. he shares buys/sells and really good deep dives on it. he was early on all the popular names and gains speak for themselves. you won't regret following him.
To celebrate Dick Capital’s one-year anniversary on Substack, we’re running a special offer for the next week. ♥️
In one year, we’ve become one of the top publications in Finance, uncovered some of the market’s biggest themes before they went mainstream, and watched subscribers completely change their financial lives.
Thank you to everyone who has been part of the journey.
Until October 4, you can subscribe for less than a dollar a day and join us.
Full post + special offer available through the link in our bio.
Jane Street Head of AI just released 1-hour course on how they uses LLMs and AI agents to make $39.6 billion/year:
• 3:12 - how Jane Street Quants adopted AI agents
• 17:34 - 230k tokens later, agents can forget the rules!
• 29:33 - Jane Street’s test loop reveals what agent code really does for trading
• 55:00 - Jane Street’s setup: one agent writes, several others hunt for flaws
One agent can find you a trade. But who checks the agent?
Jane Street shows how it’s work: tests → feedback loops → agents checking agents - inside one of the world’s most profitable trading firms
copy Jane Street’s agent workflow, adapt it to your own trading system - then build AI setup from one prompt ↓
Joy (turn back the clock to 2021 & defi summer) retweeted
SpaceXAI engineer, Lauren Tan:
"99% of people using GrokBot just for 1% of its real power. They run 1 agent without "loop" & "graph"
I'm running a team of 20+ GrokBot agents, fully autonomous. I have a Chief of Staff agent, a PM agent and 20+ workers - that's the new stack of engineer"
In a 1-hour session, a SpaceXAI engineer showed how to build a team of effective AI agents from scratch
this is worth more than a $500 agentic engineering course
watch this workshop today, then read how to build a fleet of GrokBot agents in the article below
Joy (turn back the clock to 2021 & defi summer) retweeted
THIS IS F**KING DANGEROUS
I GAVE CLAUDE ACCESS TO OPTIONS FLOW, ORDER BOOK DATA, AND GAMMA WALLS — AND IT STARTED CALLING LEVELS BETTER THAN I DO
not a chart with two moving averages. not a “buy the dip” alert. this is a full institutional desk compressed into one AI session
here’s what’s on the screen right now:
→ GEXRadar pulling Gamma Exposure across 44 strikes in real time
→ Delta Exposure flipping between bars and candles showing where dealer hedging shifts
→ Major Walls: call wall at $700 with +276M, put wall at $691 with -97M
→ Hedging Pressure gauge sitting at 48/100 — neutral, dealer flow balanced
→ Net Drift tracking $19.93B in calls against $32.39B in puts — put dominated
→ IV Surface Analysis rotating in 3D showing where volatility is mispriced by strike and expiry
→ MotiveWave footprint printing every bid and ask at every price level, tick by tick
→ DOM ladder on MNQ futures with live bid/ask stacking — 29,103.75 ask, 29,103.50 bid
Claude reads all of it
not the chart — the structure underneath the chart
gamma exposure tells it where dealers are forced to buy or sell. delta exposure tells it which direction the hedging pressure is building. the wall map tells it where price stalls. the footprint tells it whether the move is real volume or just air
entry: 1 contract MNQ at 29,054.75
TP set: 29,178.25
P&L climbs → +$64 → +$146 → +$167
the trade wasn’t a guess. the call wall at $700 held, hedging pressure was neutral, gamma flip sat at $698, and the footprint confirmed buyers absorbing every offer at the 29,050 level
Claude saw the wall. saw the flow. saw the absorption. and said: this is the entry
tested this setup with $100 starting capital — ended the session at $3,153
that’s not leverage doing the work. that’s an AI reading options structure, order flow, and gamma positioning simultaneously — something no human can do across 44 strikes in real time
you’re still drawing trendlines
Claude is reading the market’s skeleton
save this — the full setup and session walkthrough is in the article below ↓
Joy (turn back the clock to 2021 & defi summer) retweeted
I still don’t understand why everyone is not doing this yet. Elon Musk reposted my Grok Bot GUIDE, and using this exact agent setup, I made $13,100 last week alone.
Eight Grok agents on the desk, $200 a month, running a floor that usually costs a crypto fund $500K a year in analyst salaries.
The morning call is at 4am. I am not in it.
1. SEARCH scrapes real-time alpha, dev GitHubs, and unindexed Telegram signals before CT finds them
2. RISK audits contract functions, mint rights, and LP locks, flags honeypots before entry
3. SNIPER places high-speed orders on chain the exact millisecond risk clearance passes
4. WHALE tracks smart money wallets and flags insider accumulation in real time
5. RUG monitors dev wallet activity 24/7 and dumps the entire position if LP is touched
6. EXIT trails stops dynamically, scaling out as liquidity builds
7. SHILL tracks social volume, momentum velocity, and key influencer calls
8. HEAD OF DESK never trades, routes data, checks handoffs, and brings me the one decision that needs a human
142 tokens scanned, 19 qualified setups, 6 executed trades while I was asleep. Net result: +$13,100 after fees and zero bad fills.
Every agent has its own virtual browser, terminal, and local memory in the cloud. The floor stays active with my laptop shut.
The setup is dumber than it looks:
Download Grok Bot and create your Head of Desk.
Give the remaining 7 agents job descriptions like you're briefing new hires.
Run the workflow once on your screen while they watch.
Hook up Telegram and wallet webhooks.
No VPS, no code, no waiting on developers.
A crypto trading floor used to mean 16-hour screen time, paid alpha groups, and constant fatigue. Mine took one evening to set up.
Save this before your next trade. Save GUIDE.
I gave Elon Musk's new Grok Bot an org chart instead of a to-do list, and in one week I stopped being a founder who does the work and became one who assigns it.
eight bots. one org chart. nobody sleeps but me. here's the whole design, steal it.
step 1 → 0:01 What we're covering
step 2 → 2:01 Installation & Setup
step 3 → 3:14 Building the First Bot
step 4 → 8:18 Teaching by Screen Recording
step 5 → 14:18 Putting it All Together
THE ROSTER:
Atlas, chief of staff. the only bot I talk to. I give it outcomes, never tasks. it decomposes them and delegates to the team in group chat, and it never does specialist work itself. it posts the plan every morning and what shipped every night, and it only comes to me when a decision is irreversible or spends money.
Scout, research. finds and qualifies my ICP. every day: 25 verified prospects, one line on why they need us right now, and a source. if it can't verify, it marks it unverified. it never guesses.
Quill, content. turns what the company learned this week into 5 posts and 1 long piece, in my voice, matched from the last 50 things I wrote. drafts only, it never publishes.
Pitch, outbound. writes a first touch and two follow-ups for everyone Scout marks ready. 60 words max, one specific observation about their business, one clear ask. queued in drafts, I approve in bulk.
Vault, inbox and ops. triages everything into needs-me, needs-a-bot, needs-nothing. it handles the last, routes the middle, and gives me five bullets on the first by 9am.
Ledger, analyst. one report a night: what moved, what didn't, and the single number I should care about tomorrow. no dashboards, no adjectives.
HOW THEY'RE WIRED
one group chat per outcome, not per person. Atlas sits in all of them. the bots hand off inside the chat, so I only read the handoff, I never manage it.
two rules that made this actually work:
1. every charter ends with a hard "never do this without asking" line. autonomy without a fence is just chaos on a schedule.
2. show once, don't describe. I ran the full workflow on my screen one time. that single demo taught them more than a page of instructions ever could.
WEEK ONE
214 verified prospects delivered. 89 personalized outreaches queued and approved. inbox at zero every morning. 11 content pieces ready.
THE POINT
most people are still treating Grok Bot like a smarter chat window. it isn't. it's the first time one person can own an org chart instead of a to-do list. my bottleneck was never how much I could do, it was how much I could hand off.
bookmark this.
retail has 10,000 indicators and still loses 80% of the time
quants have one number they pull every morning before touching anything else
it's called options skew - the gap between how much downside protection costs versus upside exposure on any liquid name
CBOE publishes it free, every 15 minutes, on every major ticker since 2004
when put vol runs 8+ points above call vol on SPY, institutional desks are paying premium to hedge hard - that's not noise, that's where money thinks risk actually is
backtested 2009 to 2024: when skew spikes 1.5 sigma above its 90-day baseline, SPY underperforms next 15 sessions 68% of the time
not predicting direction - reading where billion-dollar desks are quietly paying to protect themselves
size with kelly at 68% edge, enter short, exit in three weeks
retail spent years memorizing moving average crossovers
quants spent the same years reading a free number nobody told them existed
same market. completely different question
bookmark this
data was public the whole time, you just didn't know to look for it
Joy (turn back the clock to 2021 & defi summer) retweeted
this is f**king dangerous
Most people treat Claude like a search bar.
The ones with a second brain treat it as an intelligence layer on everything they know.
Here's the 15-minute setup:
→ Install Claude Code
→ Install Obsidian, create a vault
→ Open it in Claude Code, drop in the CLAUDE.md schema — that one file is the engine
→ Throw any source in, say "ingest this"
Done.
From then on, everything you know is queryable by the most capable AI on earth. Claude wires each source against your whole vault and surfaces connections you forgot you had.
Build it tonight.
full A–Z guide below.
Joy (turn back the clock to 2021 & defi summer) retweeted
$BTC Liquidity/Liquidation levels to watch:
Below, $74.8K & $75.5K got a few decent clusters.
Above, the entire $78K-$80K is pretty filled due to the slow grind down before. Another big area sits at $82K.
Joy (turn back the clock to 2021 & defi summer) retweeted
This is Leopold Aschenbrenner.
He turned $220M into $13.67B in 15 months. 62x.
All by trading AI stocks as ex-OpenAI employee.
Today his Q1 13F dropped - SEC filing of every position he holds.
Here's the book of the most profitable AI trader alive 🧵👇
ANTHROPIC JUST DROPPED 13 FREE CLAUDE CERTIFICATIONS AND ALMOST NOBODY IS TALKING ABOUT IT.
Not a YouTube playlist.
Not a third-party course.
Official certifications from the team that built Claude.
Free. Forever.
Here is the full list with links:
START HERE
01. Claude 101 — Learn Claude for everyday work
anthropic.skilljar.com/claud…
02. AI Fluency: Frameworks and Foundations
anthropic.skilljar.com/ai-fl…
03. Introduction to Agent Skills
anthropic.skilljar.com/intro…
FOR DEVELOPERS
04. Building with the Claude API
anthropic.skilljar.com/claud…
05. Claude Code in Action
anthropic.skilljar.com/claud…
06. Intro to Model Context Protocol
anthropic.skilljar.com/mcp
07. MCP Advanced Topics
anthropic.skilljar.com/mcp-a…
FOR EDUCATION AND NONPROFITS
08. AI Fluency for Students
09. AI Fluency for Educators
10. Teaching AI Fluency
11. AI Fluency for Nonprofits
FOR ENTERPRISE
12. Claude with Amazon Bedrock
13. Claude with Google Cloud Vertex AI
13 courses. 6 skill levels. 5 audiences. 100% free forever.
The engineers getting hired at $150,000 to $300,000 to work with Claude at the highest level are learning exactly this material.
Anthropic's team just made it available to everyone.
Pro tip: Start with Claude 101 then go straight to Claude Code in Action. That is the fastest path from beginner to builder.
Bookmark this before you pay for another AI course.
Follow @cyrilXBT for every Anthropic resource that compounds your skills the moment it drops.
Joy (turn back the clock to 2021 & defi summer) retweeted
RIP Bloomberg Terminal.
Someone just open-sourced a desktop app that does most of what a $24,000-a-year Bloomberg seat does, and it runs on Windows, macOS, and Linux.
Fincept Terminal ships with:
→ Full CFA Level 1, 2, 3 analytics implemented in Python
→ 20+ investor AI agents (Buffett, Dalio, Graham, Soros, Lynch, Klarman)
→ Hedge fund presets: Bridgewater All-Weather, Citadel multi-strat, Renaissance stat models
→ 100+ data connectors including Polygon, Kraken, DBnomics (100M+ series), World Bank, IMF, OECD
→ Visual node editor to wire data sources to Python agents to MCP tools
→ 3D globe with live AIS ship tracking, aircraft paths, and satellite orbits
→ Local LLM support so your positions never leave your machine
2.6K stars. AGPL-3.0. 100% Opensource.
github.com/Fincept-Corporati…















































