与「arbitrage」相关的搜索结果

Korean premium for crypto is going negative, and this is very unusual, as Korea usually has a premium. By the way, Korea tends to have either a premium or discount because arbitrage is not easy due to capital controls. Here are some thoughts on why the discount is happening. 1. The speculative market is declining. The Korean market is known for speculation, which at one point pushed the premium to 20%+. However, as the crypto market has been struggling, especially altcoins, the market's interest has been declining. 2. The Korean stock market is doing much better, which is driving liquidity away from crypto. The Korean stock market has almost doubled compared to last year, and several tech companies, including Samsung and SK Hynix, have been leading the growth. This means liquidity that was previously in crypto has been moving into the Korean stock market. 3. The upcoming Korean crypto tax, likely starting next year, is also changing market appetite. Crypto tax in Korea has been delayed multiple times because Korea does not have good enough infrastructure for collecting crypto tax, and it is also a very unpopular political topic. However, the current government, at least for now, has confirmed that it will tax crypto starting next year.

This is insane. A college dropout from Massachusetts Institute of Technology reportedly made $297,000 using a simple Bitcoin arbitrage strategy. He noticed BTC was selling for thousands more on overseas exchanges due to local demand. So he would: • Buy BTC cheap on U.S. exchanges • Transfer it overseas • Sell it higher instantly • Repeat daily Once he automated the process with bots, the profits scaled fast. No meme coins. No gambling. Just spotting inefficiencies before everyone else. The biggest money is made where attention is low. I have a complete arbitration course. Action items: Learn AI + automation Study market inefficiencies Build systems, not hype Test ideas daily Follow me @mikezillionaire Comment “AI” and like + share to get the course👇

🚨 Claude Code made me 6 trading bots in 15 mins In the US alone, emotional retail traders lost more than $1.8 billion on liquidations While billions of amateur traders were staring at charts, overtrading, and getting wrecked on fees, a quiet group of algorithmic traders treated prediction markets like a hyper-liquid data engine They didn't guess outcomes -> they knew the structural price gaps in advance Here is how they did it, and why manual trading is completely dead: It's all about removing emotion and deploying cross-market statistical arbitrage Linear Spread Cointegration Formula: S_t = P_P,t - β * P_K,t - μ Ornstein-Uhlenbeck Continuous Dynamics Formula: dS_t = θ(μ - S_t)dt + σ dW_t Euler-Maruyama Discretization (MLE Calibration) Formula: S_t_i = S_t_i-1 * e^(-θΔt) + μ(1 - e^(-θΔt)) + ε_t Level 1 Order Book Imbalance (OBI) Formula: I_t = (V_b(t) - V_a(t)) / (V_b(t) + V_a(t)) Volume-Weighted Micro-Price Prediction Formula: P_micro(t) = P_mid(t) + I_t * (Δspread / 2) Cross-Venue Predictive Signal Optimization Formula: ΔP_Kalshi(t + δ) = f(I_Polymarket(t), P_micro,Polymarket(t) - P_micro,Kalshi(t)) In the era of advanced AI, the winner is not the one who guesses the score, but the one who lets automated systems execute with absolute patience and discipline. AI does the hard parts now -> you don't even need a CS degree to build this The full behind-the-scenes live system build is now available to the public 📝

Hyperliquid is being treated like the cleanest up only narrative in crypto and I want to add some nuance here. To be clear, this is not Luna. Luna was a closed loop algorithmic peg with no real revenue. Hyperliquid has real fees from real users, a genuinely best in class product, a dominant share of onchain perps volume, and a team that has shipped relentlessly. The bull case is real. I'm not dismissing it. I just don't think the risks are getting the airtime they deserve. I called Luna a ponzi here in 2021 around $50 and got absolutely hammered on the tl for daring to suggest anything negative about it as it ran to $120 before it vaporized to $0 within days. Very similar to yesterday when I simply said HL does not have remarkable tokenomics and got piled on for it. Being early on structural risk often looks wrong for a long time. And Luna was not just a retail rug. It roped in 3AC, Galaxy, Delphi, Hashed. Sophisticated money held the bag right alongside everyone else. Here is what I see with Hyperliquid. 97% of fees buy back HYPE. That sounds incredible, and in an active perps bull market it absolutely is. But fees come from perp volume, volume comes from people chasing the token, and the token is held up by buybacks funded by that same volume. Every leg moves together. It's a functional flywheel. And in the other direction every leg turns at once. However, nobody can tell you how much of the volume is organic either. If buybacks pump the token and the chart pulls in size and size funds buybacks, you cannot cleanly separate real activity from reflexive activity onchain. It doesn't mean the volume is fake. It just means you cannot prove how much of it isn't. Then the supply side. Only 25% circulates. Team and foundation together hold roughly 30% (23% team plus the foundation allocation which is essentially team with extra steps). Buybacks absorb about 90 million of unlocks a month. Actual pressure is closer to 400 million plus. Revenue keeps growing at a real clip, which is the whole bull case, but it has to roughly 4x just to keep price flat through vesting. Then the part nobody wants to touch. 31 validators, foundation controls the supermajority of stake, closed source binary, an assistance fund holding billions that we are simply told has no private key, on a chain the team built and runs. A lot of the business is regulatory arbitrage. Offshore venue, no KYC, users that shouldn't be using it are all over it. The founder is in DC right now precisely because everyone knows this. SBF was in DC lobbying for the DCCPA right up until FTX collapsed. Do Kwon was meeting Korean regulators before Terra blew up. Doesn't mean Jeff is anything like them. And to be clear, none of this means price stops going up. This is the part that matters for traders. Reflexive setups run for a long time, sometimes years. Luna ran from $5 to $120 while plenty of smart people screamed about the structure. HYPE can absolutely do the same. The flywheel is real while it's spinning, and standing in front of it is a great way to underperform. Just know what you're actually holding. Trade the tape, respect the trend, but don't fall in love and confuse a beautiful reflexive setup with a riskless cash machine. We have seen structures like this before and it tends to end the same way for usually the same reasons.

Watch @saylor 's latest thoughts on $BTC, $STRC, and $MSTR with @TheBonnieChang and @davidlin_TV at Consensus 2026. A Master Class. 0:00 - Strategy’s Bitcoin sale controversy 0:15 - Why Strategy may sell Bitcoin 2:42 - “Never sell your Bitcoin” explained 4:19 - How Strategy buys more Bitcoin than it sells 6:12 - Michael Saylor’s Bitcoin accumulation philosophy 7:45 - Using Bitcoin liquidity and market arbitrage 9:49 - Responding to Ponzi scheme criticism 12:07 - STRC trading patterns and Bitcoin buying 13:28 - What really drives Bitcoin’s price 16:34 - Bitcoin, macro risks, and Fed policy 18:21 - Bitcoin as digital capital and digital credit 20:29 - Strategy’s dominance in preferred stock issuance 21:45 - AI, digital credit, and Bitcoin’s future 23:13 - Saylor’s childhood inspiration and MIT story

个人感觉币圈kol发的这些预测市场脚本。很容易成为被套利对象 我也还在学习 toxic flow检测，latency arbitrage，跨市场对冲，anti-prediction，动态inventory平衡等知识。欢迎老师指导

Cursor is raising at a $50 billion valuation on the claim that its “in-house models generate more code than almost any other LLMs in the world.” Less than 24 hours after launching Composer 2, a developer found the model ID in the API response: kimi-k2p5-rl-0317-s515-fast. That’s Moonshot AI’s Kimi K2.5 with reinforcement learning appended. A developer named Fynn was testing Cursor’s OpenAI-compatible base URL when the identifier leaked through the response headers. Moonshot’s head of pretraining, Yulun Du, confirmed on X that the tokenizer is identical to Kimi’s and questioned Cursor’s license compliance. Two other Moonshot employees posted confirmations. All three posts have since been deleted. This is the second time. When Cursor launched Composer 1 in October 2025, users across multiple countries reported the model spontaneously switching its inner monologue to Chinese mid-session. Kenneth Auchenberg, a partner at Alley Corp, posted a screenshot calling it a smoking gun. KR-Asia and 36Kr confirmed both Cursor and Windsurf were running fine-tuned Chinese open-weight models underneath. Cursor never disclosed what Composer 1 was built on. They shipped Composer 1.5 in February and moved on. The pattern: take a Chinese open-weight model, run RL on coding tasks, ship it as a proprietary breakthrough, publish a cost-performance chart comparing yourself against Opus 4.6 and GPT-5.4 without disclosing that your base model was free, then raise another round. That chart from the Composer 2 announcement deserves its own paragraph. Cursor plotted Composer 2 against frontier models on a price-vs-quality axis to argue they’d hit a superior tradeoff. What the chart doesn’t show is that Anthropic and OpenAI trained their models from scratch. Cursor took an open-weight model that Moonshot spent hundreds of millions developing, ran RL on top, and presented the output as evidence of in-house research. That’s margin arbitrage on someone else’s R&D dressed up as a benchmark slide. The license makes this more than an attribution oversight. Kimi K2.5 ships under a Modified MIT License with one clause designed for exactly this scenario: if your product exceeds $20 million in monthly revenue, you must prominently display “Kimi K2.5” on the user interface. Cursor’s ARR crossed $2 billion in February. That’s roughly $167 million per month, 8x the threshold. The clause covers derivative works explicitly. Cursor is valued at $29.3 billion and raising at $50 billion. Moonshot’s last reported valuation was $4.3 billion. The company worth 12x more took the smaller company’s model and shipped it as proprietary technology to justify a valuation built on the frontier lab narrative. Three Composer releases in five months. Composer 1 caught speaking Chinese. Composer 2 caught with a Kimi model ID in the API. A P0 incident this year. And a benchmark chart that compares an RL fine-tune against models requiring billions in training compute without disclosing the base was free. The question for investors in the $50 billion round: what exactly are you buying? A VS Code fork with strong distribution, or a frontier research lab? The model ID in the API answers that. If Moonshot doesn’t enforce this license against a company generating $2 billion annually from a derivative of their model, the attribution clause becomes decoration for every future open-weight release. Every AI lab watching this is running the same math: why open-source your model if companies with better distribution can strip attribution, call it proprietary, and raise at 12x your valuation? kimi-k2p5-rl-0317-s515-fast is the most expensive model ID leak in the history of AI licensing.

@steipete wait, cursor for free tokens? as in routing through cursor's API to avoid paying directly? that's either genius arbitrage or the kind of thing that gets patched in a week. how stable has it been?

“In 1776 the American colonists rebelled against what they saw as the arbitrary and tyrannical British monarchy,” writes Rosa Brooks in a guest essay. Today, “it’s hard not to imagine Mad King George gazing out at Donald Trump’s America—and laughing”