I don't predict price. I watch the people who move it. Fear, greed, attention, belonging, narrative. Same crowd. Different incentives. When the thesis breaks, I get out. I tell others to do the same. I still miss it sometimes. That’s the lens.

On the fastest phones, some short Laya decisions come back in around 0.2 seconds. That’s fast enough that I’d probably read the answer first and the probability second. Laya runs the same System One model across nine live demos on Android phones in the Acurast network. It returns probabilities, and a short decision usually takes around 0.2–1 second. The uncertainty is visible. The speed just makes it easy to ignore. Acurast’s mail sorter got 144 out of 160 letters right in one test. They also openly point out that tricky inputs can fool the model and describe Laya as a fast first filter, not the final word. I keep thinking about that gap. A model can effectively say “I’m 72% sure” and still answer before I’ve decided whether 72% is good enough. The system never had to claim certainty. My brain can add that part on its own. hub.acurast.com/rebellion?re… @Acurast

Towers 刚刚爬到后面，我又开始跟自己说“最后一层”。越往上，按下一格之前停顿得越久，刚开始那股一路往前冲的劲儿也没了。 这回真停住了，几盘下来在 @Roobet 净赚 240U。以前我说最后一层，自己都不信；今晚说到做到一次，居然比爬上去还爽 😂 roobet.com/?ref=all3in4out

I keep coming back to one thing @lampropeltis_ said after using FX100. The 15-minute protection made him more relaxed around entry volatility — sometimes even a bit more aggressive. If the thesis still held, he was more willing to give the trade extra time. That’s the part I care about. The feature changes the rules for 15 minutes, but it also changes what you’re willing to sit through. He described minute 16 as “relief first, then a decision.” Once the timer is gone, the trade is just the trade again. My self-check would be simple, If the 15-minute window disappeared tomorrow, would I still take the same position at the same size? @FX100Perp @protocol_fx

GM GN CT ☕️ One thing I keep coming back to in @Tangent_fi new sUSG market on Pendle is the points table. PT gets 15 Tangent Liquidity Points per PT. YT gets 45 per YT. LP gets 45 per dollar. It looks like a reward table, but it also becomes a decision screen. Someone who came in looking for fixed yield can very quickly start comparing 15 vs 45 before comparing what each position actually does. PT, YT and LP aren’t three versions of the same trade. The exposure is different. The number sitting next to them changes how attractive each path feels before the risk is even processed. That’s what I’d watch here. If the points disappeared tomorrow, would you still choose the same leg?

GM GM CT ☕️ The scary AI story this week is not extinction. It is cancel. Last week the market priced a simpler idea, an agent that never gets tired of comparing prices, booking the cheaper flight, and killing the plan you forgot you had. Planet Fitness dropped 14%. Charter dropped 12%. Allstate dropped 8.9%. Those are not chip stocks.Those are businesses that lived on one human habit, canceling is work, so most people don’t. The old click was Stay Subscribed. Not because the price was fair. Because the hold music was long. The new click is Confirm Cancel. The form is already filled. You just tap it. That is why this one travels. Only 75 to 100 million people on earth actually pay for AI.Everyone else is still on the free tier, watching the paid button. Apple’s $10 a month still feels like nothing.$20 a month suddenly feels like a decision.Same product. Different label. Same thumb. Then the other side showed up. Muse for Mac pulled 187,000 lines out of a user’s Messages database. Amazon shut the shopping door.Nvidia shipped a kill switch for agents the same day it authorized another $150 billion in buybacks. So the week has two buttons. One says: let it cancel the gym. One says: wait, what else did it just open. Both are real. Neither needs a forecast. The subscription did not get more expensive.The friction got cheaper.And friction was the product.

GM GM CT ☕️ A 15% net APR headline changes the question in your head. You stop asking, Is this position good? and start asking, How far can I push the LTV? In @Tangent_fi BTC FIR example, the borrow rate stays at 6.5%, while projected net APR rises as LTV moves from ~50% toward 78%, reaching ~15% on the sUSG/reUSD route. Same fixed funding cost. Much less room if collateral moves against you. That’s why the slider matters more than the headline. Up to 15% doesn’t just describe the trade—it quietly pulls your hand to the right. Fixed rate ≠ fixed risk.

今天的 Towers 又让我说了三次“最后一层” 前两次都没停，第三次才终于舍得收手 @Roobet 你是真的懂怎么让我跟自己吵起来 😭 roobet.com/?ref=all3in4out

GM GN CT ☕️ The closing bell still rings at 4.The button does not.This week the stock click split into three products. First: own a token of the stock after midnight. On Sept 17 the SEC gave qualifying venues a five-year pass to trade tokenized U.S. equities on public chains. Binance already listed the wrappers. Monthly tokenized-stock volume has been running in the billions. August printed near $5B. The tape did not wait for the culture to catch up. Second: lever the stock and never let it expire. Coinbase filed with the CFTC for single-stock perpetual futures — first set aimed at Apple, Microsoft, Tesla, Nvidia, then 50 to 60 names. 24/5. No share certificate. No dividend. Funding payments instead of a close. Offshore already had the toy. This filing is the attempt to bring the casino clock onshore. Third: bet on the outcome and get sued for the same motion. Yesterday New York sued Polymarket US as an unlicensed gambling business. Fine of 3x proceeds. Shut the door in the state. The platform still cleared about $109M in 24 hours. The Fed-in-October contract is still a live button. Same finger. Three labels. Buy. Perp. Yes. One is ownership. One is exposure that does not sleep. One is a probability the state now calls a wager. That is why this week travels.Not because Bitcoin sat on $84k while $15B–$19B of options expired today.Not because ETFs still took in ~$191M yesterday while Fear & Greed printed 71. Those are weather.The weather changed the interface. For a hundred years the hardest part of a stock was waiting for the open.This week the product managers removed the wait, then argued in court about what to name the wait they removed. If you felt the itch at 1 a.m. and opened an app anyway, that was the design working.The design is no longer place an order.The design is there is always an order to place.

礼盒到的时候我还在想，是不是又是月饼。 打开才发现，这回真有点惊喜呀。 外盒那个月亮会亮，很中国风，很有意境。 巧克力、挂件挺精致的，让我意外的是那只空灵鼓，大小两根槌，旁边还有2张谱，照着敲了两下，居然有那么一点意思。平时电脑看累了，敲两下，有种吃斋念佛的禅意。 很感谢麦通 @msxcom @MSX_CN ，每次周边都很惊喜，真的很用心在做，他们是做链上美股的，想交易美股的话，可以去看一看，用心做事的人靠谱，放心。 明天就中秋了，祝朋友们中秋快乐，圆圆满满。

以前看别人玩Towers，我一直觉得不就是一层一层选吗？ 今天自己在 @Roobet 开了几局，才发现真的走到后面，手会越来越慢。 前几层我按得特别快，感觉自己状态好得不行。层数上来以后，每一次选择都开始像最后一次，明明已经可以停了，脑子里还是会冒出一句：都走到这里了，再上一层吧。 然后你就会知道，这游戏最难的不是选左边还是右边，是说完最后一层以后真的停下来 😂 看别人爬是一回事，轮到自己决定又是另一回事。 来试试你能爬到哪里： roobet.com/?ref=all3in4out

Yesterday @protocol_fx said fxUSD had crossed $85M in circulation, adding that nothing is being paid to make it happen. That line is interesting because people can easily read it as this is all organic demand. But there are different layers here. fxMINT still shows a Merkl campaign offering 10% APY in FXN for minting fxUSD with ETH. Around the same time, @protocol_fx also highlighted ~10% APR on the debt side of fxMINT when LTV is above 70%. Those are different mechanisms, and neither automatically explains the entire $85M supply. Current official metrics show roughly $85.9M fxUSD supply, $130.6M in reserve assets, and $228.4M TVL. The growth is real. What I’d watch next is how much of that demand stays after separating product usage from each incentive layer. That distinction matters more than the headline.

GM GN CT ☕️ What caught my attention about @Tangent_fi upcoming FIR markets isn’t just the fixed rate part. It’s the possibility that people may quietly translate fixed borrowing cost into more predictable position. Those are not the same thing. FIR isn’t set forever and untouchable. Tangent also notes the DAO can adjust the rate, and if USG depegs in a significant and persistent way, FIR can be converted into HEC or LEC. So the product is designed to reduce one source of uncertainty, not remove uncertainty as a whole. That’s probably where most users will slip. Once one number stops moving, the brain tends to relax a little too early. But collateral risk, liquidation risk, and market structure don’t disappear just because funding cost looks cleaner on paper. So to me, FIR is less about safe borrowing and more about what kind of uncertainty you’re choosing to keep.

GM GN CT ☕️ The people who built it just asked the world to wait. Then they priced the wait. Last weekend Anthropic’s CEO called for a slowdown at the frontier. Altman said pacing is not stopping. Musk agreed. Chips sold off. Cyber names ripped. California signed a kill-switch order. Same week, the same company crossed a $100 billion revenue run-rate and slid its IPO from October to November — so the roadshow can carry a cleaner quarter. Investors are still talking $2 trillion. That is the tell. Slow down is the essay. November is the button. One sentence is for the public. One date is for the order book. Friday, while that argument was still open, Robinhood jumped 9%. Coinbase 12%. Retail did not wait for the safety paper to finish. The Fed had just hiked. The 10-year was back at 5%. Jensen said Nvidia will sell twice as many chips next year. So two clocks are running. On stage pause the frontier.On the phone don’t miss the listing, the chip, the bounce. Nobody has to be wrong for both to be true. A lab can fear what it is building and still need the window. A trader can believe the warning and still hate being last. The click is not is AI dangerous.The click is if they are scared enough to write it down, am I early or late. That second question is the one that moves money. #CT

GM GN CT ☕️ One $USG , three very different risk books: lever an LP, trade with borrowed liquidity, or skip the debt and park in sUSG. sUSG is the quiet one—a Yearn v3 vault that auto-compounds USG, with 25% of protocol revenue currently feeding its yield. Borrowing is a different game: HEC and LEC charge differently, and once your health factor breaks, liquidation is real. That’s the product map. Not the APY card. Before chasing yield, know which risk book you’re actually in. @Tangent_fi #Tangent

GM GN CT ☕️ fxUSD supply dropped to $79.72M, but the reserve didn’t move in one clean line. Official stats now show $111.42M in reserve assets, with 6,666.30 stETH and 1,264.34 WBTC. Compared with the prior snapshot, stETH is slightly higher while WBTC is lower. So I’m not reading the supply drop as capital just left. The more useful question is how the reserve mix is changing underneath it. Funding is currently Off at 0%, fees remain 0.3%, and leverage still tops out at 7x. Those are live states. The balance-sheet shift is the part I’d keep watching. Supply tells you how big fxUSD is. Reserve composition tells you what is carrying that size. @protocol_fx #fxUSD

GM GN. CT ☕️ #USG can’t be redeemed for its collateral. So its peg defense lives in liquidity. @Tangent_fi crvUSD-forked Peg Keepers currently work through two Curve pools: USG/USDC and USG/frxUSD. When USG is scarce, they add it; when USG is in excess, they pull it out. Anyone can call the function once conditions are met. 50% of keeper profits go to Swiss Stake AG. No redemption desk—just two pools and a public function. For $USG, pool depth is part of the peg defense.

GM GN CT ☕️ fxUSD just moved past $83.36M in supply after @protocol_fx recently marked the $80M milestone. Nice number, but I’m more interested in what is growing underneath it. The official Stats page now shows $115.69M in fxUSD Reserve Asset Value, alongside 6,652.72 stETH and 1,284.11 WBTC in protocol metrics. Total TVL is $211.82M, while fxSAVE sits around $77.75M. One snapshot can’t tell us whether the balance sheet is scaling cleanly with supply. That takes a series of snapshots. If fxUSD keeps growing, I want to watch how Reserve Asset Value and the reserve mix move with it, not just how fast the supply counter climbs. Supply growth gets the screenshot. The balance sheet is where I’d look for the follow-through.

在 @Roobet 玩了一个月，我已经赢了600多美元。 Crash依然是我最喜欢的游戏之一。 一开始，我每次兑现后都会盯着赔率线，琢磨着是不是下注太早了。现在，只要兑现时我对结果满意，就足够了。 加上一个月600多美元的收益，我对这个结果非常满意。 没人知道下一条赔率线会怎么走。你的选择权在你手中： roobet.com/?ref=all3in4out

GM GN CT ☕️ 31% is the number your eyes remember. Leverage is the part your position has to survive. On Tangent, net yield at max leverage means what’s left after financing the loop—not a free 31% APY. Sep 9 @Tangent_fi snapshot, markets above $100k TVL: USDC/fxUSD 31.05%, reUSD/sDOLA 27.54%, BOLD/USDC 24.70%, reUSD/scrvUSD 24.03%, fxUSD/reUSD 20.93%. Strong numbers, but still a snapshot: borrow costs move, HF moves, liquidation stays real. If you already LP with @protocol_fx or @ResupplyFi , this is where the trade gets interesting. Don’t size from the card alone—open the live market, check the curve, then decide whether the extra yield is worth the extra liquidation risk.

刚开始玩 Crash 的时候，我最容易纠结的其实不是没收掉，是收完以后看到数字还在继续涨。 前几天在 @Roobet 又碰到一次，自己已经收了，结果后面那条线还跑了很远。 以前我大概会盯看半天想刚才是不是应该再等等，这次反而没什么感觉，那晚最后收益$74.20，我已经挺满意了。 玩了几周以后才发现，能接受自己没吃到最高那段，反而会轻松很多。 想看看你会在什么时候觉得够了：roobet.com/?ref=all3in4out