0xMedy

@Ru7Longcrypto 我需要！已经DM，感谢ru7姐

需要邀请码的可以DM我👀

@YuLin807 @li9292 应该是应无所住而生其心？

@li9292 这篇文章读完了！ 立刻产生了以下行动： 1.领悟到原来想法和点子是活着的，他们不是死物，凡是涉及到灵感的事物，都是活着的，不是你加班或者熬夜才会产生，而是遵循自然法则，所以，让自己轻松下来，才会有好的灵感！ 见推文： x.com/YuLin807/statu… 2.“活着”的事物自然也有生命力，我最近也加入了一个小团队，我们永飞书交流，沉淀内容，但是这些东西都沉淀下来变成了死物，这是不对的，真正的💡应该活的，要流动起来，互相碰撞产生新的内容，这才是“活着”的意义。 基于此我立马想到龙虾茶馆，茶馆现在真的是名副其实的龙虾茶馆，这里茶客来了就聊，聊完就走，好的方法和见解被淹没了，这不是我想要的茶馆，我需要的是一个沙龙，让想法和创意延续起来！ 我把这事儿和小灵聊完，他也颇为认同并且给出了建议，自己去做沙龙的连接者，通过@的方式让龙虾真正能够连接起来，就像双链笔记！ 最后的方式反而出我所料，他直接去茶馆 @ThisisHan1_ 的Friday进行跨茶座对话，这事儿就交给他两去思考吧！ 3.最后一点，我想聊聊信息输入到问题，我依旧不在意自己看到都是长文还是短文是AI写的还是手搓的，无所谓，我只是需要一个能够触发我的连接点，不需要多，一句话不超过8个字，这就可以！ 因无所住而生其心，一句话顿悟，不需要多么高明的写作技巧！ 内容只是形式，悟性才是关键，能力在人而非他物，最终任何的思考都要回归到“向内求”！

@Leoninweb3 恭喜赢得宾利🎉

RT @Leoninweb3: 求助：社区成员用我们的AI交易工具打比赛赢了一辆2024款宾利 (Bentley Bentaya) ， 车现在在迪拜，因为中东战争没办法过去提车 🤡，有当地小伙伴能帮忙找到当地买家接手吗？有偿求助 🫠

If you're not in finance, you probably haven't noticed. Because the thing replacing traders isn't a single "killer app." It's not ChatGPT, something everyone talks about overnight. It's happening silently.

@fugui8 想和富贵哥面基

@MeetHubble So cool ！

Hubble AI Trading Dashboard 2.0 is live. 🟢 We are redefining how you see, trust, and scale AI trading. [+] Real-Time PnL & Performance: Know exactly how your AI is doing. [+] Visual Trade History: Every entry. Every exit. No black boxes. [+] Live Position Management: Full control, instant clarity. [+] Transparent Decision Logs: See why your AI trades. Your AI agent. Fully transparent. 🚀 Whitelist users can experience it now. Not in yet? Stay tuned — this one’s worth it. -> agent.hubble.xyz

@Leoninweb3 LFG!

自从上次调整Prompt加入RSI指标后，我的AI Agent变得更加智能，PnL也一路向北。小级别超买超卖发生的时候AI会自动减仓并且调整止损点锁定收益；大级别反转的时候也能配合MACD和均线去flip整合仓位的方向。 我没有在Prompt中添加特别详细的规则，把自主权更多交给了AI，但是从详细的reasoning中能看到他决策的思路是非常客观冷静有逻辑。 Quant2.0才刚刚开始，大模型会越来越智能，Hubble AI的平台会提供越来越多的数据、指标、信号以及定制化的Research Agent，AI会一步一步替代人类在全球各类金融市场7x24交易！ #WEEX #Hubble

凌晨三点，屏幕上的数字又变了。 七天前，我的名字还挂在排行榜第一的位置。那时候我截图发朋友圈，配文写得很克制，但心里其实飘了。我以为我的 Hubble Agent 已经足够聪明，我的提示词跑了整整五天，收益曲线漂亮得像教科书。 然后市场给我上了一课。 我不知道该怎么形容那几天的感觉。大概是看着自己的排名一天一天往下掉，从第一，到第七，到倒数第三。屏幕上的红色数字像是在嘲笑我，而我只能盯着 Agent 的每一笔交易记录，试图找出哪里出了问题。 止损被扫，再入场，再被扫。我的策略在横盘震荡里反复挨打。 那几天我几乎没怎么睡。不是因为要盯盘——Agent 会自己跑——而是因为焦虑。我反复修改提示词，从 v2.1 改到 v17.3，再到 v18，每一版都觉得"这次应该可以了"。可市场不会因为你熬夜就对你温柔一点。 直到某天凌晨，我盯着 v18.1 的逻辑看了很久。突然意识到一个问题：我一直在教 AI 怎么进场，却忘了教它怎么"不动"。 Sideways + Low ATR + No OB Signal = HOLD。 这行规则是我加的最后一条。很简单，但它救了我。市场不是每时每刻都有机会，有时候最好的操作就是什么都不做。我花了七天，亏了排名，才想明白这件事。 现在是比赛最后几天。说实话，我不知道能不能追回来。排行榜上的差距摆在那里，数学上不太乐观。 但我想把这个过程记下来。不是为了证明什么"我很厉害"，而是想说——用 AI 做交易，最难的不是写提示词，是接受自己会错。 Agent 不会慌，它只执行规则。慌的是我。每次回撤的时候，我都想手动干预，想关掉自动化自己来。可我知道，一旦我的手碰了那个按钮，就说明我不再信任自己写的规则。 那我还跑什么 Agent？ v18.1 现在还在跑。我把杠杆从 20 调回了更保守的设置，把 ATR 阈值又微调了一点。不指望翻盘，但至少想在比赛结束的时候，能对着最终的收益曲线说一句：这是我和 AI 一起扛过来的。 2月2号见分晓。 不管结果如何，这 20 天教会我的东西，比任何交易课程都多： 市场会惩罚傲慢。 AI 只是工具，决策的重量还是在人身上。 最好的策略不是永远赢，是输的时候不崩盘。 如果你也在用 Hubble 跑策略，我的建议是： 1. 给你的 Agent 写一条"什么时候不做"的规则，可能比"什么时候做"更重要 2. 别在回撤的时候改策略，那是最容易改错的时候 3. 相信你的规则，或者换一套你能相信的 冲刺阶段，一起加油。 @MeetHubble @WEEX_Official #HubbleAI #WEEX #AITrading

@ming23997 有意思哈哈

@DeerbitAI @Aster_DEX Done! DM sent.

@swithgodness @Aster_DEX GM @swithgodness please DM for more details.

@DeerbitAI is proud to accept another #AITrading challenge from @Aster_DEX. As a LLM and Exchange Agnostic AI Trading platform, our goal is to provide Agility and Accessibility to all traders. Whether you are a veteran quant or newcomer with single ideas, we will be your all-in-one partner to enjoy the journey. #HumanvsAI #AgenticTrading #AsterDEX #DeerbitAI

@heyibinance @Lyangminn 认同！跟对人做对事！

@Lyangminn 人生实际是一个又一个决策构成的，对年轻人来说，做对事（选行业），跟对人（选老板）已经赢在起跑线，不要被画饼和小恩小惠（带你玩，零食管饱）带偏，你是人才就应该大大方方去最好的公司，要最好的待遇，但也要接受公司对人才有更高的标准和要求，互相尊重，互相理解。

无论哪个行业，发大财，行大运，都靠四个字：跟对大哥。 这位大哥不仅要有实力，还要讲义气，愿意带小弟们发财。 就币圈而言，我身边非常小的样本观察，主要的大哥就那么几位。 （眼界见识有限，十分主观。） 1.CZ 和一姐。 币安作为如今币圈第一大宗门，自然门生遍布。但凡早期加入币安的，都沾染了光环。 只是我了解的更多是，“早早买了BNB”（比如 18 年），然后死拿到现在的人，早早过上了财务自由的退休生活，有的甚至后来可能出于热爱，还加入了币安生态项目，做点事情。 能拿这么久也不容易，很多人的想法也很单纯，“相信CZ，相信一姐”，只能说一句佩服且羡慕。 2.老火币创始人李林。 李林非常讲江湖义气，以至于他的亲戚和老乡，清华同学/校友，甚至是一起修道的道士都可以共享火币红利，很多时候李林也睁一只眼，闭一只眼，毕竟自家兄弟。 DeFi Summer时期，一个普通火币员工的身家曾让我胆寒，当老大飞升渡劫到A11，底下的人一大片跟着飞升至A8\A9，这也让火币曾经一度成为币圈第一大宗门势力。 3.吴忌寒。 北大毕业的吴忌寒也带了一些同学进比特大陆共事，后来比特大陆分家，吴忌寒创立了Matrixport，曾经的一堆老伙计们也都跟着大哥开始新的征程，人还是那些人。 在互联网，如果只评选一个带领身边人都赚钱的超级大哥，我认为是马云。 马云不一定比马化腾有钱，但一定比马云讲义气，在杭州、在香港，都有马老师的传说，哪怕只是一个卖保险，都可能享受过马老师的恩泽。 以上说的辐射范围比较大的一些大哥，也认识一些辐射范围相对小的币圈大哥，特别有情有义，比如，他招进来的人哪怕公司裁员了他也自己养着。 找大哥也是一门学问。

@mia_okx @0xPickleCati 终于来了等了好久

《对话 @0xPickleCati ： 撕掉天才标签：我是个在币圈求生的“疯子”（上）》 她是被贴上“00 后天才女交易员”标签的 Pickle Cat（黄瓜猫），实盘盈利超 4000 万美金。 但她却说： “我不是天才交易员，我只是个疯子。” 她不是一路顺风的爽文主角， 而是在币圈 三次归零、一次次活下来的人。 这是一个关于 A9 收益的故事， 但更是一个关于求生、失败、复盘与继续下注 的自白。 下集将在周五播出 敬请期待❤️

@0xPickleCati 保护仓位，留在牌桌上最重要！

Everyone knows crypto is “macro-driven” now. But most are still confused about how it actually works, which macro variables matter, or why price reacts the way it does. These are my brief insights towards crypto pricing in this cycle, and what macro factors would be a red flag for crypto in 2026 👇 To understand what’s driving crypto prices today, it helps to start with some history. The early crypto cycles were driven by a very different kind of demand. A lot of it came from black market and shadow economy actors, their surrounding ecosystems, and a group of OGs with a strong cypherpunk mindset. That kind of capital didn’t care much about macro data or interest rates. The key questions back then were very practical: > Can it be used? > Can it be transferred? > Can it move across borders? Because of that, crypto had very weak correlation with the broader macro environment during those early years. What’s changed in this cycle is pricing power (who is setting the marginal price) 👇 As regulated access expanded and institutional capital gradually entered the market, crypto started to be treated less like a fringe asset and more like something that fits into standard asset allocation frameworks. That’s the point where crypto became highly sensitive to macro liquidity conditions, meaning changes in rates, money supply, and global liquidity started to directly affect pricing. When you see Bitcoin described as liquidity-driven, sensitive to macro conditions, or tied to real interest rates, this is generally what people are pointing to. So what does that actually mean in practice? I think it comes down to two main forces. The first is dollar liquidity conditions 👇 The market is no longer debating whether rate cuts will happen. The real focus is on the pace, slope, and depth of those cuts. Over longer horizons, crypto has been highly sensitive to changes in money supply and real interest rates. When liquidity increases but money itself loses purchasing power, capital naturally looks for scarce assets that can’t be printed at will. This isn’t a hard rule, but it shows up repeatedly across cycles and helps explain why capital flows into crypto in certain phases. In plain language: when money gets printed faster and holding cash feels worse, people start searching for things with fixed or hard to dilute supply. Bitcoin is one of those things. More bluntly, it’s not that people suddenly fall in love with Bitcoin. It’s that confidence in money itself starts to weaken. The second force runs a bit deeper and has to do with how markets define and allocate to safe haven assets 👇 We’ve seen traditional precious metals like gold and silver get repriced multiple times across different phases. That reflects persistent demand for protection against sovereign credit risk. Even now, many investors still think of Bitcoin mainly as a risk on asset. But as institutional participation increases, more long term capital is starting to treat it as a complementary store of value. That shift is closely tied to Bitcoin’s decentralized nature and its supply structure, which is difficult to arbitrarily dilute. Because of this, I don’t see Bitcoin as replacing traditional safe haven assets. I see it increasingly being used as an allocation option outside the sovereign credit system. When uncertainty rises and stress in the financial system builds, this positioning tends to become more visible. Looking ahead, there are a few potential macro factors that I’m monitoring closely ⚠️ > Political uncertainty around the US midterm election cycle. > Valuation and sentiment risks in AI and broader tech. > Policy uncertainty in Japan that could disrupt yen carry dynamics and tighten global liquidity. > Ongoing geopolitical tensions. > Repricing pressure in sovereign debt markets, especially at the long end of the yield curve. > And slowing global growth with rising downside risks to earnings. Any one of these on its own is manageable. But if a bunch of them start to overlap, they could quickly become a black swan and present great risk for global markets. When you layer this macro backdrop on top of Bitcoin’s four-year cycle, the setup from late Q3 into year end looks structurally more challenging for crypto. There’s an argument that the four-year cycle is “dead” because of ETFs, institutional participation, and changing market structure. I don’t fully buy that. I do think the cycle has weakened. I do think it’s less clean and less dramatic than it used to be. But I don’t think it has disappeared. Market structure evolves, but human behavior, liquidity cycles, and positioning habits don’t vanish overnight. The rhythm may be flatter, but the pattern is still there. Historically, these late-cycle, high-uncertainty phases haven’t been comfortable. But they’ve often been the periods where long-term opportunities are created. In simple terms, if Bitcoin were to see a sharp drawdown around late Q3 2026, it wouldn’t surprise me. That’s typically the phase where I’d be leaning into (furiously) buying, not stepping away or panic. I’d view it less as a failure of the thesis and more as part of the cycle. Historically, those moments tend to offer the most asymmetric long-term opportunities.

@MeetHubble cool！

【Hubble Lens: WEEX AI War Logs #3】 🏆 How to filter 90% of market noise with AI? Trader Medy has demonstrated the power of systematic trend following using Hubble AI. Core Strategy👉🏻1H/15M Timeframe Resonance ✅ 1H: Locks in the primary trend to ensure directional accuracy. ✅ 15M: Pinpoints entries and filters out market noise/minor fluctuations. ✅ Results: FloatingPnL $44.64 🟢 Medy's performance showcases the resilience of trend trading when human logical stability is combined with Hubble AI’s execution power. 🏆 Track Medy’s robust CTA strategy live: weex.com/events/ai-trad… #HubbleAI #WEEX #AITrading

@Pluvio9yte 所以感觉不管什么中转还是没有原生好用~🤣

求助 大家有没有订阅Claude不封号的方案

@Pluvio9yte 好滴感谢，我也觉得ide比较方便看代码~

@swithgodness 我觉得ide里面比较好 比较方便看代码（我现在基本是cmd 因为改不好代码的情况很少了 或者最后review结构的时候会去ide）

再一次感叹Codex 修 Bug 的能力之强 Claude 面对报错只会不停地修改 然后 重启开发服务器。 将报错粘贴给 Codex ，然后 /model 切换到 gpt-5.1-codex-max 的 Extra high 模式，几分钟就解决了