二营长 | Finn | tech/acc (@xpcrack) header
二营长 | Finn | tech/acc (@xpcrack) profile picture

Degen follow Degen | vibe coder | member of DSH club | 方大同十年老粉 $trump 450K $swarms 360K

code is law
Joined March 2022
Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)Photo by 二营长 | Finn | tech/acc (@xpcrack)
二营长 | Finn | tech/acc retweeted
一些已经财务自由的朋友问,不要工资,有没有门路去 frontier lab 工作。这些 frontier lab 的回答是,每个人消耗的 GPU 成本和 token 成本远超他们的工资,一个不能胜任的人浪费的资源不是不要工资就能弥补的。 DeepSeek 刘胜宇把今年夏天以来的算子开发比作 “机甲驾驶员”。LLM 正在逐渐制造业化,机器的成本超过人的成本(虽然人仍然很贵),工程经验比 raw intelligence 更重要,social closure 正在形成。 创业以来接触了很多财务自由的人,财务自由只能给你拒绝工作的自由、消费的自由,不能给你创造的自由。 很多财务自由的人甚至连拒绝工作的自由都没有,比如很多上市公司高管,公司就像自己的孩子,必须一直负责下去,没有心思赶新的技术浪潮。 要获得创造的自由,也就是自由进入自己想做的领域,我只想到三条路: 1. 自己深入学习另一个领域,转行,比如从天坑专业转码;AI 可以让我们更快地入门和探索新领域,但要达到顶尖水平,仍然很难,因为专业圈子都像 LLM 这样存在 social closure。 2. 在一个领域取得成果、声誉和影响力,从而可以通过项目合作、顾问等方式进入感兴趣的其他领域,甚至横跨多个领域;(可以是公司内,比如做网络的专家也需要参与芯片、OS 等部门的讨论和合作;也可以是跨公司的) 3. 具备讲故事、拉团队、整合资源的能力,比如很多优秀创始人并非技术背景。
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel. - Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all. - Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract. - Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code. - And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while. The funnel is driven by a combination of factors: 1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel. 2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel. 3. Access. Free-tier, paid-tier, internal. So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
41
102
25
929
143,923
二营长 | Finn | tech/acc retweeted
stuckfunds.eth.limo finds crypto you bridged but never claimed. $135M+ sits unclaimed across 84 bridges on over 500k wallets (Arbitrum, Optimism, Base, zkSync, Wormhole, CCTP...).
217
260
229
2,386
1,709,218
二营长 | Finn | tech/acc retweeted
Following agreement between the community and team, All remaining $short tokens in the airdrop reserve will be burned, except for 500M reserved for future CEX liquidity needs. Always standing with our community. 🙏🩳
31
10
3
81
10,860
二营长 | Finn | tech/acc retweeted
My dear Western followers, buy gstock:native and thank me later
19
5
3
86
33,499
二营长 | Finn | tech/acc retweeted
If Shiba Inu proceeds seeded the biggest EA / AI Doomer org (FLI), then it's fitting that the e/acc think tank uses whatever capital on hand to get going. We have to scale this fight and we need a war chest. It's wartime.
Any sorts of proceeds and donations will go towards a new non-profit I will be building with my friend @mjdramstead We will be fleshing out the science of e/acc in an interdisciplinary manner, and have a think tank / policy advisory portion of the institute. It's time.
46
59
22
529
248,862
二营长 | Finn | tech/acc retweeted
So if FLI (EA AI Doomer org) was seeded with $600M in Shiba Inu proceeds injected into Vitalik's, wallet... how much should I aim for to seed an e/acc think tank to take on the Doomers? 🤔 Time to build a war chest.
E/ACC vs. D/ACC: THE DEBATE @VitalikButerin thinks slowing down AGI by four years is worth it. @beffjezos thinks that's exponential opportunity cost. They debated it live, moderated by @eddylazzarin and @shawmakesmagic. 00:00 Opening 07:02 Thermodynamics and first principles 16:04 Acceleration, entropy, and civilization 28:29 The core disagreement 32:42 Comparing and contrasting e/acc and d/acc 36:20 Open source, open hardware, and local intelligence 54:18 Should AI be slowed down? 1:02:35 Autonomous agents and artificial life 1:21:07 Crypto as the trust layer between humans and AI 1:35:37 Closing arguments
54
38
11
435
106,511
二营长 | Finn | tech/acc retweeted
为啥一直坚持买 gstock:native ,叙事想象力足够高肯定是第一位的,还有就是bsc这几个发射台我肯定要赌CZ担任顾问的发射台能成功啊,如果它能成功那龙头20m不是送钱吗? 对比下罗宾汉的long平台的龙头AI,GStock的叙事和它同类型,但AI那只是口号,GStock是实打实的和Genius平台一起做这个事情,实现这个叙事。那20m对比AI 300m,空间就足够大值得出手了。 如果你觉得币安上的叙事会比罗宾汉的叙事天花板低这么多,只能说你没见识过币安mm的威力。
49
21
11
170
129,085
二营长 | Finn | tech/acc retweeted
if you're currently 25-32, please read this carefully.
111
654
30
13,154
7,983,999
二营长 | Finn | tech/acc retweeted
There's a bull in all of us we just have to find it.
78
176
14
992
36,187
二营长 | Finn | tech/acc retweeted
看到这么多人蹭孙哥流量,大家都在猎奇,我其实有点唏嘘。 本来想写点东西,讲讲自己几个前任,后来想想还是算了。 我和我朋友都谈过捞女。 我朋友银行卡被风控了。那天出去买个 20 块钱的雪糕,女朋友就在原地等着,看着他现场 C2C 换完钱,再回来把账结了。 我自己也遇到过类似的。以前和一个前女友在停车场门口买 15 块钱的烤串,我临时接了个电话,十分钟以后回来,她还站在原地等我回来付那 15 块。 这种事其实最能看出一个人的底层观念。 有些女生眼里,男朋友跟父母差不多。回家衣服是父母洗,饭是父母做,出去消费默认就是父母买单。谈恋爱以后,只不过把这个角色从父母换成了男朋友。 她们会天然觉得,男朋友存在的意义之一就是改善自己的生活。 更麻烦的是,这些人还带着一种隐性的补偿心理。她会觉得,这段关系从一开始就是自己吃亏了,是你追的我,是你想跟我在一起,我只是勉强和你在一起,那就应该由你持续付出。自己只要花一点钱、承担一点成本,就会觉得吃亏。 甚至会把亲密关系理解成一种交换,我已经卖身体给你了,你做什么都是应该的。 这种关系最危险的地方,不是那 15 块、20 块。 而是她从一开始就没把你当成一个需要互相照顾、互相承担的人,而是把你当成了一个新的供养者来吸血。 钱很小,观念很大。 孙宇晨和景甜是活生生的例子。 关系里最残酷的一件事。 就是一个人被上一段感情伤得太重以后,往往会把伤口带到下一段。 女人年轻的时候会爱上一个黄毛,为他付出所有。做饭、养家、做家务、给男的花钱、拍裸照,把对方放进自己未来的人生规划里,最后却什么都没得到。 被辜负以后,人就开始变了。 变得尖酸、现实、利益导向,觉得这个世界都欠她。内心其实非常自卑,但表面又会用极端自负把这种自卑压住。 等到下一任出现,她不敢再像以前一样敞开心扉了。她会开始算投入产出,开始看这个男人能提供什么,开始把工具人属性放在感情前面。 她可能真的会遇到一个真心爱她、对她很好的人。 但她已经不会像当年那样爱了。 其实傻子都知道,一个人真正爱另一个人的时候是什么样。会主动,会付出,会关心,会把你放进未来,而不是永远计算自己得到了什么。 男的也一样。 有些男人年轻的时候为了一个女人什么都愿意做,甚至贷款都要给她花。后来发现,自己拼命付出的东西,对方拿到以后可能根本不珍惜。 于是他也开始变得现实,开始害怕被捞,开始算钱,算风险,算这个女人值不值得。 如果两个都被上一段关系伤过、都已经把防御练满的人碰到一起,基本就是灾难。 一个怕被利用,一个怕被辜负;一个不敢付出,一个不敢相信。两个人都在等对方先证明自己,最后谁都不肯先走一步。 十年前、十五年前的感情不是这样的。 至少那个时候的社会,经济还在上升,大家普遍相信明天会更好。房子会涨,工资会涨,事业会变好,人生还有很多可能性。 所以人更愿意享受当下,也更愿意和另一个人一起成长。 现在不是。 现在每个人都更焦虑、更现实,忍耐度也更低。 所有人身上都像带着刺,稍微碰一下,就先想着保护自己。 乱世里的感情就是这样。 从社会经济传导到职场,传导到家庭。利益的榨取,情感的博弈罢了。 大家不相信爱了,谁先认真谁就输了。
就像公司招员工一样,很多公司为什么喜欢应届生? 因为是一张白纸,可塑性强,愿意接受企业文化,也更容易跟着团队一起成长。 反过来,工作很多年的老油子,如果已经形成了一套自己的生存方式,就容易不服管理、不认价值观、喜欢摸鱼、搞小动作,甚至把跳槽,破坏公司规则,中饱私囊当常态。 能力可能更成熟,但对组织来说,可能是毁灭性的灾难。 谈恋爱其实也一样。 长期关系里单纯是很重要的品质。 真正要避开的,不是有太多前任的人,而是经历了很多关系,却从来没有处理过自己创伤的人。 身上还带着上一段关系留下来的不信任、攻击性、控制欲、回避和报复心理,却把这些东西全部带进下一段关系。 谈过很多段、每一段都没真正消化的人,最后往往会把前任留下来的创伤、算计和防御机制带到下一任身上。 这种人最容易让真正爱她的人产生救世主心态。 你觉得她只是受过伤,只要自己够真诚、够有耐心、给够安全感,就能把她感化回来。 最后很可能不是你治愈了她,而是她上一段关系留下来的伤口,反过来继续伤害你。 至于年龄,我反而更看重阶段。 我更倾向于找刚进入社会不久、但又不是完全没有社会经验的人。 一个人内心没有被复杂关系和长期博弈污染得太重,仍然有正常的信任能力,也愿意认真进入一段关系。 18—22 岁的大学生,本身生活圈子就主要在校园里。一个年级真正条件好的女生其实不超过 20 个,她们优先接触的天然还是同龄人。 你大学的时候,如果班里有女生长期和社会上三十多岁的男人谈恋爱,周围人大概率也会觉得这件事很离谱。 反而 24、25 岁这个阶段很有意思。 刚研究生毕业,或者工作了一两年,已经开始知道成年人的世界是什么样,也开始面对房租、工作、职业发展这些现实问题,但很多东西还没有彻底固化。 她通常已经开始理解成年人世界的规则,但很多价值观还没有完全固化。这个阶段如果双方成熟度、生活节奏和未来预期匹配,关系反而比较容易建立。 这个时候,如果一个更成熟的男人能提供稳定的生活、明确的未来、职业上的帮助和情绪价值,那种吸引力其实很现实。 至少相较很多还在起步阶段的同龄男生,你能提供的确定性完全不是一个量级。 所以我说找单纯的,不是找幼稚的人,而是找一个社会化已经开始、但还没有被太多烂关系和复杂博弈塑形的人。 人一旦被反复伤害、反复博弈,最后学会的往往不是怎么爱,而是怎么保护自己和模仿施暴者去伤害别人。 不要找一个没有把上一段关系的烂账,带到你这里来的人。
123
83
7
626
221,224
二营长 | Finn | tech/acc retweeted
最近带一个量化新人做 RWA 套利监控系统。 小伙子底子相当不错,也用着最顶级的模型,每天从早到晚敲 prompt,commit 效率很高。但几周下来,监控系统的跨链价格对齐、清算逻辑和资金费率计算依然千疮百孔,项目看似进展神速,但其实在原地打转。 我看了下他的工作流,瞬间明白了问题所在:他把最不该外包的规划与验收,统统交给了 AI。 他的典型日常是:丢一段模糊的需求让 AI 搭建框架,跑不通就把错误日志原封不动贴回去让 AI 修。AI 飞速打补丁,引入新 Bug,再继续喂给 AI……表面上是在高强度协作,本质上完全放弃了主动思考能力。 这正好印证了最近看到的一篇文章中提到的核心概念:倒置学习模型。 用 LLM 写代码是一场角色反转,经验尚浅的学生在指挥一个无所不知的导师,随后还要对导师的输出做决策。如果连底层的资金流向和边界条件都没摸透,拿什么去审查和校准 AI? 有效驾驭 AI Agent 所需的核心能力(系统架构、边界防御、代码审计、全局推演等等),恰恰是被持续依赖 AI 所逐步侵蚀和削弱的能力。重度依赖 AI 的人最容易产生能力错觉,例如,看到屏幕上迅速吐出成百上千行规范的代码,就误以为自己已经掌握了系统。这种专家级新手看似效率极高,实则经不起任何复杂业务的冲击。 在量化和 Web3 这种高容错率极低的领域,系统的灵魂永远不是怎么把 API 调通。 RWA 套利监控的真正难点在于:链上延迟与 Oracles 预言机喂价的时滞差、跨链流动性深度的滑点模拟、以及极端行情下的清算断崖。 这些硬核的领域逻辑和异常边界,AI 不可能替你凭空构建出 Mental Model。如果开发者自身不先建立清晰严谨的架构 Spec,AI 只会用最高效的方式,帮你快速写出一堆架构脆弱的技术垃圾。 今天我给他定了一条硬性规矩:把方向盘重新夺回自己手里。 1. 规划前置:写代码之前,先规划数据流图、状态机以及核心套利数学公式; 2. 拒绝黑盒验收:任何由 AI 生成的代码,必须自己主动思考有哪些测试用例可以应对各类边界条件,说不清为什么这么写的代码,一律不能当作生产代码; 3. 把 AI 当执行者,而非架构师:让 AI 充当高效的执行者和结对编程搭子,而不是让它决定系统的走向。 AI 编程工具就像一台大马力跑车:它能放大顶级车手的速度,但如果副驾驶坐着一个不会看路况和仪表盘的新手,油门踩到底的结果只有原地打转甚至车毁人亡。 真正的技术成熟度,也不在于用 Prompt 跑出了多少行代码,而在于系统崩溃时,驾驶员脑子里有多清醒的全局图景,以及对系统每一处细节敢不敢负责。
36
32
2
282
43,496
二营长 | Finn | tech/acc retweeted
We promised open weights for Qwen3.8. Now, time to meet them! 🎉 ⚡ Qwen3.8-27B: - A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows. - 262K native context, easily extendable to 1M tokens via YaRN. - Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0. 🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently. Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now! Download, deploy, and build something we haven't imagined yet. 👀👇 - Hugging Face: huggingface.co/collections/Q… - ModelScope: modelscope.cn/collections/Qw…
849
2,107
975
15,890
6,052,108
二营长 | Finn | tech/acc retweeted
The curse of discipline is that every day looks the same. The curse of indiscipline is that every year does.
565
32,998
1,474
159,051
9,897,801
二营长 | Finn | tech/acc retweeted
We ran DeepSeek V4 Flash through 4 agent harnesses (Claude Code, Codex, OpenCode, Oh My Pi) on 30 agentic tasks. A different harness won on each metric: success rate, cost, and speed. 🧵🧵🧵
151
107
66
1,454
317,516
二营长 | Finn | tech/acc retweeted
solana is home
solana is home
117
75
18
528
35,223
二营长 | Finn | tech/acc retweeted
It's officially Jimothy Summer. UCLA's Helen Deutsch explains why people have responded so strongly and what our collective interest in a raccoon tells us about ourselves. newsroom.ucla.edu/stories/hu…
14
69
9
219
16,112
二营长 | Finn | tech/acc retweeted
Who Is ‘Jimothy’? The Viral, Round Raccoon, Explained forbes.com/sites/danidiplaci… Photo: Getty Images
173
301
113
1,015
224,976
二营长 | Finn | tech/acc retweeted
#TRENDING: Hundreds packed Ballard today for Jimothy's official proclamation. 🦝🎨 A live art battle, huge crowds, and a neighborhood-wide celebration proved Seattle's favorite raccoon isn't slowing down anytime soon.
40
239
62
1,744
128,246
二营长 | Finn | tech/acc retweeted
Jimothy the raccoon is an idea—a symbol of being unapologetically different and free, despite dealing with setbacks. "I think he's an example to all of us that you don't have to be perfect to be loved," said one Seattle City Council member. "It is a dream of mine to encounter Jimothy in real life and then keep my measured distance from him so he can live his good little life." But what about humans? Seattle isn't exactly known for lawmakers keeping their measured distance and letting their constituents live their good little lives: reason.pub/4wZeX8r
18
62
22
306
26,911
二营长 | Finn | tech/acc retweeted
Jimothy the Raccoon has received official honors from Washington's governor, being crowned the "Washingtonian of the Day" earlier this week. They're trying to track Jimothy down so they can give him the award.
55
138
26
722
70,676