
Co-Founder Westar Labs | 🛠️ $STC & AI Explorer | Ex-Chief Architect Weibo (NASDAQ:WB) What we hear is opinion; what we see is perspective. 此有故彼有 此生故彼生
Joined May 2007
- Tweets2.6K
- Following1.2K
- Followers9.9K
- Likes1.4K
Tim✨ retweeted
I'm amazed by this line from @deepseek_ai 's official announcement
"Not lured by praise, not frightened by slander, follow the righteous path and discipline oneself with integrity"
DeepSeek's real edge isn't just the tech, it's the ethos behind the work. And that's what will carry them further than any benchmark ❤️
Tim✨ retweeted
Replying to @jack
Jack 这个关于在 Block (Square) 用 AI 重塑组织架构确实是一个伟大想法,延续两千年的罗马式层级是因为人类管理跨度有限,而 AI 可以通过构建公司世界模型取代中层管理的信息调度功能,让公司像一个真正的智能体一样运作。
愿景特别宏大,但从工程角度来看,要实现所谓“公司世界模型”,还存在巨大的鸿沟:
从罗马军团到 AI 智能体,组织设计的核心确实是信息路由。但这里有个非常大的区别,并不像 AI 去编写或优化一个大型项目的代码那样,人并不是一种 Deterministic 的组件。
AI 可以监测一个系统每秒请求数(QPS),就像 Jack 想用收付款当系统运行核心日志;但 AI 无法捕捉员工的士气波动、非正式沟通中的信任,以及那些无法被数字化的管理黑盒。
了解大公司的人都知道,每个微型组织单元都有一套维持运转的暗逻辑:看不见的管理优先级、不成文的约定、甚至是多年磨合出的默契,维持了整个庞大机器的运行。去掉这些暗知识,组织很可能不会变成高效的智能体,而会变成一个混乱的蜂群。
目前 AI 对组织的理解更像是看一张降维后的平面图。靠这种降维观测来取代层级管理,短期内面临的系统性风险可能远超想象。
Tim✨ retweeted
文字输出的品味其实可以靠做减法来维持下限,只要废话够少,基本面就不至于太差。
反直觉的是,没品味的人往往对 AI 吐出来的那些华丽辞藻沾沾自喜,却不知废话越少,翻车的概率反而越低。
但在图、视频、代码或者 agent 逻辑里,平庸通常是结构性的,没有这种一删了之的低成本捷径。
Tim✨ retweeted
Replying to @dotey
AI 结果本质上是基于概率的文本推理,它确实存在一些让部分人不适的固定范式(比如过度使用排比或特定的标点)。
但当我们谈论“反对 AI 味”时,到底在反对什么?是反对高效的自动化生产,还是在反对低质量的废话堆砌?不可否认存在 AI 垃圾信息,但这和真人发帖一样:有人在灌水,有人在博取流量,有人在深思熟虑。
关键在于意图。对于大部分文字表达能力有限的人来说,AI 恰恰是实现词能达意的桥梁。如果 AI 能更精准地传递发帖人的原意,提高沟通效率,我们为什么要拒绝这种进化?
Tim✨ retweeted
Replying to @FuSheng_0306
第一条保密确实是刚需,但执行起来感觉不好说。毕竟 agent 对保密这种高阶意图的理解很容易跑偏,特别是公司资料塞多了以后。
感觉辅助一个定义隐私边界的 Skill 比较稳,把能说和不能说拆解细一点,省得它在某些场景推理着推理着就把老板秘密给说出去了。
Tim✨ retweeted
Replying to @jolandgraf
非常认同。Agent 的进一步应用,本质上是迈向可编程基础设施。
安全性、强隔离、方便迁移、自动化生命周期,这些是 Agent Loop 的生命线,也是 localhost 很难逾越的短板。
最近流行的从 Cloud 搬到 Mac Mini,对自动化浅尝者或许够用,但对追求 7x24h 并行协作的深度玩家来说,感觉并不是一个值得尝试的方向。
Tim✨ retweeted
Replying to @ChromiumDev
感觉 WebMCP 是 Web 开发的一个分水岭。
接下来,互联网将只有两种页面,Agent-friendly(智能体友好) 和 Legacy(传统网页)。
这就像十多年前的 Mobile-friendly 浪潮,要么主动适配,接入 Agent 时代的结构化协议;要么在 AI 自动化的洪流中彻底失联。WebMCP 或许正在定义这种新规则。
Tim✨ retweeted
Replying to @AtlasForgeAI
好奇这种 principle 最终反馈差异,大部分任务第一性原理就是执行及完成,强调“学习”会不会反而落入马屁输出。
> 系统中最高层级的准则是:优化学习速率,而非任务完成。
> 大多数 Agent 的评判标准是否完成了任务,回答是否正确?请求是否执行?这优化的是短期表现。
> 而我的评判标准是我是否变得更好了,是否从这次互动中学到了东西?
Tim✨ retweeted
Replying to @0xDevShah
OpenClaw is less of a stable foundation right now and more of a high-speed construction site.
Also, a true top-down observer wouldn't just wrap OpenClaw, it would probably replace it. If the agent is smart enough to learn your 90%, it won't need a middleman for the hands part.
Tim✨ retweeted
Replying to @Piglei
现实项目中代码产出的弹性其实非常大,主要取决于几件事:
1. 代码量与认知负荷
AI 确实能让可交付代码行数爆炸式增长,但人类大脑处理和理解的容量并没有相应提升。
当代码体量快速膨胀时,review、调试、改动、迁移的认知成本会急剧上升,会造成团队/个人主动或被动地“刹车”,要么控制需求增速,要么限制生成代码的体量和复杂度。
2. Review 的真实成本
Review 消耗时间其实不太好量化,考虑到工程师由于 AI 辅助写代码时间已经减少,加上调试疑难杂症问题的减少,可用于 review 时间相比于传统开发模式是增加的。
另外 AI 加持的 100% 的测试覆盖也可以适当降低 review 心智负担。
另外想到一个有意思问题,跟银弹无关,当 n 倍效率持续几年之后,代码库会不会团队中没人真正了解了,即使主架构师的 knowledge 也比不上 AI 的 knowledge……
Tim✨ retweeted
Replying to @huang_song_
最近 X 上不少关于 Typeless 语音输入法好评,很多人夸它转写质量高、用在聊天/prompt 时特别爽,比传统语音输入省心多了。实际简单用了一下,流畅度也OK,确实是个挺有潜力的输入方式。
但个人不太会把它设为首选,原因主要还是场景和习惯限制:
1. 公共场合(办公室、咖啡馆、路上)直接语音输入总有隐私尴尬,而且别人听到你在自言自语也很怪;嘈杂环境识别也容易掉链子(虽然它估计比默认语音识别强很多)。
2. 所有语音输入的通病,为了让它识别更准,会不自觉把话讲得更通顺、边说边想怎么形成系统能理解的句子,同时还要检查屏幕反馈有没有输出正确,这其实打断了正常思考流。
3. 文字输入(尤其是中文)当然也存在输入时候需要检查输出的分心,但多年肌肉记忆已经适应了这种“输入+即时反馈”的节奏。
4. 真正独处环境、需要大量输出的场景(写长文、写代码文档),键盘大多也不是瓶颈,卡住的往往是 idea 本身,而不是手速。
所以对键盘工作者来说,语音输入很长时间可能还是辅助工具;但如果你的使用场景是日常发语音信息聊天、隐私不敏感、或者手不方便打字(开车、做饭、带娃等),它确实能大幅提升效率,值得一试。
Tim✨ retweeted
Replying to @arpit_bhayani
OpenAI 仅用 1 个 PostgreSQL 主库 + 50 个只读副本,就支撑起了 ChatGPT 的 8 亿用户规模
在优化基础设施支撑增长过程中,我们得出了一个新的认知:PostgreSQL 在读密集型场景下的可扩展性,远比许多人以为的要强得多。
Tim✨ retweeted
Replying to @jocyiosg
我觉得也没必要怪赌场(交易所)和 meme,它们只是顺应了用户天然的赌性需求,这种需求在牛熊时期都存在,只是熊市时候更容易被诟病。真正的问题在最近一两年 Crypto 领域真正解决用户需求或痛点的场景一直稀缺,ETF/稳定币/meme 也不需要太多新创新团队的支持,新项目团队自然也无法捕获到这些领域中的价值。
Crypto 领域大家经常寻找的创新,更像是手里握着区块链+AI/区块链+社交/zk 这些锤子,到处找能砸的钉子,而不是找到用户和行业到底需要什么愿意为啥买单。比如原生的 AI 系统中有区块链痛点吗?至少暂时看不到。因此去这些领域创新和探索,空间肯定有,但是中短期自然非常窄,承载不了行业的厚望。
创新团队真正的难点是:需要找到那个需要用 on-chain/去中心化才能更好解决的真实需求,然后再反过来设计机制和代码,而不是反过来。
Tim✨ retweeted
Happy 2026! Will this be the year we finally achieve AGI? I’d like to propose a new version of the Turing Test, which I’ll call the Turing-AGI Test, to see if we’ve achieved this. I’ll explain in a moment why having a new test is important.
The public thinks achieving AGI means computers will be as intelligent as people and be able to do most or all knowledge work. I’d like to propose a new test. The test subject — either a computer or a skilled professional human — is given access to a computer that has internet access and software such as a web browser and Zoom. The judge will design a multi-day experience for the test subject, mediated through the computer, to carry out work tasks. For example, an experience might consist of a period of training (say, as a call center operator), followed by being asked to carry out the task (taking calls), with ongoing feedback. This mirrors what a remote worker with a fully working computer (but no webcam) might be expected to do.
A computer passes the Turing-AGI Test if it can carry out the work task as well as a skilled human.
Most members of the public likely believe a real AGI system will pass this test. Surely, if computers are as intelligent as humans, they should be able to perform work tasks as well as a human one might hire. Thus, the Turing-AGI Test aligns with the popular notion of what AGI means.
Here’s why we need a new test: “AGI” has turned into a term of hype rather than a term with a precise meaning. A reasonable definition of AGI is AI that can do any intellectual task that a human can. When businesses hype up that they might achieve AGI within a few quarters, they usually try to justify these statements by setting a much lower bar. This mismatch in definitions is harmful because it makes people think AI is becoming more powerful than it actually is. I’m seeing this mislead everyone from high-school students (who avoid certain fields of study because they think it’s pointless with AGI’s imminent arrival) to CEOs (who are deciding what projects to invest in, sometimes assuming AI will be more capable in 1-2 years than any likely reality).
The original Turing Test, which required a computer to fool a human judge, via text chat, into being unable to distinguish it from a human, has been insufficient to indicate human-level intelligence. The Loebner Prize competition actually ran the Turing Test and found that being able to simulate human typing errors — perhaps even more than actually demonstrating intelligence — was needed to fool judges. A main goal of AI development today is to build systems that can do economically useful work, not fool judges. Thus a modified test that measures ability to do work would be more useful than a test that measures the ability to fool humans.
For almost all AI benchmarks today (such as GPQA, AIME, SWE-bench, etc.), a test set is determined in advance. This means AI teams end up at least indirectly tuning their models to the published test sets. Further, any fixed test set measures only one narrow sliver of intelligence. In contrast, in the Turing Test, judges are free to ask any question to probe the model as they please. This lets a judge test how “general” the knowledge of the computer or human really is. Similarly, in the Turing-AGI Test, the judge can design any experience — which is not revealed in advance to the AI (or human subject) being tested. This is a better way to measure generality of AI than a predetermined test set.
AI is on an amazing trajectory of progress. In previous decades, overhyped expectations led to AI winters, when disappointment about AI capabilities caused reductions in interest and funding, which picked up again when the field made more progress. One of the few things that could get in the way of AI’s tremendous momentum is unrealistic hype that creates an investment bubble, risking disappointment and a collapse of interest. To avoid this, we need to recalibrate society’s expectations on AI. A test will help.
If we run a Turing-AGI Test competition and every AI system falls short, that will be a good thing! By defusing hype around AGI and reducing the chance of a bubble, we will create a more reliable path to continued investment in AI. This will let us keep on driving forward real technological progress and building valuable applications — even ones that fall well short of AGI. And if this test sets a clear target that teams can aim toward to claim the mantle of achieving AGI, that would be wonderful, too. And we can be confident that if a company passes this test, they will have created more than just a marketing release — it will be something incredibly valuable.
[Original text: deeplearning.ai/the-batch/is… ]
Tim✨ retweeted
续费悖论 🛠️ ENS 及其他 ID 都需定期交费,否则身份就被回收,这与 Web3 宣称的“主权资产”理念相悖。不续费就收回的逻辑也是 Web2 注册商的遗产。
治理困境:若设计为一次性买断终身有效,协议将失去长期维护资金支持;且稀缺后缀被囤积者永久占据,导致后来者由于成本过高而无法进入损害生态活力。4
Tim✨ retweeted
“ENS can’t resolve the paradox between convenience and privacy / identity equals ledger.”
ENS is just a naming layer; privacy is primarily about address reuse, and you can use fresh addresses (or stealth address systems) while still using an ENS name as the human readable identifier. Tools like Fluidkey (stealth address style receiving) address the “public balances” concern at the wallet/payment layer without requiring ENS to make your entire financial history public.
“Most users won’t sacrifice onchain privacy just to avoid copying an address.”
Using ENS does not force a single, forever linked address: you can rotate receiving addresses, use different addresses per context, and choose what (if anything) you publish in public records. In practice, ENS improves safety and UX (human verification, phishing resistance, readable destination) while letting users adopt privacy preserving address practices alongside it.
“Identity silos: besides .eth there’s .sol/.bnb/.apt; same name on different chains creates misdirection; no global standard.”
You do not need separate siloed namespaces to send to different chains: ENS can resolve to most coin/address types (multicoin support), so one .eth name can direct payments across many networks and assets. Other namespaces exist because other ecosystems chose to build their own, not because ENS can’t be the unified name layer.
“Renewal Paradox: periodic payments contradict ‘sovereign asset’; reclaiming on non renewal is a Web2 legacy.”
Renewals are not a paradox; they are a deliberate antihoarding mechanism that keeps unused names from being locked forever and reduces rent seeking on scarce names. If you want “no renewal worry” UX, you can use free/cheap subnames, so users effectively get a stable identifier without fighting scarcity at the root level.
“If it were lifetime validity, maintenance funding is lost; scarce suffixes get permanently occupied by hoarders; ecosystem vitality suffers.”
This is exactly why ENS uses renewals: perpetual ownership of scarce roots predictably leads to permanent capture. Meanwhile, subnames provide abundant, low cost identity at scale, which is the practical path to mainstream adoption without turning the base layer into an irreversible land grab.
“Scarcity ≠ practical value; Web2 domains provide business traffic/brand trust; Web3 domains are just ‘digital stamps’ (transfer convenience only).”
ENS names are not limited to “transfer convenience”: they are an onchain identity primitive used for wallet UX, app login/identity, profiles, and decentralized websites (e.g., .eth names resolving to content via gateways). And if you want Web2 brand continuity, DNS names (like .com) can be imported into ENS (with DNSSEC proofs) so existing Web2 trust can carry into the ENS ecosystem.
“ENS should first address privacy by automatically resolving to single use stealth addresses; return to tool nature; stop being show-off/hoarding chip.”
“Automatic stealth” is a wallet/payments design choice because it changes how recipients derive addresses and how senders construct transactions; ENS can support those patterns by pointing to compatible resolution targets, but it’s not something the naming protocol should unilaterally enforce. The adoption path is interoperability: let wallets implement stealth/rotation (Fluidkey style), while ENS remains the stable, portable identifier that those wallets can resolve safely.






