Developer Product Operation / Product Manager. AI List: https://xitter.cf/t.co/4AX7HgbyHO Indie Hacker List: https://xitter.cf/t.co/piwCEswHW7

Joined October 2012
Do you remember when you joined X? I do for the 14th year. #MyXAnniversary
6
Block +1
“这么好的前端提示词,你怎么不早点告诉我?” 被群友@了,发现被别人洗稿了,还是同平台😂
There's a new version of this post
7
端上来端上来
Can’t wait to pro and tell next week.
11
出一个Tibo重置计划排期Official版😊
I bought a new domain and I’m very excited for it. Very
26
thanks for ur sharing!
how do you trust your agents' work more? copy this post/link and give it to your @bot or agent! 1. codebase improvements: this is just good software engineering - better architecture, stronger types and invariants, clear boundaries. agents love to copy what exists and take the shortest path, so make the easy thing the right thing. categorically eliminate problems through code and you won't need to think about it again. 2. static analysis: lint rules, tests, CI, verification (eg TLA+). very token efficient ways to catch issues ahead of time. 3. prose: comments, rules, AGENTS.md, skills - anything that can't be enforced and requires agent judgment. use it minimally, but do use /create-verification-skill. 4. human review: the hardest to scale. at the rate agents land code, it's pretty much impossible to thoroughly review every PR, so spend that effort on the highest leverage point, designing an environment where agents do the right thing by default. this forms an inverse pyramid: you want most of your effort at the top, least at the bottom. codebase improvements are deterministic, making it more token efficient, and lets you sleep more soundly at night. install github.com/cursor/plugins/tr… and create a new Project: 1. look through my slack/chat messages and code review comments where i've given feedback on PRs, code, architecture 2. use /recall on my past transcripts to learn how i steer agents 3. /poteto-mode use /correct and /architect on past commits to find the most common pitfalls agents fall into, then propose a plan to rearchitect the code i own so it's more agent friendly. answer open questions with prototyping rather than defer to me, and back the plan with real data. 4. before you begin, use /sharpen + /in-your-own-words and confirm this prompt with me
23
Just retweeted
Antigravity 向付费 Google AI Pro 用户开放 Gemini 4 Argon,Argon 有单独的每周额度,不占 3.8 Flash 额度 Google 的 AI 编程工具 Antigravity 今天开始向付费的 Google AI Pro 订阅用户推出 Gemini 4 Argon。官方模型页写明只限正式付费的 Pro,试用期的 Pro 用不了,免费用户和 Google AI Plus 用户也不在范围内。能用的地方是 Antigravity 桌面 App、命令行工具 CLI 和 IDE 扩展(装在 VS Code、JetBrains 等编辑器里的插件),Antigravity IDE 里暂时选不到。目前在逐步推送,官方建议先更新到最新版。 Pro 用户用 Argon 有一份单独的额度,只按周计算,不像原有额度那样每 5 小时刷新一次;它和原来给 Gemini 3.8 Flash 用的额度分开,用 Argon 不会吃掉 3.8 Flash 的额度。每周到底给多少、哪天重置,推文和官方文档都没写。文档对所有额度的说明是,消耗多少取决于智能体实际干了多少活,任务越复杂用得越快。 官方模型说明:antigravity.google/docs/mode…
We’re rolling out Gemini 4 Argon in Antigravity to paid Google AI Pro subscribers today. Demand has been incredible and we've scaled capacity deliberately to get here. We’ve rolled out a separate weekly-only Argon quota for Pro on top of the existing Gemini bucket for 3.8 Flash so you have the best of Gemini for all your tasks.
119
8
15
132
53,348
The new Gemma docs! gemma.ai/ 🥁🥁🥁 From capabilities, to fine-tuning guides, inference optimizations and educational content on how the models work. All at central place!
10
9
140
8,275
Just retweeted
Magpie 现在全面插件化了 🧩 以前社区插件只能接入新的订阅;现在 Magpie 自己的几乎每一层都向社区开放,一个 npm 包就能扩展: • 订阅 / Provider:登录、刷新、模型、请求、用量全由插件负责 • 网关中间件:改写经过 Magpie 的请求和响应 • 路由策略:自己决定请求怎么在多个账号、多个模型之间分配 • onUsage:每一条用量记录插件都能收到 • 工具:插件写的工具通过 Magpie 的 MCP 服务器,给每一个接入的 Agent 用 • Agent:让 Magpie 支持一个新的 Agent • 页面:插件在 Magpie 窗口里有自己的页面,跑在沙盒里 入口:Magpie › 插件 › 发现,每类插件各有分区;点「编写插件」直接打开文档。 文档:usemagpie.ai/docs/plugins 社区仓库:github.com/magpie-community/… 三个示例已经发到 npm:tools-calc(精确计算、单位换算、日期计算工具)、page-usage(用量一览页面)、middleware-reserve(留余量的路由策略)。 magpie plugin add @magpie-community/tools-calc 一行装好。来写插件吧 👇
66
45
8
448
54,314
旧的放转转🤗(乱入
哈哈,这就是我给自己的赛马模式,我为了加新的产品链接,会淘汰一下旧的产品链接(
1
17
DeepSeek V4.1 Flash is free on OpenCode, the duration is unclear and the context is 200k
DeepSeek V4.1 Flash is now free in OpenCode
43
DMed👀
Our sister companies within @shanda_group have a few openings in DevRel, GTM, and Growth in Shanghai and Singapore. If you love tech and AI, can yap, enjoy giving demos, and actually like meeting people, come talk to me. Especially if you’re the person who shows up at an event knowing nobody and leaves with plans to organize the next one. I’ve been here for ten months now, so here’s my very personal recruitment pitch. The food is awesome. We go out for barbecue and beer probably more often than we should. There’s always someone ordering more skewers after everyone says they’re full. In my role, I’ve had business-class travel covered, attended events around the world, and met people whose work I used to follow online. I get to spend my days talking to people, trying things with AI, and using all the tokens I could ask for. Honestly, it’s been a pretty good ten months. There’s also room to try things beyond whatever your job title says. Write something, build a demo, get on stage, bring people together, then take what you hear back to the product team. If you’re curious and know your way around AI, there’s a lot you can do here. And I’m sweetening the deal. Join us and I’ll give you one of my favorite cars from my collection. Yes, you can take it home. I’m putting my own cars on the line for this recruitment effort. DM me if you’re interested, or send this to that friend who knows everyone. Preferably before I change my mind about the cars.
24
Just retweeted
Argon has this amazing natural understanding of the world. It can do so many things from first principles. Here's a pelican riding a bike, using fourier epicycles. Video made by Argon. Time-synced Lyria 3.5 music prompted by Argon.
235
593
238
8,406
2,210,336
great guide 👍👍
our team shipped a bunch of guides to help you build with Gemini 4 Argon: - Dev guide: aistudio.google.com/learn/ge… - What's new + migration checklist: aistudio.google.com/docs/lat… - Building 3D games: aistudio.google.com/learn/bu… - Long decode guide (1M output): aistudio.google.com/docs/lon…
1
37
Hi, I've been using your product @hark_labs recently, and I'd like to suggest a small improvement to enhance the user experience: When a user reaches their usage limit, they wouldn't receive any task executions or be able to respond to Hark's subsequent execution strategies. Though it's as expected, but I think you may allow users to tell Hark to automatically execute previous tasks and retrieve results after the usage limit is reset (or let them send a yes/no statement or provide a Y/N option to choose). Could you please review if it's necessary? Thank you very much for your time and consideration. @adcock_brett
49
Gemini的Win客户端是不是用的人比较少?今天打开提示更新,但一点Restart to update程序就消失不见了,得在任务管理器杀掉了才能重开,然后点restart开始进入循环...我想用用Gemini 4 Argon呀
44
iOS 创建PWA应用图标展示异常的问题,已经好几个Beta版本没修复了,最新的iOS 27.2 Developer Beta 4 依旧没解决。 The issue of abnormal icon display when creating a PWA app on iOS has persisted for several beta versions, and the latest iOS 27.2 Developer Beta 4 still hasn't resolved it. reddit.com/r/PWA/comments/1w…
1
121
开的普通premium没有bot权限,之前在用Hark直接实现这个效果🤗,我是直接把我建的AI list链接发它,让他抓list里的热门信息和主流科技媒体的报道,不过速度和效果肯定比不上X官方自己的,刚刚用Tw老师的提示词优化了一下,效果如下:
教大家如何享受纯净舒服的推特时间线,不错过你的好友的高质量信息流。 给到 Grok Bot 的提示词大概是这样,大伙也可以玩玩。 每天晚上8点,帮我整理今天中文推特圈的精选 来源是我关注的中文账号,加上你觉得不错的中文账号。只用 X 官方 API 读帖子,跟前一天去重。 排版: - 开头放「今日主线」,3 条短句 - 每个人单独一节,写上名字和 @,节和节之间加分隔线 - 每节 2 到 4 条要点,格式是「加粗主题:内容 + 原帖链接」,每条写两到四句,讲清楚他说了什么、关键细节和背景 - 每节最后用一句话写你的看法 - 零碎工具放到最后的「小工具」一节 - 大约 10 到 15 节,不写破折号,少用句号,不要有 AI 腔,简单清晰就好 然后我每天就可以收到这个非常有用的消息了,让我少刷推特,更加专注一点。
51