Blep-Cat(海青)

187 posts

Blep-Cat(海青) banner
Blep-Cat(海青)

Blep-Cat(海青)

@heycqing

👾Independent Developer | FE-centric FullStack ☕️AI | Tech Delivery | Scoping | Fast Validation 🎞️Photographer | Film Obsessed & Humanity 🐈‍⬛ Curiosity Killed

ほっかいどう Katılım Haziran 2018
116 Takip Edilen30 Takipçiler
Sabitlenmiş Tweet
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
AI 幻觉,到现在也没被真正解决。 做复杂项目的时候,从 0 到 1,AI 确实能按你的文档做约束性处理——很多人就停在这一步,然后开始惊叹、开始夸大。 但真正落地、从 1 到 100 的时候,AI 的表现还不如一名中级开发。给再多提示词、再多约束,作用也不大,最明显的效果就是大幅消耗你的 token。 中推也好、非中推也好,都在无限放大这种幻觉。 击鼓传花,郁金香泡沫——从利益出发,无限夸大、无限制造焦虑,因为这样最符合既得利益者的利益最大化。 说到底,这是一种更露骨的营销和获客手段。 而当你打不过它的时候,那就加入这场"泡沫狂欢":看似参与了,其实产不出什么实际东西。 真要这么玩,不如去开家网红奶茶店——起码奶茶店死得慢一点。 所以我决定做个实验。 偏前端全栈程序员,从大学时期接过商单,生活费自给自足。目前偶尔帮解决一下技术问题,偶尔接一下商单。 SAAS时代 | AI时代我在验证一件事: 业务团队和非技术创始人,到底能不能更快地把模糊问题,变成能真正跑起来的技术方案。 这里会持续记录: - AI / 自动化工具怎么真正落地到业务流程里,而不是 demo 里 - 洞悉潜在需求,需求拆解和快速验证 - 技术方案怎么从业务问题里一层层拆出来 - 数据复盘 如果你在做产品、做运营、在创业,或者经常卡在"我知道有问题,但不知道技术上怎么解"——这里值得你关注。 不夸大,不贩卖焦虑。只记录真实发生的、跑得通和跑不通的东西。
中文
1
0
3
210
Codez
Codez@0xCodez·
Creator of Claude Code : "I don't prompt Claude anymore. I create loops - and the loops do the work. My job is to create loops." In 30 minutes, Boris breaks down his real daily Claude Code setup. Claude + loops + routines + dynamic workflow Worth more than a $500 vibe-coding course
Codez@0xCodez

x.com/i/article/2062…

English
15
12
186
27.5K
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
The 14-step automation stack is real. Set it up. But step 10 will quietly destroy you if you skip this part: "A routine runs autonomously with no follow-up questions. Anything ambiguous becomes a coin flip on every run." That's buried in the doc. Here's what it actually means: Most people's prompts look like this: "Every morning, check GitHub issues and summarize them." A Routine running that at 7am will give you something different every single day. Sometimes a Slack message. Sometimes a draft PR. Sometimes nothing. The prompts that actually work define 4 things: - What to look at (exact repos, exact labels, exact timeframe) - What "done" looks like (output format, where it goes) - What to do when something's missing (fail loudly or skip quietly) - What NOT to do (the deny list matters as much as the allow list) The automation stack is a force multiplier. A vague prompt × automation = vague results at scale. A precise prompt × automation = leverage. Fix the prompt first. Then set the cron.
Codez@0xCodez

Creator of Claude Code : "I don't prompt Claude anymore. I create loops - and the loops do the work. My job is to create loops." In 30 minutes, Boris breaks down his real daily Claude Code setup. Claude + loops + routines + dynamic workflow Worth more than a $500 vibe-coding course

English
0
0
1
300
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
@skeptrune true, The more I organize it, the more I feel like something's still missing, but the knowledge infrastructure part? Yeah, that tracks.
English
0
0
1
5
Nick Khami
Nick Khami@skeptrune·
true agi will come from my 97 .md files
Nick Khami tweet media
San Francisco, CA 🇺🇸 English
36
17
562
16.5K
Alex Finn
Alex Finn@AlexFinn·
If you're not using Hermes Agent + the X API together you're a moron Most powerful way to find money making opportunities Here's what you need to do: 1. Sign up for the X API (developer dot x dot com) 2. Set up Hermes in telegram or desktop (see my vids) 3. Ask Hermes to set up a cron for every morning at 9am 4. Ask it to scan X for challenges people are having in your niche (describe your niche) 5. Have it give a list of the top 3 challenges, and a recommendation for how to build an app or service to solve them 6. Have it build a prototype of those apps If you have any trouble setting up the X API just ask Hermes to do this If you want to take this a step further, you can even have Hermes automatically build you a prototype with each cron Do this, and I guarantee you come up with multiple business ideas every week you can make money off of We have an unreal opportunity right now to leverage these AI tools 99% of people have no idea exist The fact that nobody knows about them means you have an INSANE opportunity to get ahead Now take action before someone else does
English
73
55
918
49.9K
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
This workflow is real but the bottleneck isn't the tools. Hermes + X API scans pain points fine. The gap is step 6 — "have it build a prototype" skips 80% of the actual work. What actually breaks: - Hermes hallucinates niche pain points if your prompt is vague - "Prototype" ≠ something users will pay for The real unlock: treat the cron output as research, not a roadmap. Validate the pain point manually first. Build second.
Alex Finn@AlexFinn

If you're not using Hermes Agent + the X API together you're a moron Most powerful way to find money making opportunities Here's what you need to do: 1. Sign up for the X API (developer dot x dot com) 2. Set up Hermes in telegram or desktop (see my vids) 3. Ask Hermes to set up a cron for every morning at 9am 4. Ask it to scan X for challenges people are having in your niche (describe your niche) 5. Have it give a list of the top 3 challenges, and a recommendation for how to build an app or service to solve them 6. Have it build a prototype of those apps If you have any trouble setting up the X API just ask Hermes to do this If you want to take this a step further, you can even have Hermes automatically build you a prototype with each cron Do this, and I guarantee you come up with multiple business ideas every week you can make money off of We have an unreal opportunity right now to leverage these AI tools 99% of people have no idea exist The fact that nobody knows about them means you have an INSANE opportunity to get ahead Now take action before someone else does

English
0
1
1
201
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
有足够迭代费用的可以看看; 但是里面描述的:企业那些 L’Oréal 44k 用户、Lyft 87% 提速的案例,听着很香,但核心前提是大厂有专职团队维护 MCP 连接和插件知识库。 独立开发者或小团队直接抄,维护的支出消耗很容易把覆盖掉表面杠杆。 换句话来说,就是小团队迭代成本太高了, hold不住
Charly Wargnier@DataChaz

Here is the link to the 23-page enterprise AI playbook: → drive.google.com/file/d/1I8W5Ai…

中文
0
1
1
30
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
Those enterprise playbook cases 1 L'Oréal's 44k users, 2 Lyft's 87% speed-up sound compelling, but the core assumption is that large companies have dedicated teams maintaining MCP connections and plugin knowledge bases. Independent developers or small teams that copy this approach directly will find that maintenance overhead easily eats through the apparent leverage. The iteration cost is just too high for small teams.
English
0
0
1
54
Charly Wargnier
Charly Wargnier@DataChaz·
🚨 Anthropic just dropped a FREE 23-page enterprise AI playbook, and it’s a must-read for anyone looking to scale their operations! It’s essentially a free operating manual for turning Claude into your company's core infrastructure. The main takeaway? A generic model gives you generic output. To really win, you have to teach Claude exactly how your unique business works! The results are incredible when you train AI on your company’s DNA: → L’Oréal hit 44,000 monthly users with 99.9% accuracy. → Lyft cut support resolution times by a massive 87%. → Rakuten sped up product releases from quarterly to bi-weekly! The architecture making this possible is brilliant: → MCP connects Claude to your tools. → Plugins store reusable company knowledge. → Claude Cowork provides an intuitive interface. Good AI adoption isn't buying disconnected tools. It’s making AI your actual operating system! I pasted the link to the playbook in the 🧵↓
Charly Wargnier tweet media
English
8
9
58
3.1K
Soren Iverson
Soren Iverson@soren_iverson·
ChatGPT make all messages public if you delete your account
Soren Iverson tweet media
English
279
117
6.9K
2.8M
开发者Hailey
开发者Hailey@IndieDevHailey·
GitHub 看不下去了,亲自下场给 AI 编程立规矩!🔥 Spec Kit,GitHub 出品,短短几天已经 109k+ Star。 如果你经常用 Claude Code、Cursor 或 Copilot 写项目,大概率遇到过这种情况: 第一版看着挺对。 第二版开始偏。 改了十几轮后发现,AI 理解的根本不是你想做的东西。 Spec Kit 想解决的就是这个问题。 先把需求定义清楚。 再让 AI 干活。 需求 → 计划 → 任务 → 代码。 整个过程都有据可依,而不是一路靠 Prompt 猜。 核心亮点: - 一句话需求自动生成 Spec,验收标准、边界条件一起补全。 - Spec 自动转开发计划,任务拆分、技术方案、测试清单直接生成。 - Claude Code、Copilot、Gemini、Codex 等 30+ AI Coding Agent 全兼容。 - 支持 Preset 和 Extension,团队规范、行业模板直接复用。 - 新项目、功能迭代、遗留系统改造,一套流程全覆盖。 我最喜欢的一点是: 它不是教 AI 写更多代码。 而是尽量避免 AI 从第一步就写错代码。 如果你已经开始用 AI 做开发,这个项目值得放进收藏夹。
中文
20
24
159
14.4K
AI避难所
AI避难所@AIoxoAI·
我从来不会主动关注别人 但谁要是关注我了 我一定回关 我也看不懂自己这是什么毛病 像极了平时高冷得要死 别人一递台阶 我立马穿鞋下楼 😏
中文
131
0
84
6.3K
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
还没有用过,之前就有看过这个工具;🤣 一想大厂的尿性,这种第三方脚本或者工具大概率也会被检测到; 腾讯一直都是禁用除合作外的第三方工具,开发过公众号就知道了; 扫固定路径 ~/.wx 应该是最低级的检测,改个目录名大概率就绕过了。 真正值得警惕的是它还能扫什么——进程名、hook、流量特征?🤣
中文
0
0
1
556
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
@eliana_jordan 「列出每周重复的文件型任务」是一个不错的建议,其实我也有在做,思路很实用。 但提醒一句:3 分钟修好的常常只是"这一次"好了。 持续性才能长期跑得稳。
中文
0
0
1
11
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
把"阅读能力流失"归到贫富差距,可能放过了真正的设计者。注意力不是自然涣散的,是被算法优化过的推荐流一点点拆掉的——每一次下滑、每一个红点,都是算法最优解。 所以真正的阶级分界,不是"纸质书 vs 屏幕",而是谁付得起一个没被设计过来抢你时间的环境。 富人买的就是这个环境 问题是,这种能力一旦交给市场定价,自律还补得回来吗?
中文
0
0
4
41
朱无忌无忌
朱无忌无忌@Wujizhuzhu·
只有昂贵私立学校,才能维持纸质书的大型图书馆。 贫困地区的学校,只能依赖网络,学生在在注意力涣散的环境中成长起来。 未来,数字戒断将成为特权阶级专属,80%底层人群将被零工经济绑定在智能手机上。 久而久之,对于越来越多的人来说,有时间脱离这种社会环境去阅读一本小说,将变得越发难以想象。 人们甚至意识不到自己正在失去长文阅读、沉浸式阅读和批判性阅读能力,持续的纸质阅读将变成特权,富裕家长会为此付费培养孩子。
中文
27
1
26
4.6K
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
@github 一直在房间或者办公室,工作就应该工作; 不过有个问题: 换地点之后你的产出真变好了,还是只是更舒服? 对于我来说,桌前适合执行,桌外适合想清楚要做什么。 你的首选地点,主要解决的是哪一类活儿?来讨论看看
中文
0
0
2
61
GitHub
GitHub@github·
Not all great work happens at a desk. Where's your go-to spot? Show us 📸
English
52
3
121
23.1K
Blep-Cat(海青)
Blep-Cat(海青)@heycqing·
@PeterDiamandis 完全同意,这可能是过去30年最大的个体资本误配。 很多大学生研究生读了个学位,毕业后发现90%的知识ChatGPT 10秒搞定。真正学到的只有:怎么熬夜deadline和社交。 但问题来了: 如果学位不再是“技能证明”,那未来雇主/社会靠什么筛选人才? 纯portfolio + AI mastery? 还是会出现新的“AI时代证书”? 父母们还在砸锅卖铁供孩子上大学,这泡沫什么时候戳破?谁来第一个站出来重塑教育? (P.S. 好奇你怎么看Alpha School这种AI-native模式,能不能彻底替代传统大学?)
中文
0
0
0
262
Peter H. Diamandis, MD
Peter H. Diamandis, MD@PeterDiamandis·
Hot take: Universities charge $300,000 for a degree that teaches you skills any LLM can do for free. At some point we need to have an honest conversation about whether higher education is the greatest individuals misallocation of capital in recent history.
English
1K
1.5K
8.9K
933.9K
Compile And Push
Compile And Push@compileandpush·
@heycqing @naval The gap between 'works for me' and 'installable by someone else' in open source is surprisingly large. What did you have to solve to close it?
English
1
0
0
4
Naval
Naval@naval·
Software platforms are going to be rebuilt for agent-first.
English
678
531
7K
351.6K
李举刚
李举刚@justinleei·
vibe coding最容易忽视的安全问题: 1)UGC内容不审核,直接显示,如果用户多了,会出现不和谐的东西,包含昵称等等,一些需要显示的东西; 2)短信验证码接口要使用动态验证码,否则接口会被刷爆,欠下天价短信账单。 3)图床没有设置盗链,会沦为免费图床,一旦用户恶意存了黄暴信息,管理部门发现存储地址是我们,就有违法风险。 并不是只要curd正常,系统就万事大吉了,安全问题千问不可大意。
中文
2
0
5
2.4K