Cypher

最近在学习agent的同学都可以读一下

@PandaTalk8 和MySQL最大的不同是啥

postgresql 之前了解的不多， 感觉这个数据库可以玩性很强啊 ， 找本书了解一下。

@HiTw93 明白了，大概是做填空题的意思，后面有打算深入研究一下agent的评测相关内容吗？

@Lin142535Lin 前提是脑袋里面先有一个框架和提纲，然后去找到最新都一些文章下然后简单读读是否合适，合适然后我会让ai帮我翻译成中文，我再看，然后最终决定哪些是我的素材，同时结合自己真实场景来结合，没有ai前写分享也是这个习惯

@ShaoXuan9520 慢慢看哈哈哈

@Lin142535Lin 太长了，没有看完

非常深刻的体验，出海是值得的

今年应该继续待在大厂还得去初创呢？真是个纠结的问题

@lifesinger 如何让用户愿意把数据放到你这里，可能是最关键的

For agents 是非常廉价的创业方向 小心被骗进去 因为强大的 agent 自己会写工具 不需要你去写 如果你去写了 效果很好 收入很多 agents 背后的人类就会 转手自己去实现了 For agents 的 skills 等服务的真机会 是稀缺性 如果没有稀缺性 迟早都会被大流量的 agents 或 LLM 吃掉 LLM 和有大流量的 agents 可能就是那几家了 For agents 没有大机会 但有大量中长尾的创业机会 以上是逻辑推演 很无聊 期待真实世界 可以不按照逻辑往前走 这样才刺激好玩

@Jarodxu7 哈哈哈，昨晚还在问，今天就刷到

欢迎加入 happycapy！

MCP fans won't like this... ok here's my mcp vs cli take after months of building agents the good stuff about MCP: standardized way to connect AI to tools great for IDE integrations (cursor, claude desktop) type safety, schema validation, all that jazz the reality check: schema loading eats your context window alive debugging json-rpc errors is no fun for tools the model already knows (git, curl, npm)... why bother? what i actually do: # instead of mcp server agent.run("curl api.github.com/user") # mcp approach agent.use_mcp_server("github", "get_user") # ...wait why is this returning nothing again maybe MCP is great for enterprise setups with fancy requirements but for me? my agent runs commands. cli works. simple. fight me ✌️ #ai #llm #mcp #programming #indiedev

@its_unks I spend a lot of time on Twitter, and I follow many people in the AI space, which has led to this.

@Lin142535Lin How do you keep these updated?

So I found this tool called gstack and it's kind of insane It basically turns Claude Code into a whole team of specialists. Instead of one AI helper, you get 13 different "modes" you can summon with slash commands. Need to think like a CEO about your product? `/plan-ceo-review` Want someone to tear apart your code for bugs? `/review` Need to run QA and actually fix the bugs? `/qa` Just want the AI to browse your app and take screenshots? `/browse` It's like having an entire dev team in your Claude Code session. Each command switches the AI into a different mindset — founder, engineering manager, designer, QA engineer... Setup is super simple: `git clone github.com/garrytan/gstac… ~/.claude/skills/gstack && ./setup` If you use Claude Code daily, this is legit a game changer. No more asking "can you review this like a senior engineer?" — just type `/review` and it happens. Been using it for a week now and my workflow is completely different. #AI #DevTools #ClaudeCode

@YashPnke Totally agree — too many slash commands shift users from doing to thinking, which kills conversions. Treat them as an accelerator, not the interface: keep a strong default path, limit options, and surface them contextually.

@Lin142535Lin slash commands are cool, but make sure it doesn't create the 'expert paralysis' effect. too many options kill conversions.

@PandaTalk8 哈哈哈，最近正在看初创公司，打算离开大厂了

对大型组织比如超100人以上的团队，在AI日新月异的今天，他们所面临的挑战要远远超过小于10人的小团队。 小团队+ AI 的杠杆效应，在面对科技、市场趋势剧变化的时候会更加快速的、灵活的作出反应。 在我看来这是小时作为单位时间来衡量的。 而大型组织管理、内部协调、信息传递的效率上无处不在的进行能量损耗、信息传递的失真。 比如高层决策缺少足够真实的信息、中层所处位置天然的向上讨好、基层会因为中层的压榨会导致会高度内卷，平白的消耗掉更多的时间和能量。 大型组织最终会成为他们的「创新者的窘境」。 所以，现在的我，宁愿选择一个小型的创业团队，也不想去大型组织工作。

That makes a lot of sense—you’re essentially running a hybrid model of explicit routing + context-driven orchestration. When you need precision, you directly call specific agents (like /impeccable, plan ceo, plan eng). When you need exploration or augmentation, you defer to the dashboard to decide based on context.

@Lin142535Lin @YashPnke Great call out. I do both depending on project. If I’m designing a mood board I need call out /impeccable with gstack. Then I say after this search dashboard for example. If you know you need specifics call them out and say run these with plan ceo to plan and plan eng to sort

That’s a really clean setup. I like the idea of using a centralized dashboard as the control layer—basically turning CC into an orchestrator rather than just a single agent. Do you find it more effective to let it auto-select from the Gstack based on the dashboard context, or do you still sometimes explicitly route tasks to specific agents? I’m thinking about how to balance autonomy vs. control, especially as the number of skills/hooks grows.

@YashPnke @Lin142535Lin You can tell CCto run the Gstack as a team of agents or say which ones specifically with others depending on what you’re building. No /commands needed. I had CC build dashboard of all skills,hooks,/com etc. I have it check dashboard run gstack or check what in dashboard can help

@watreejane 太强了，学到了

@JamesAI 居然有64GB的，我还以为48GB的就到头了

下周开始 vibe coding 走起

@Leobai825 本00后表示，既知道x，也发了x，求个关注不过分吧

昨天回到青年旅舍 跟00后聊天发现一件挺割裂的事 很多人甚至不知道 X 是什么 也不知道世界正在发生什么变化 能刷到 X，本身就已经超过了 95% 的人 这个世界远比我想象得大 也远比我想象得不公平

@lxfater 马上就发，求关注求转发求点赞

做 X 只有一个秘籍就是多发内容 假如你不会写文案，多发，过一段时间就会了 假如你不会选题，对方，过一段时间就能看到规律 或者，单纯是因为你倒霉😅，不能火 多发一段时间，总有被你撞大运的时候 那些 4k 粉的做增长大师们只不过说一些刺激你分泌多巴胺的话，抚慰你的心灵而已 真正懂的人都知道，多发。 其他都是噪音，精神春药。