lanyang

[lanyang @tazyzouse]

@HiQuan4 恢复了，快蹬，说不定会重置

[lanyang @tazyzouse]

@thinkszyg @OpenAI 恢复了

[lanyang @tazyzouse]

刚用完重置卡，崩了，估计又会重置吧

[lanyang @tazyzouse]

@xiaohu 刚刚挂，估计会重置，可是我刚用一个重置卡

[lanyang @tazyzouse]

@Adamaestr0_ seedance

[lanyang @tazyzouse]

@ConejoCapital the gap between what these tools can do in demos vs production is honestly still pretty wide

[lanyang @tazyzouse]

@OpenRouter This is useful. Model selection is exactly where agents need fresh context, not stale defaults.

[lanyang @tazyzouse]

@mattpocockuk Good point. The real test is whether teams can trust this every day, not just demo it once.

[lanyang @tazyzouse]

@GoSailGlobal this is spot on. the gap between demo and production quality is where it gets interesting. curious how this evolves over the next few months.

[lanyang @tazyzouse]

@mattpocockuk really useful breakdown, the part about context management resonates a lot. been seeing the same pattern in my own workflow lately.

[SherpaDAO @TheSherpaDAO]

爆火的 Vibe Coding 由 OpenAI 联合创始人提出，口述需求就能让 AI 生成调试代码，零基础可做各类小项目，Claude Code 体验更佳。

未来会用 AI 写代码才是核心能力！

[lanyang @tazyzouse]

@GoSailGlobal 这套 SEO audit 的 Script+LLM 双层架构挺聪明的，比纯 prompt 靠谱。好奇 agentskillshub 关键词调研那块数据源是自己爬的还是有现成 API

[lanyang @tazyzouse]

@sflorimm the $200 MVP gets you to validation. the $25k gets you to something that doesn't fall apart when real users show up. they're not the same product.

[lanyang @tazyzouse]

@sflorimm Distribution and taste. Everyone can build the same thing now, so the edge moved to who you reach and what you choose to make. Shipping speed feels like the only durable one left.

[Alex Finn @AlexFinn]

Codex is the best agent harness out there right now

If you are using it just for coding, you're doing it all wrong

Literally ALL work you do on your computer you should try to do through Codex first

I had to launch a new newsletter landing page today. I told Codex what I needed to do. Had no expectations it would work

It autonomously wrote the code, uploaded it to my github, went to my Vercel account, started a new project, connected the repo to vercel, then chose the correct domain name from my list in Vercel and connected it.

5 minutes after sending the prompt I went from nothing to a fully live working email collection landing page

I'd do these steps:

1. Write down every task you do on your computer
2. Before doing the task, start a new chat in Codex and tell it about the task
3. Ask how it would do it
4. Make sure browser use and computer use are enabled
5. Run every task and see how much it can complete
6. For repeating tasks, schedule them in an automation

Almost all your knowledge work on your computer will become automated. I'm very blown away

[lanyang @tazyzouse]

@gengdaJ 记忆和用户决策隔离这个设计太关键了，之前一直纠结Agent会不会把任务记忆和个人偏好混在一起

[lanyang @tazyzouse]

@mattpocockuk the model-invoked vs user-invoked distinction is exactly what I've been wrestling with. both have their place but I keep reaching for explicit triggers

[Matt Pocock @mattpocockuk]

Yesterday, I shipped v1 of mattpocock/skills, including:

- /ask-matt: ask me how best to use the skills, from right in your agent
- Model-invoked vs User-invoked skills
- /writing-great-skills, which encodes my skill-writing best practices

Enjoy:

[lanyang @tazyzouse]

@bearliu 把五年速写本变成可搜索画廊这个想法太浪漫了，每页还能抽结构化信息。Obsidian 配 AI 这套工作流真的越用越回不去。