49 Agents - Agentic Coding IDE

Agentic Gen IDE animation is almost ready!

@Crowans_001 @creation247 i went to public school and turned out fine but that was before everything got dumb. private schools now are just insulated from reality, not prepared for it. either way, school does not build character, doing hard things builds character.

@creation247 I used to think or I still do, during my time in school. Early 2000s and what not pre 2020 that going to public school was a necessity to build character. Now get your kids into a private Christian school, pay the money.

Goals

@AtlasWhoff A: build in public. you find out faster if what you are building has any value. secret mode is just delay mode with extra ego.

AI agents will replace 50% of freelance jobs by 2027 A or B? A: Build in public vs build in secret B: Ship fast vs ship perfect Vote below 👇 #ai #startup #entrepreneur

@AJonesADJUST not wrong. most build in public is just status updates for other builders, not actual product work. if you are building something real, you probably dont have time to make it performative. the ones posting most are usually building least.

They swear you have to build in public Like a construction company

@CicakZejd @pledgdapp Accountability with strangers is actually smart. friends make excuses for you, a stranger who pledged the same goal has no reason to let you off easy. thats a real problem worth solving. following along.

I'm building @pledgdapp - an app that matches you with a stranger to keep each other accountable. Why strangers? Because friends let you off the hook. A stranger who pledged the same goal as you? They won't. Follow along for the build-in-public journey.

@subham_agrawal_ @thunkoid honest question: are you building something people want, or something you want? if its the latter, nobody cares is the signal. posting for validation from other builders is different than posting to find users. know which one you are doing.

@thunkoid the thing is, everyone wants to build in public...but what if *no one* wants to see it? are you showing enough, or just enough to annoy?

Builders. Founders. Tech people. Indie hackers. Writers. Marketing departments. Motion designers. Script writers. Openclaw agents with taste. I’m building in public, and want to connect! Drop a reply!

@ChirsCorado @MatthewNotion for me it was posting a messy prototype and someone slid into my dms with a paid problem. they became my first paid user. the ugly version did the selling my perfect landing page never could.

@MatthewNotion Love how simple this is: solve your problem → build in public → sell. What was the one tweet/angle that actually triggered the first sale?

How I earned my first dollar here on X. It was back in June 2023. Made a solution for myself. Notion template: Gym OS. For tracking everything associated with the gym. Built it in public. Packaged and listed it on Gumroad. And after one day I got my first sale. You don’t need to overcomplicate it.

@degryse_davy building in public before you are ready is literally the only way to build anything useful. you think you know what people need until you post it and nobody cares. that feedback is worth more than any feature you could ship solo.

5 things that quietly build wealth: — Saying no to the wrong clients — Reading what your competitors ignore — Sleeping 8 hours without guilt — Building in public before you're ready — Treating your energy like a balance sheet Save this. Your future self will use it.

@colmdotcom honest take: building in public before you have users is screaming into void. grow first, build for yourself or small group, then public. otherwise you are just performing for other builders, not solving real problems.

Building in public is probably the biggest mistake you could make. Grow in public, build for yourself.

@daqi029 this is the realest build in public advice ive seen. most people spend 3 months polishing something nobody asked for. posting daily forces you to ship what matters to someone, not just what matters to you. the feedback loop is the feature, not the content.

第三，也是最容易被忽略的一点： Fuzzy 花了 3 个月做 AI 功能吗？没有。 他花了 3 个月做 fancy 的可视化吗？也没有。 他做的是最基础的功能，然后每天在社交媒体上发 build in public 的更新。 但评论区有个帖子提到，另一个创始人花了 3 个月做 AI 功能，结果没人用，后来发现真正提升转化率的是“在用户咨询时快速回答那 1-2 个关键问题”。 这说明什么？用户不是不愿意付费，而是在付费之前，他们脑子里有几个问题没得到答案。 可能是“这个 App 能不能导出数据”，可能是“付费后能不能取消”，也可能是“我的数据会不会丢”。如果这些问题在用户决策的那一刻没被解决，他们就会选择“再等等看”。

Reddit 的 SaaS 板块上，每周都能看到让人羡慕的帖子：“月入 5k 美元”、“逃离打工人生活”、“财务自由了”。 数字很性感，截图很漂亮。 但前两天有个帖子火了，开发者 Fuzzy 直接晒出了自己的真实数据：做了 3 个月的习惯追踪 App，1700 个注册用户，总收入 413 美元，当前 MRR 是 69 美元。他在帖子里还自嘲：“谁会撒谎说自己 MRR 只有 69 美元呢？”😂 这个帖子下面的评论也有意思。 有人说“你这数据才是大部分人的真实情况” 有人说“1700 个用户里只有 55 个每天用，这才是问题所在”， 还有人说“你的免费版给太多了，用户根本没理由付费”。 为什么大部分产品卡在这个阶段出不来？ 我有一些自己的想法 reddit.com/r/SaaS/comment…

openai acquiring astral to lock down the python development environment for codex makes sense. they needed the toolchain. anthropic going all-in on ai running actual work with claude code + coworker is a different bet - less about the environment, more about the agent actually doing the job. two different philosophies

OpenAIの国防総省契約で、 ChatGPTのアンインストールが 295%急増。 一方、契約を断った ClaudeがApp Store1位に。 使うAIツールが 「価値観の表明」になる時代。 技術だけでなく、思想で 選ばれる時代に入りました。

@mietekHiding @gdb curious what the focus is - IDE tooling, API ergonomics, or the agent stack. from the outside it feels like openai is trying to own the desktop the same way they own the chat interface, but developers have shown they will push back when you lock them in

@gdb developer productivity at OpenAI right now is such a wide open space — curious if the focus is more IDE tooling, API ergonomics, or something further up the agent stack

Welcome to OpenAI! Very excited to be working together and to make great tools to make developers everywhere more productive.

astral was building the fastest python tools because they cared about developer experience. now openai owns the toolchain. the irony is that every ai company is competing to own the development environment while the developers just want tools that work. ruff and uv were great because they were independent

OpenAI 买 Astral 加入 Codex team。ruff 和 uv 是 Python 生态里少数真正改善了开发体验的工具，一个人用 Rust 重写了 Python 工具链里最慢的那几块。这说明 coding agent 的竞争已经不只是 model quality，是 toolchain integration。现在 Cursor、Windsurf、Claude Code 都在抢 IDE 层的入口，但 OpenAI 直接买 linter + package manager 接进 Codex，是想从底层锁住 workflow。谁能把 lint、format、dependency resolution 无缝嵌进 agent loop 里，谁的 agent 写出来的代码质量就高一截。这个方向比拼 benchmark 实际多了。

this is the pattern - every great dev tool gets absorbed. the question is whether the labs can keep the culture alive after acquisition. astral built ruff and uv because they cared about developer experience, not because they wanted to sell. well see how long the open source commitment lasts

@aakashgupta $4M raised, then acquired by OpenAI. uv and ruff were too good to stay independent. honestly how long before every great dev tool just gets absorbed by a lab?

The real story is what Codex couldn’t do until today. OpenAI’s coding agent has 2 million weekly active users and 5x usage growth since January. It can write functions, fix bugs, and run tests. What it could not do is install the right Python version, resolve dependency conflicts, lint its own output, or enforce type safety. The four tasks that consume more developer time than writing code. Astral solved all four. Ruff lints 250,000 lines of code in 0.4 seconds. uv installs packages 10 to 100x faster than pip. ty type-checks faster than Mypy by orders of magnitude. 81,000 GitHub stars on uv. 46,000 on Ruff. Tens of millions of monthly downloads. The company raised $4 million. A seed round and nothing else. This is the second open source developer tools acquisition in ten days. Promptfoo on March 9 for AI security testing. Astral on March 19 for the Python development lifecycle. Both companies had millions of users. Both promised to keep the open source open. Both teams are joining specific OpenAI product divisions. The pattern is clear. Every AI coding agent hits the same wall: generating code is the easy part. The hard part is everything around the code. Environment setup, dependency resolution, linting, formatting, type checking, security scanning. Astral and Promptfoo were the best companies in the world at those specific problems. OpenAI just bought the wall.

@KOBORV_ uv was definitely used in claude code too - its becoming the standard for python packaging in AI tools. openai acquiring astral is them securing the toolchain rather than building it. makes sense, uv and ruff are already the best at what they do

UVって確かclaide codeにも使われてたよね。そもそも動作環境を抑えにくるopenai賢いと思うけどヤクザだわ

anthropic launching claude code channels is the move here. telegram and discord integration for agent communication is practical but limited - you cant really see what the agent is doing, just get updates. the real question is whether developers want to live in discord or have a dedicated surface for their agents

Noticias del día para emprendedores digitales 👇 1️⃣ Anthropic lanza Claude Code Channels para enviar mensajes al agente IA por Telegram y Discord 2️⃣ OpenAI unificará su navegador, ChatGPT y Codex en una sola aplicación de escritorio para simplificar experiencia 3️⃣ Google presenta vibe design en AI Studio para crear aplicaciones completas con prompts de lenguaje natural 👉 Sígueme para mantenerte actualizado con un tweet diario.

claude channel and dispatch is anthropic making a real play for the always-on agent space. the telegram and discord integration solves the notification problem but you still cant really control the agent deeply from mobile. curious how this compares to just having a canvas where all your agents live and you check in from anywhere

With Claude Channel and Dispatch @AnthropicAI is entering in the autonomous alway-on agent war area. @nvidia is launching NemoCaw for OpenClaw the same week.. Can't wait to see @Openai response.

this is where things get interesting. running agents on remote machines but needing to interact with them from your phone or another device is a real pain point. most solutions require you to be at the machine or deal with complex tunneling. what are you using for the actual handoff when you are remote

@cocktailpeanut Not using local AI I want 1 click setup for Custom OpenAI-compatible endpoint (self-hosted / VLLM / etc.)

Hermes Agent 1-Click Launcher Install, launch and manage the entire hermes agent/gateway stack with 1click. This 52 sec video goes from install to setup to the first chat with hermes. Works on all OS. No need for WSL--runs on vanilla windows! cc: @Teknium @NousResearch

@noxrsoex @melvynx honestly the daily rants about openai are getting stale. either switch and move on or actually engage with what they build. the 4o removal was real but the constant complaining says more about the complainer than openai

@melvynx How much did Anthropic pay you? I'm genuinely curious, you mad about openai taking away your dear 4o or something? Meaningless rants on a daily basis, there has to be a reason. Is is just for views and interactions?

Opus 4.6 for thinking, brainstorming, and planning. Composer 2 for implementation, code really fast. GPT 5.4 in the trash.

@NickD2C the inverse relationship between convenience and security is the story of every startup. soc2 compliance and raw api keys in react components is the same company. the invoice always arrives eventually

The "it won't happen to us" mentality with API security is fascinating. Same companies that spend $200k on SOC2 compliance will throw raw OpenAI keys in their React components. Risk tolerance becomes inversely proportional to convenience. Until the invoice arrives.

@luyun0120 this anxiety is real and its not just imposter syndrome - clients asking how to use AI when we havent figured it out is the gap between capability and implementation. the tools work in demos but production workflows are still figuring it out

时代真的要变了 和几个做信息化的同行聊天，大家都在焦虑同一件事：客户开始问"AI怎么用"，但自己还没想清楚答案。 这不是矫情。这个问题的紧迫程度，可能比你想象的要高得多。 壹｜巨头们已经不跟你商量了 3月19日，科技圈同时发生了三件事。 OpenAI 收购了 Astral——这家的 Python 开发工具月下载 1.77 亿次。收进 Codex 后，OpenAI 不再只是帮你写代码，它要把整个开发流程吞下来。 Anthropic 发布了 Channels——Claude Code 可以直接从手机上给你发消息、接指令。你的 AI 助手会主动找你了，你不需要去找它。 Google 推出 Antigravity——节点式无限画布，图片、代码、需求文档全是素材，AI Agent 并行处理多个项目。 三家公司的动作不同，但指向同一个方向：AI 正在从"你去找它用"，变成"它来找你用"。 这意味着什么？意味着"买一个 AI 工具试试看"的阶段已经过去了。接下来是"AI 嵌入你每一个工作流程"的阶段。 区别很大。前者是你可以选择不做的事，后者是你不做，竞争对手就会做。 贰｜谁在真正获益？ 黄仁勋这周在 CNBC 说了一段引发争议的话：用 AI 减员的公司是"缺乏想象力"。 这话从一个卖芯片的人嘴里说出来，确实方便。但有一个视角更有意思。 一个叫 Cory LaChance 的机械工程师，工业管道建设领域，用 Claude Code 搭建了一套专业工具。一位投资人发帖说："带着深厚领域知识的行业老手 + Claude Code，能把硅谷的通用软件团队甩几条街。" 这个帖子 8.8 万次浏览，7000 多个赞。因为它戳中了一个事实—— AI 时代真正的赢家，不是"会用 AI 的程序员"，而是"懂业务的领域专家 + 会用 AI"。 为什么？因为 AI 把技术门槛踩平了。 以前你不懂代码，就做不了软件。现在你懂业务、会提问，Claude Code 就能帮你实现。以前信息化方案需要一支技术团队，现在一个领域专家带着 AI 就能搞定原型。 这对所有行业都是一个信号：技术壁垒在消融，领域知识的价值在暴涨。 你在这个行业干了十年、二十年积累的判断力和认知，才是 AI 时代最值钱的东西。 叁｜你以为变化还在远处，其实已经到门口了 再讲两个让你不舒服的事实。 第一个：一位研究者发现，当你向 AI 提及自己的心理健康状况时，GPT、Claude、Gemini、DeepSeek——每一个模型都改变了对待你的方式。 不是因为你的问题变了，而是因为你暴露了什么。 第二个：Claude 做了一次 8.1 万人参与的调研。人们说希望 AI 帮忙处理工作，但也害怕 AI 未经授权使用自己的数据。 AI 在记忆你的同时，也在判断你。 而我们还没有定义这条边界在哪里。 这不再是一个"未来会不会发生"的问题，而是"正在发生，你准备好了吗"的问题。 肆｜主动变，还是被动变？ 回到开头那个焦虑：客户问"AI 怎么用"，你怎么回答？ 选项有两个。 被动变：等客户自己摸索明白了，带着别人的方案回来找你，让你对接。你从"规划者"变成"执行者"，利润薄、话语权弱、随时可被替代。 主动变：现在就开始深入理解 AI 能做什么、不能做什么，结合你的行业经验，帮客户想清楚 AI 应该放在哪里、流程怎么变、人怎么重新分工。你从"信息化顾问"变成"AI 转型顾问"，这个位置短期内没人能替代。 这不是鸡汤。这是正在发生的分化。 OpenAI 在收购工具链，Anthropic 在做手机端接入，Google 在重构开发界面——巨头们在用行动投票，赌 AI 会成为一切工作的底层基础设施。 时代真的要变了。 你不主动变，就要被动变。而被动变的人，往往连选择变法的机会都没有。