Ming

133 posts

Ming

Ming

@mingyu8h

Independent Developer | Building openteams, an open-source collaboration platform for AI agents. https://t.co/VKR2IRtVO2

Katılım Şubat 2025
353 Takip Edilen98 Takipçiler
Sabitlenmiş Tweet
Ming
Ming@mingyu8h·
我认为要看使用场景: 1. 协作场景下要用AI团队必须是类似于claude teams一样的架构,每个Agent负责一个任务,做好上下文隔离,只做有限的消息传递,能够实现并行提效和任务自动化; 2. 开放场景下可以让不同Agent扮演不同的角色,用不同的视角看待同一个问题,这个我们证明了在学术研究上很有价值。 所以关键不是把它们定义成什么角色,而是赋予它们相应的工具和技能以及明确其职责。 我做了openteams很好地适配了上面两种常见,分别是工作流模式和自由聊天模式,欢迎体验: github.com/openteams-lab/…
中文
0
2
9
1.7K
Ming
Ming@mingyu8h·
@wangray 分享分享经验吧大佬
中文
1
0
1
31
Ray Wang
Ray Wang@wangray·
回头望 4个多月前我还在庆祝自己有了500个蓝V关注 当时的目标是3个月内到500万流量 谁能想4个月后我已经有超过4000万的流量了呢? 所以目标还是得定的高一点,万一实现了呢?
Ray Wang tweet media
Ray Wang@wangray

Two weeks in on 𝕏 and I just hit a small milestone.🥳 500 verified followers. Seriously, thank you for being here and for the support. I really believe we can make this place better when creators show up consistently. Next goal: 5M+ impressions in the next 3 months. Let’s keep building and enjoy the ride together.

中文
44
0
28
3.6K
Ming
Ming@mingyu8h·
When knowledge itself is no longer valuable, who should be producing it?
English
3
0
1
19
Rayane
Rayane@FlippedRay·
Time to promote your startup Drop your project URL Let’s drive some traffic
English
299
7
131
13.1K
Ming
Ming@mingyu8h·
Kiro Code’s spec mode really didn’t work for me. Gave it a fairly complex task, it ran for 30 mins, burned 300 credits, backend wouldn’t compile, frontend opened to a blank screen. If it weren’t free credits, I wouldn’t bother. openteams feels much better: clear workflow, fresh context per step, no context pollution, and agent review built in.
Ming tweet mediaMing tweet mediaMing tweet mediaMing tweet media
English
3
0
1
116
Ming
Ming@mingyu8h·
ok this is a good one — new kiro users get a month free right now. opus 4.8, ~$100 of usage. steps: 1. fresh account, credit card ready 2. pick the $20 plan at checkout — auto-discounts to $0 3. flip to Free immediately so next month isn't billed
Ming tweet mediaMing tweet media
English
5
0
4
664
Ming retweetledi
Aron厚玉
Aron厚玉@aronhouyu·
longflow这个项目 安装到codex里!!!!! 你就可以爽了!!! longflow = 可视化通用工作流
Aron厚玉 tweet media
中文
23
46
314
38.5K
Ming
Ming@mingyu8h·
Claude实现完UI以后,再让gemini评价一下,综合两种模式在UI上的优势,最后效果看着还可以。
Ming tweet mediaMing tweet media
中文
0
0
1
41
GrowthExtra
GrowthExtra@GrowthExtra·
Bro to Bro: Let’s grow our X accounts together. Just comment “HELLO” below and start connecting with everyone.🙋‍♂️🔔🔔🔔
GrowthExtra tweet media
English
89
33
66
4.8K
Ming
Ming@mingyu8h·
@AngryPayload I’m in China, but this doesn’t seem to have any regional restrictions.
English
1
0
0
160
Ming
Ming@mingyu8h·
MiniMax-M3 is probably the most cost-effective model right now. Based on my tests, its coding ability is basically on par with GLM-5. In openteams, you can let M3 implement the code first, then have GPT-5.5 review it. Overall token usage is about 30% lower than using GPT-5.5 directly, while the code quality is basically comparable to GPT-5.5’s own implementation.
Ming tweet mediaMing tweet media
English
1
0
1
126
mia
mia@miiamiav·
eğerr takipçi kazanmak istiyorsan bu gönderiyi beğen ve bir nokta bırak
Türkçe
536
20
801
55.6K
七喜 | 7UP
七喜 | 7UP@bitcoin136·
上个蓝V互关帖子涨了1000+粉丝 再来最后一次蓝V互关 冲击2W粉 还是老样子 我的内容你可能不喜欢 我这个人你可能不喜欢 但是我能为你的XHunt KOL关注数量 +1 仅此而已 留言必回关 没回的直接私信我
中文
566
2
200
30.4K
Ming
Ming@mingyu8h·
have you used multi-agent setups with codex? would love to hear what problems you've run into — maybe i can help.
English
0
0
2
48
Ming
Ming@mingyu8h·
@Foresight_OS 大家蓝V互关起来,报团才是起号的答案
中文
10
0
3
89
@Foresight远见
@Foresight远见@Foresight_OS·
如果你也在做X,应该懂那种发出去没人理的感觉🥲 所以蓝V朋友可以互相关注一下,大家先把第一批真实互动做起来。内容合适我都会看,也会认真回。
中文
53
0
32
1.9K
Ming retweetledi
Ian (伊恩)
Ian (伊恩)@ianneo_ai·
看到好些小伙伴问这个 skill 能不能分享,来啦,我直接丢 GitHub 开源了!!! 它不是插画 prompt,也不是 PPT 模板,更像一个专门帮中文长文找画面的人: 先读文章,抓里面那个真正值得画出来的判断、流程、卡点、隐喻,再生成 16:9 白底配图 这次我把小黑 IP 拆得很细,放在 references 里,你可以直接改成自己的角色:小猫、小狗、机器人、品牌吉祥物都行 别只换皮,关键是让它参与核心动作,掉坑、递材料、过桥、接住用户,而不是站旁边卖萌 我自己用 Codex 跑了,出来就挺能用,红橙蓝批注、流程线、留白都稳 现在写长文舒服很多,以前抽象概念只能硬塞截图,现在终于能画成人话了 适合写文章、知识卡片、Notion 文档、方法论内容的人 GitHub 自取,喜欢点个小⭐️ 链接放评论区👇
Ian (伊恩) tweet mediaIan (伊恩) tweet mediaIan (伊恩) tweet media
Ian (伊恩)@ianneo_ai

现在写长文真的离不开这种小破图了。小黑、红蓝手写字、几根箭头,完全不精致,但特别有脑子在现场转的感觉。比那种 AI 生成的高级封面耐看多了

中文
14
24
142
11.9K
Ming retweetledi
Ian (伊恩)
Ian (伊恩)@ianneo_ai·
继续,这次我准备更新极简高级模式啦!!! 配图这件事拼的不是画得多牛逼,而是人愿不愿意停一下 小黑好用就好用在,它不像 AI 封面那种一眼在喊我很高级,它更像你在白板旁边随手画给朋友看的草图,有人性,有停顿,有哦我懂了的那一下 这次更新了极简高级模式,留白更狠,线条更少,信息密度反而更高。黑色小人负责情绪,红橙蓝手写字负责判断,几根箭头把流程讲完 长文配图最怕装,越装越像课件。真正能传播的,是让人感觉作者脑子在现场转 小黑就是这个味儿 大家更喜欢前面那种常规正文模式,还是这次这个极简高级模式?
Ian (伊恩) tweet media
中文
11
4
38
7K
Ming
Ming@mingyu8h·
@romanbuildsaas I feel like the advice above only works if you've already pinpointed a real user need.
English
0
0
0
33
Romàn
Romàn@romanbuildsaas·
I'm 30. I built an AI startup called GojiberryAI to $2.5M ARR. Got accepted into YC. If I had to start from 0, here's exactly what I'd do: 1. Sell it before I build it. No code. Just a simple slide deck (mine was 6 ugly slides) explaining the problem, the solution, the outcome, and the price. I made my first $10k that way, before writing a single line. 2. Pick a painfully specific customer. Not "B2B SaaS." Something like "founders at 20-person SaaS companies about to hire their first SDR." So specific that the right person reads it and thinks "that's me." 3. Start outbound on day one, but only to people showing intent. Not scraped lists. People engaging with competitors, changing roles, raising money, or publicly posting about the exact problem I solve. That's the gap between a 1-2% reply rate and 25-40%. 4. Lead with value, never a calendar link. Send a blueprint, not "got 15 minutes?" Let the resource do the selling, and the trial becomes the obvious next step instead of a pitch. 5. Pick ONE channel and go deep. For us it was outbound first, then Reddit (10M+ organic views), then LinkedIn lead magnets. I wouldn't touch a second channel until the first one was clearly working. 6. Talk to customers every single day. The product doesn't matter until you understand the problem better than they do. Spend 90% of every early call listening, not demoing. 7. Only build once people are actually paying. Then keep it dead simple and price it to sell itself. We landed on $99/mo with a free trial, so the funnel runs without me dragging anyone onto a call. 8. Do this relentlessly for about 12 months. That’s roughly how long $0 to $2.5M took us. Bootstrapped. No outside funding. Most founders don’t lose because they can’t build. They lose because they build too early, sell too late, and quit the channel before it compounds.
Romàn tweet media
English
124
87
1.4K
96.5K