← 精选
Nan Yu
算力过剩的奢侈。
第一次听说这么神仙的 App，FindMy 团队真该好好学学。
当然是英文
从圣克鲁斯往北，整个美国西海岸。旧金山、马林、西雅图、波特兰……
要是我们盖的房子真能让人们住得起就好了。
走在最前沿的问题在于，随着行业不断摸索，你必须频繁调整工作方式。
最佳切入点在于稍晚几步：即那些相对较新、但已度过动荡期的技术。
信不信由你，这其实只是换种说法而已。
想想在 AI 时代之前，你在 Behance 和 Dribbble 上见过多少“天气”应用和“仪表盘”。
全是幻觉。
设计的核心在于理解，而非产出。
设计是一种意图，而非图像或原型。缺乏意图的产出，只是幻觉。
大教堂。随处可见。
Venmo到底什么时候才能关掉那个能让我看到共同好友是不是在约炮的功能？
如果各大实验室都发布新闻稿，你觉得哪家的公开言论与其实际想法的差距最小？
我一直也有这种感觉，但直到读到这段话，我才终于能把它说清楚。
这就像在问“我该买红色的兰博基尼还是绿色的？”，却预设了正确答案应该是“你不该买兰博基尼，因为这在财务上是不明智的”。
如果人们觉得你不擅长公关，反而会相信你没在撒谎。仔细想想，这其实说明你公关做得很好。
就像加密行业重塑了所有现有的金融概念，只不过这次速度快得多。恭喜我们。
当然还有工程师
我们正从第一性原理出发，重塑设计师（乐观地）
今天和一位企业客户聊了聊，对方说他们 Linear 的头号重度用户几乎完全通过 MS Teams 来操作。对于这类用户来说，聊天界面绝对是与所有软件交互的最佳方式。
这个说法很到位
这很合理。 https://t.co/OUQYrhld4w
品味
从这张图中能学到的道理，已与五年前大不相同。
提醒一下：复杂性并非护城河。
它护不住你的业务，因为更简单的对手会跑得更快并赢得胜利；它也挡不住 AI，因为 AI 的进化终将让这种复杂性变得毫无意义。
每当有人喷 Google 或 Meta 时，就得提醒他们，这些公司的规模大得惊人。由于其业务基数巨大，哪怕是些许失误，看起来也会显得规模宏大。就像大多数生意一样：看赢了多少，而不是输了多少。
还有正则表达式。
现在要是能拿下 skillshare 这个域名，简直绝了。
很多设计师和工程师在抽象思考产品问题时表现很出色。但你得收走他们的 IDE 或 Figma，否则他们会直接扎进编码或设计里，只顾着完成既定需求。设计师应该多去当产品经理，他们会很擅长。
翻转时刻
死神之眼
主要是因为他们很擅长。
自己去看吧。
我觉得应该有更多设计师和工程师考虑转行做 PM。
这里有一个试金石。我不再用这个问题面试了，因为 GPT-4 之后的大模型给出的答案，比 90% 的 PM 候选人都要好：
“用户经常要求在 Linear 的 Issue 评论中增加‘定时发送’功能，主要是担心在周末打扰同事。我们应该做这个功能吗？”
思考并写下你的答案，然后去问问 Claude 或 ChatGPT。如果你觉得自己的答案比它更好或不相上下，那你就是前 10% 的顶级 PM。
这才是加州高铁真正的目的：变相将洛杉矶和旧金山融为一座城市。
https://x.com/EedenEnne/status/2040651790295658510?s=20
今天在纽约
让我想起 iOS 的天气应用，设计者们一年有 300 天都待在 21 摄氏度的气候里。
依然如此
虽不想泼冷水，但你的品味可能并不比 AI 高。
如果你给顶尖模型一个指令来做产品决策：“别让我去问用户，你自己推理一下。”它给出的答案会比 90% 的产品经理都要好。
很多人不敢尝试，是因为他们担心答案真的会非常出色。人类确实还有很多独特的价值，但“品味更好”显然不属于其中之一。
I've always dreaded these big bang quarterly and annual planning exercises. The larger an organization gets, the more performative and less anchored to reality it becomes.
And the worst thing is that productive work tends to just stop during this period.
The problem with most existing AI solutions is they expect you to micromanage them. This creates tons of new mental burden and overhead and it's a big reason why AI adoption efforts fail.
We're taking a different approach at Linear, which we're sharing today. Our goal is to
I've really been enjoying this workflow lately. Whenever a PM or a designer finds a small change to be made, we just create the issue and assign it to a coding agent.
It would have been much more annoying for us to put it into the work queue, wait for an engineer to take time ht
Talent is still the bottleneck —
Even for a startup with a great reputation like Linear, competing for top talent with all of the mega cap tech companies is *really* hard.
One of the things that AI has been able to do for us is to take all of the ticky-tack stuff off of the
This quote from Teller (of P&T) lives in my head rent-free:
"Sometimes, magic is just someone spending more time on something than anyone else might reasonably expect."
Do people understand that 90% of the time a faster horse is the right answer?
Seems like a good time to bring up some previous thoughts I’ve had about remote work.
I love being able to work with colleagues from across the country and the world, without any of us needing to uproot our lives.
Our 3 co-founders Iive in California, New York, and Finland.
I’ve noticed the liberal arts aspect of my education playing an increasingly larger role in my career as time goes on and AI tools have supercharged this.
They are steerable enough at this point that you can be specific and expressive.
Now you have to know what you want!
https://x.com/IamJoshJackson/status/1966262228736442720
Something that we all need to recognize is that engineers end up making 50%+ of the product decisions by volume.
There's no practical way PMs and designers can specify every single edge case, all the subtleties of the loo
If you've ever thought "I wish issue trackers had this feature" just come and build it. This is software you've used your entire career
If you are an engineer, and you are pretty sure you're a better product person than most or all of the PMs you've worked with, we would like to hire you at Linear.
Apply on the website and mention this tweet and we will make sure we get to your application quickly
People often come to me for advice about product management, and the most consistent thing I say to them is: "You should probably be 10x more specific than you're being right now."
Who is your customer? "IT managers"
No! It's Bob Smith who leads IT at Figma
What's the benefit
Interview design is becoming an interesting game.
It's basically assumed you're going to be using AI to generate answers, so we're actually selecting for your judgement to select the best answers and arguments, and your ability to identify tradeoffs and practical