关注科技, 投资, 和个人成长.
Calgary, Alberta
Joined October 2021
- Tweets4.6K
- Following582
- Followers43.1K
- Likes5.7K
Are there anyone like me who gets excited whenever hear the fan spins up on your Mac mini when vibe coding 😂 I get annoyed when it’s on my MacBook Pro but weirdly satisfying on Mac minis.
Kylon to me is like Grok Bot with Opus 5.5 without limit: I bring my own Claude Ultra and Codex sub into Kylon and it runs like Grok Bot but better. That’s my secret for the past several months. And that’s why I don’t really find Grok Bot too attractive to me. Plus you get real team feature in Kylon. All my collaborators and employees are on Kylon. Kylon is like Slack + Claude Tag + Grok Bot
Grok Bot without limits?
Elon says it costs too much to run. So in Kylon, Grok Bot plans the project, Claude Code builds it, Codex reviews it.
Three agents, one shared workspace, each doing what it does best.
xitter.gq/i/status/2116992258520…
一口新饭 retweeted
“There is no moat.”
But moats are everywhere:
Ambition.
Taste.
Agency.
Curiosity.
Judgment.
Courage.
Obsession.
Persistence.
Integrity.
Patience.
This era belongs to those who refuse to be mediocre.
This is definitely the correct direction for harnesses because software needs happen on the interface of user and AI. And that interface is the harness. Basically, a harness plugin is a customized AI native software created by and for the user. For example, you don’t need a Notion/a video editor with MCP. What you really need is a Notion/video editor attached natively to your harness (deepseek harness, Claude code, or Codex/chat gpt).
Vacation is over and work starts. Did some iPhone Duo adaptations for QCam at the airport. Now we have waist level mode and camcorder mode. Pretty fun! – at Santa María Huatulco, Oaxaca
AI allows people to waste time on the low priority “small things”so that they can avoid what’s truly important.
The least used accessory for my iPad Pro was the Apple Pencil. So no, I won’t buy an Apple Pencil usb for the iPhone Duo. It will be even more finicky.
一口新饭 retweeted
我给自己的 App 做了一轮 ASO：
从只跟踪几个词，到找出上百个相关关键词；从几乎没有词排在前列，到多个词排到第一，绝大多数进入前十，核心关键词上升超过 200 位。
我把具体方法整理成了指南：用 Astro、Google Trends 和 ChatGPT 选词、写文案、做截图、看效果。步骤具体，欢迎大家批评指正
ourcoders.com/guide/show/app…
Sam’s argument is like: personal computers are a terrible allocation of scarce resources. You need mainframes. Actually, many data centers will have the same fate of mainframes. You will have personal devices do better jobs than today’s data centers in the “cloud” not so far in the future.
sam: local inference is a terrible allocation of scarce resources
“a lot of people want local inference, and somehow they assume that the problem is solved now because we have open-source models, we're going to be able to run them on our computers, and so we're not going to need all the stuff in the data centre
but from a global allocation perspective, actually it's quite terrible to have everybody, like, have a bunch of the world's limited supply of DRAM sitting in their own computer and being used five percent of the time
you want maximum utilisation in the cloud if you're trying to deal with this very supply-constrained environment
somehow many people's intuition is the opposite of that”
