开源工具 no-ai-slop,专治英文稿里的“AI味”。
它不负责把文章写得高级,只管删掉机器腔。目前 GitHub 上已经拿了五千多颗星。
精准剔除套路:内置了二十多种 AI 常见句式。比如强行对比、废话开头、假深刻和堆砌同义词。
保留你的原味:它只删模板句,不动你的词汇和幽默感。改完还会列个清单,让你看清动了哪里。
两种使用模式:直接清理就输入 /no-ai-slop。只想找毛病不下手,就用 is this slop。
还能反向教学:让它故意写一段最油的 AI 文。拿来当反面教材培训新人特别直观。
安装也简单,在 Claude Code 等工具里粘贴一行命令就行。不过要注意,这套规则专门针对英文。纯中文稿子用它不太对口。
AI 帮你写得快,但能不能写出你自己的味道,才是真本事。
项目地址:https://t.co/U4BeQA6eRK
详细解读:https://t.co/PZuIp2wZfp
JUST IN: 🇺🇸 White House crypto advisor Patrick Witt says the SEC and CFTC will deliver crypto regulation.
“We’re full steam ahead.”
WITH OR WITHOUT CONGRESS, CLARITY IS COMING! 🚀
"Making strategic long-term decisions accrues more value than just doing dumb buybacks all the time."
Avichal Garg on the difference between short-term and long-term greedy when it comes to buybacks
"Right now there are sort of only two stages. There is the prospective fee switch with prospective buybacks stage, where people believe that at some point the fees will switch on after there's enough market share, and that will somehow accrue back to the token."
"The second is you've actually decided to do that. And the expectation from the market right now is you should take as much of your fees as possible and do the buybacks. I think it's because people are short-term greedy."
"When you're in growth mode, you don't want to give back a lot. You want to own market share in a large and growing market before you start to give that back to the token holders. So you want to take all those fees and give them back to LPs, give them back to users, invest in customer acquisition, invest in product development."
"At a certain point of maturity you can potentially start to do some buybacks, and you can turn the dial on how aggressive you get. And that's long-term greedy."
"We're basically speed running the evolution of companies. Joint stock companies and LLCs, eventually C corps, we got boards and we got CEOs. That evolution took 150 or 200 years, and we're just going to speed run it in 10 years. But we're not quite there yet."
Every backend service should have a /feedback endpoint.
Agents are quickly becoming the heaviest users of most APIs. When one hits a missing feature or a bug, it should be able to say so right there, in a structured way.
If the request makes sense, another agent drafts the PR and a human approves it. Software that improves itself based on what its users actually tried to do speeds up recursive self improvement.
so far crimsonland decomp took 510 agent hours. it's very close, sitting at 96.4% with only two functions remaining. the larger one is 97.75% matched. you can be the one to solve this decomp.
https://t.co/3j4TFyCtxG
AI token costs are going to continue falling aggressively.
This will bring low-cost, high-powered intelligence to every business.
An incredible economic boom of productivity and prosperity will follow.