虽然我很讨厌 Anthropic,但这 5 篇文章真的值得所有 AI 新手阅读
Anthropic 这家公司虽然真的很让人讨厌,要不就是直接给你封号,要不就是直接让你不能使用,但是你不能说他写的内容差
昨天我搜索资料的时候,发现 Anthropic 官网上提供了非常多优秀的文章,尤其是针对 AI 领域的内容都写得比较好,而且最大的优点就是通俗易懂
我从里面选了 5 篇最值得新手阅读的文章,如果你也想学习 AI Agent,可以从这里面开始看
1、Building effective agents|构建高效的 AI Agent
我最推荐新手阅读的一篇文章,讲清楚了 Agent 和 Workflow 的区别,以及 5 种常见的工作流设计模式,适合刚开始学习 Agent 的人
https://t.co/9RF4vgj6sN
2、Claude Code: Best practices for agentic coding|AI 编程最佳实践
主要介绍怎么正确使用 Claude Code,包括项目配置、任务规划、工具调用和代码验证,适合想学习 AI Coding 的新手
https://t.co/rWbZzkVafK
3、Equipping agents for the real world with Agent Skills|Agent Skills 入门
如果你还不理解 Skill 是什么,可以看看这篇,讲清楚了 Skill 的基本结构、SKILL.md 和按需加载机制,以及怎么给 Agent 扩展能力
https://t.co/0f5kU0YB6f
4、Effective context engineering for AI agents|Agent 上下文工程
主要讲解上下文管理,为什么上下文不是越长越好,以及怎么通过信息检索、工具管理和上下文压缩,让 Agent 运行得更加稳定
https://t.co/PBG11sK6yU
5、How Anthropic teams use Claude Code|Anthropic 内部 AI 使用案例
这篇更偏向实际应用,分享 Anthropic 内部不同团队怎么使用 Claude Code,包括编程、产品、设计和日常工作,即使没有编程基础也可以参考
https://t.co/pqUSRZ7nHm
最后我想说的是,公司虽然很垃圾,但是内容还是值得看的,取其精华,去其糟粕
尤其是刚开始学习 AI Agent 的新手,我觉得这 5 篇文章已经足够作为一个不错的入门起点了
This is f*cking insane: OpenAI co-founder Andrej Karpathy says a tiny 1B model can beat most of today's giant LLMs.
"A 1B model can be much more effective than most current LLMs. 90% of current models waste a ton of capacity and money on remembering stuff that doesn't matter"
In 64 seconds he lays out why intelligence lives in the system around the model, not in its size.
Memorized Internet → Cognitive Core → Tools → Feedback Loop → Agent That Works
Save this before the algorithm buries it.
Everything below is free:
Neural Networks: Zero to Hero
Karpathy's own course, from scratch
https://t.co/Pk7Sz5tiap
nanochat
Karpathy's "best ChatGPT that $100 can buy"
https://t.co/JWMZ6s0itw
Effective Harnesses for Long-Running Agents
The system around the model, by Anthropic
https://t.co/i3fgVnsLuu
Effective Context Engineering for AI Agents
What to feed the model and what to cut
https://t.co/8Wcs0IHTSb
All of this costs $0
99% will keep waiting for a bigger model.
The 1% who build better systems will win with a smaller one.
Save it and start with the first link tonight
then read the article below ↓
This is f*cking insane: OpenAI co-founder Andrej Karpathy says a tiny 1B model can beat most of today's giant LLMs.
"A 1B model can be much more effective than most current LLMs. 90% of current models waste a ton of capacity and money on remembering stuff that doesn't matter"
In 64 seconds he lays out why intelligence lives in the system around the model, not in its size.
Memorized Internet → Cognitive Core → Tools → Feedback Loop → Agent That Works
Save this before the algorithm buries it.
Everything below is free:
Neural Networks: Zero to Hero
Karpathy's own course, from scratch
https://t.co/Pk7Sz5tiap
nanochat
Karpathy's "best ChatGPT that $100 can buy"
https://t.co/JWMZ6s0itw
Effective Harnesses for Long-Running Agents
The system around the model, by Anthropic
https://t.co/i3fgVnsLuu
Effective Context Engineering for AI Agents
What to feed the model and what to cut
https://t.co/8Wcs0IHTSb
All of this costs $0
99% will keep waiting for a bigger model.
The 1% who build better systems will win with a smaller one.
Save it and start with the first link tonight
then read the article below ↓
物理AI——
是黄仁勋再创业的目标
除了吃瓜以外,其实我想到的更多是,如果黄仁勋重新创业,他会做什么?
以老黄的能力、精力,可以直接面对 60 多位高管,没有中间层,直接领导。以他对 AI 世界的理解,加上目前大模型的发展速度,数字世界的奇点已经只是时间问题,那剩下的就只有物理世界了。
目前的 Omniverse 已经成为数字世界的练兵场,那下一步将会是跟真正物理世界的结合(Cosmos),也就是成为物理世界的 AI。而物理世界必然要跟 IoT(Internet of Things)结合,所以下一个发展目标应该就是 Internet of Things
第一就是传统的微小电路mcu,第二是生物级别的芯片,第三就是端侧的最小的 LLM。只要从这三个领域里边去找到任意一个方向,都会是无限的。
所以,英伟达的董事们正在后怕吧?所以那么多人邀请他回来,他的员工,他的合作者。
Keep building on @BNBCHAIN , and you’ll be seen.
Eight months ago, TagAI was deployed on @BNBCHAIN , and people discovered its strengths as a token launch platform.
Eight months later, they’re beginning to see TagAI’s innovation as a social prediction market and community social network.
Of course, we’re especially grateful to @BNBChain for their support during this stage.
@BNBChainDevs@cz_binance@heyibinance@yzilabs@ellazhang516