sluke (@_sluke_) header
sluke (@_sluke_) profile picture

推短emo长 ex创业者|开源| AI Infra | 投融资 | 行业研究

Joined May 2008
Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)Photo by sluke (@_sluke_)
sluke retweeted
Finch 1.8 发布,你可以创建不同的 Bot 来完成长期任务。大量的社区运营工作,服务器运维,网站更新,写文章,这样的重复上下文的工作用长会话更适合。没想到用 Agent 的体验被 Bot 改变了。除了个别项目的迭代,大多数日常场景并不需要多 session。 欢迎下载体验 finchwork.app/
4
2
1
17
2,369
关于大模型直接调小模型的权重,提升小模型能力,问了一下专业人士,反馈是“直接编辑效果比想象中好,比直接训练弱太多”
92
新的 muxy 也很好用,跟 herdr 客户端有相似之处
75
github.com/sluke/fx-desktop Community desktop client for fx — a native MyGo app that connects to the fx agent over ACP. Conversations, tool output and permissions are rendered natively; noninteractive fx commands run behind native forms, and only interactive authentication (fx login, setup, teams) uses the in-window console. This is not an official Vercel Labs release. Platforms: macOS and Linux, matching the fx CLI.
78
从通用大模型中,特化出来小模型,是个在大规模高频场景下能自洽的商业模式。即使在特定任务中能超越通用大模型的效果,没规模也不划算,又是个大厂的自留地。
76
sluke retweeted
Typecho 1.3.1-rc is out. Security fixes + a full XML-RPC rewrite (fixes trimmed code indents, false parsed as true, >7.5MB uploads). ⚠️ Test on a copy: passwords migrate on login, and downgrading will lock you out. github.com/typecho/typecho/r…
1
1
1
213
LayerFS v1.0.0 is out 🎉 Open-source (MIT) Recursive filesystem for AI agents. 6 core ideas: 1. Ephemeral computing: agents spend ~80% thinking, so each Bash call mounts a Workspace, then commits 2. Decentralized agents: every agent commits to one Store, no DB server 3. Multiagent coordination: every tool call is a Commit; same-file clashes refused 4. Recursive filesystem: forks copy no files, at any depth 5. Ultra storage: dedup, then compression: 10 copies → 0.31 6. Instant bootstrap: ignored files stay in every Commit. ~5 min install vs 0.05 s mount And 4 more: 7. Auto checkpoints: fork from any step 8. Append-only objects: content-addressed, never rewritten 9. .lww: lockfiles and caches take the newest edit 10. Big file, small edit: a write keeps only its bytes Here are our links: github: github.com/Ephemeral-AI-Lab/… website: layerfs.ai/ Watch the video below, and have Fun!!!
147
173
85
1,293
161,564
各种支持 claude code 的壳子,是不是可以增加一个前置小插件,输入中文,用个便宜或者免费的模型翻译成英文,再提交给 cc?
131
最近的agent发展,总让我想起强殖装甲和巨人殖装,一个可强化可迁移共享的加强套件
80
vibe coding之前,让ai读一遍相近产品的文档,可能就不需要另造轮子了
83
sluke retweeted
Holon(@holonrun) 源于我自己日常使用 AI Agent 时的几个真实痛点,从一个满足自用的小玩具,变成了我每天都在用的主力系统。但它离成熟的通用工具还有不少距离。 如果你对这种长驻、异步、由工作项驱动的 Agent 感兴趣,非常欢迎体验并给我反馈。
9
6
1
34
6,279
ClickHouse(chDB)管队列事实,DuckDB 管注释与便携查询
vcfclick scores variant calls hap.py-style, then keeps the concordance as SQL tables instead of a static report - so TP/FP can be stratified by gnomAD frequency, ClinVar class or gene after the run. chDB is the default backend. github.com/nuin/vcfclick
1
1
2
513
在陪伴孩子成长的过程中,我想记住重要的瞬间,也想看清自己在亲子沟通上做得不够的地方。我和伙伴们一起做了 Pastok ,帮我记住日常中的语音沟通,助我日省吾身。当前支持 Apple Watch 和 iPhone。有类似需求的朋友不妨试试,欢迎反馈意见。 Let the past talk. pastok.com
12
9
5
67
31,692
感谢 Tom 的推荐, projectbrain.md/zh/ 是我们 @mindmux 团队前不久刚刚开源的一个项目上下文的标准,README 是给人看的;AGENTS.md / CLAUDE.md 教 Agent 怎么干活;BRAIN.md 回答为什么这么干。灵感来自 LLM Wiki,欢迎大家体验,最近会更新优化更多自动化编译的场景。
Brain.md creates a persistent memory layer for coding agents by capturing project knowledge as plain Markdown files inside the repository. github.com/mindmuxai/brain.m…
2
1
6
1,040
咱花了 ~$2000 额度写的 mac 原生终端 Kero 正式发布: kero.sh
102
18
7
254
82,979
Gork 4.5 进行了几个前端小任务,改了几个小bug,周限额就到了
315