Josh Jeong

🚨 OpenAI 's own engineers just showed how to actually use OpenAI Codex properly. 60 minutes. free. built by the people who contribute to made it. watch the masterclass. bookmark it. worth more than every $900 coding course you almost bought. you’ve been using Codex like a simple coding tool… while it’s actually a full software engineering system. watch this, it could the best 62 minutes of your life:

Today, we kill influencer marketing. Just enter your website and create your brand’s own hyper-realistic AI influencer. Fastlane then creates thousands of videos of your influencer promoting your product, all cloned from viral content. Deploy your own AI UGC army in seconds:

Absolutely game changer for anyone running paid ads @motionapp_ simply must have tool

Create with a Custom GPT - Short Film Prompt Generator : Seedance 2.0 and ChatGPT 2.0 Grok , used Dreamina

i probably shouldn’t be saying this publicly… but i’m tired of seeing people struggle with AI videos seedance 2.0 + claude + tiktok is literally broken i’m talking hyper-realistic videos that make people stop scrolling and think: “nah… this has to be a real person” i spent weeks testing EVERYTHING and finally locked in a system that: - makes AI sound human (no robotic nonsense) - adds natural body movement + micro expressions - keeps videos consistent no matter how long they are - cuts tool costs down like crazy i wasn’t going to share this yet… but f*ck it. RT + reply “Ai UGC” and i’ll send the full step-by-step system (if you’re serious, follow so i can dm you)

Most refreshing thing about working for @xai: no pointless meetings, no stupid acronyms, no annoying politics. Everyone is an IC. And it works.

所有使用 Obsidian 的人 请立刻安装 Notebook Navigator 插件 并在「插件设置 - 桌面外观」中 开启「双窗格模式」 真的无敌了 试用了半小时找不到任何槽点 也许会完全改变你的 Obsidian 使用习惯 可以替代一大批零散小插件 并且非常的克制且有品味 这个真的很重要 相信我 你会在 5 分钟内感受到这个插件的美好 我真的很少有这样的评价

无法用言语形容的画面，太美了 油画般的世界真的很治愈 可能这就是大家向往的西方浪漫吧

🚨This guy just broke down how to actually beat the new Meta Andromeda algorithm. 14 minutes. free. watch the video. bookmark it. worth more than every $1000 course you almost bought. you've been running Meta ads without knowing these 8 steps. Then read the guide below.

You asked for it, and here it is! If you're not adopting this in your creative workflows, you're going to wake up very soon in a world you won't understand:

Your Hermes Agent can now build full videos with the official HyperFrames skill by @HeyGen HyperFrames videos are HTML-native, so your agent has total control over the final output Video made entirely by Hermes using the HyperFrames skill

Hermes users after reading this:

The best LLM setup there is right now: Frontend: Opus 4.7 Backend: GPT-5.5 Debugging: Sonnet 4.6 Research: Grok 4.2 Writing: GPT-5 Mini Image: Gemini 3.1 This is all you need.

It's time to cancel your @OpusClip subscription. Use @claudeai to create 9:16 clips instead. Here's the super fast demo, while I get ready 💋

一个可以下载任何 YouTube 视频、干净地去除人声、进行转录、翻译成 100 多种语言、克隆原声并完成全自动配音的工具。全程不到 2 分钟。100% 本地运行。免费。它叫 Voice-Pro，我会把链接留在评论区。

Claude Code at 9 AM: "great architectural plan, I love it, this is the way" Claude Code at 2 PM: writes 400 lines that contradict the morning's plan, half of which weren't needed Claude Code at 6 PM: "you're absolutely right, let me rewrite that" Claude Code at 8 PM: contradicts itself again I fixed this with 3 plugins and 1 scanner. 5 minutes of setup.

I cut my Claude Code bill by 70% with 3 GitHub repos Tested 10 repos that claim to save tokens. Ran each one for a full day on real projects Most were hype. 3 actually changed my workflow Here's what they do and how to set each one up: 1. RTK (Rust Token Killer) The problem: every terminal command you run like "npm install", "git log", test suites, build output It gets dumped raw into Claude's context. You're paying for Claude to read thousands of lines of noise The fix: RTK is a CLI proxy that sits between your terminal and Claude. It filters output before it hits context How to set it up: - install the binary (one command, zero dependencies) - it auto-detects Claude Code sessions - runs silently in the background - no config needed, works out of the box My result: 60-90% reduction on terminal-heavy sessions Best for: developers who run builds, tests, or installs frequently during Claude sessions 2. Caveman Claude The problem: Claude responds like a teacher by default, full explanations, caveats, context. Helpful when learning, but expensive when shipping The fix: one line in your CLAUDE.md that rewrites Claude's output behavior to be compressed and terse How to set it up: - clone the repo - copy the prompt snippet into your project's CLAUDE.md - DONE! Claude immediately responds in compressed mode - Toggle it off anytime by removing the line My result: 65-75% fewer output tokens with identical code quality I compared outputs side by side for a full day, same code, same logic, half the tokens Best for: experienced developers or code users who don't need Claude to explain what it's doing 3. Context Mode The problem: if you use MCP tools (Playwright, GitHub API, browser, any external tool), their raw responses flood your context silently, literally single page read can eat thousands of tokens The fix: sandboxes all tool output into SQLite and only passes clean summaries into your conversation How to set it up: - install as a Claude Code plugin - it auto-intercepts MCP tool responses - raw data goes to local SQLite - Claude only sees the summary My result: 98% context reduction on tool-heavy workflows Best for: anyone running Playwright, browser tools, API calls, or multiple MCP servers How to stack all 3: - RTK cleans the input (terminal noise) - Caveman cleans the output (verbose responses) - Context Mode cleans the tools (MCP bloat) Each fixes a different layer and together they compound You don't need all 3, just pick based on your workflow: heavy terminal output? → start with RTK output too verbose? → start with Caveman lots of MCP tools? → start with Context Mode Quick test right now: run /context in a fresh Claude Code session and check how much context is already used before you type anything If it's over 20%, you're leaking tokens All repo links in the comments Save it.

I'm deleting Notion AI because of this. It's called claude-obsidian and it turns Obsidian into a running notetaker that files, cross-references, and maintains itself while you work. The difference from every other Obsidian AI plugin is the architecture. Smart Connections is a chat interface on top of your notes. This is a knowledge engine that creates, organizes, and evolves the notes themselves. → Auto-organizes everything you ingest into entities, concepts, and sources → Flags contradictions across your vault with callouts linking to sources → Runs autonomous 3-round web research and files the output automatically → Parallel batch ingestion for multiple sources at once → Queries cite specific wiki pages, not fuzzy similarity matches → Native Obsidian Bases dashboard plus visual canvas companion → Cross-project setup lets one vault feed all your Claude Code projects 100% Opensource. github.com/AgriciDaniel/c…

10 INSANE GITHUB REPOS THAT REDUCE CLAUDE CODE TOKENS BY UP TO 90% 1. RTK (Rust Token Killer) Filters terminal output before feeding it into context. Can reduce by 60-90% with commonly used dev commands. github.com/rtk-ai/rtk 2. Context Mode Offloads Playwright or GitHub tool outputs to SQLite, passing only clean summaries to the conversation. 98% reduction. github.com/mksglu/context… 3. code-review-graph Uses Tree-sitter to build a local knowledge graph of the codebase. Claude only reads the necessary parts, achieving 49x reduction for large monorepos. github.com/tirth8205/code… 4. Token Savior An MCP server that references code by symbols instead of entire files. 97% reduction with code navigation, zero external dependencies. github.com/Mibayy/token-s… 5. Caveman Claude Makes Claude speak in caveman-style tone to cut output tokens by 65-75%. Maintains technical accuracy. github.com/JuliusBrussee/… 6. claude-token-efficient Just place a CLAUDE.md file to keep outputs concise. No code changes needed, ideal for verbose sessions. github.com/drona23/claude… 7. token-optimizer-mcp MCP server with caching, compression, and smart tool control. Compresses repeated tool outputs for 95%+ reduction. github.com/ooples/token-o… 8. claude-token-optimizer Reusable setup prompt collection to optimize projects. Save 90% of tokens in just 5 minutes. github.com/nadimtuhin/cla… 9. token-optimizer Discovers ghost tokens that silently devour context. No quality degradation even after compaction. github.com/alexgreensh/to… 10. claude-context (by Zilliz) MCP that brings the entire codebase into context with BM25 + vector search. ~40% reduction with equivalent search accuracy. github.com/zilliztech/cla… Save this for later🔖