zero @Zero_cool5

Most Codex tasks don’t need Astra. I let Sol High coordinate and pick the model for each subtask: Luna → quick lookups Terra → everyday coding Sol → deep debugging Astra → the hard stuff up to 80% less Astra usage, depending on the workload. here’s the setup 👇

GitHub acaba de solucionar el mayor problema del vibe coding. Acaban de lanzar Spec Kit y en poco tiempo ya tiene +126K estrellas. ¿La idea? En vez de tirar prompts vagos y rezar para que el agente no rompa tu proyecto… Spec Kit obliga a la IA a crear una especificación estructurada ANTES de tocar código. La IA primero entiende lo que quieres construir, pregunta lo que falta, organiza el proyecto y después empieza a programar. Eso significa menos tiempo arreglando errores absurdos, menos código inconsistente y resultados mucho más predecibles cuando trabajas con agentes. El flujo es simple: /constitution → reglas y estándares /specify → qué quieres construir /clarify → dudas antes de empezar /plan → arquitectura y stack /tasks → tareas ordenadas /implement → ejecución Compatible con Claude Code, Cursor, Copilot, Codex, Gemini CLI y +25 agentes. 95K estrellas. 8K forks. Open source. Publicado por GitHub. Enlace al repositorio👇

Codex tip: A cost-efficient Luna + Sol agent tree, orchestrated by Astra. Effort levels chosen by weighing DeepSWE’s pass rates, average cost per task, and agent steps. Hand this to Codex to set it up 👇

I'm a Principal engineer & I passed system design rounds of Amazon, Atlassian, Walmart, Saleforce, and Deliveroo. Trust me, learning system is not hard. Start from these fundamental concepts: 1) Load Balancing: lnkd.in/gH9rdjCx 2) CDN: lnkd.in/g83A7-rM 3) Caching: lnkd.in/gTjxhv2V 4) Cache Invalidation: lnkd.in/geC955AY 5) Rate Limiting: lnkd.in/gWqJzCNJ 6) API Gateway: lnkd.in/gBNKpecH 7) CAP Theorem: lnkd.in/g4yFYkEi 8) Sharding: lnkd.in/gFi23iNV 9) Replication: lnkd.in/gikkrmNp 10) Partitioning: lnkd.in/gQhJS8ii 11) Queues: lnkd.in/gPGiuxtu 12) Microservices: lnkd.in/gZfYV2Qu 13) Microservices Vs Monoliths: lnkd.in/gM-dKE3D 14) Fault Tolerance: lnkd.in/gdamMmtc 15) Database Scaling: lnkd.in/ghq4v_gQ 16) Service Discovery: lnkd.in/gjfbNVBe 17) Consistency models: lnkd.in/gGkMENA3 18) Eventual Consistency: lnkd.in/gdSn54SK 19) Distributed Transactions: lnkd.in/gTc8pSbH 20) Leader Election: lnkd.in/g-kwhzSb 21) Horizontal vs Vertical Scaling: lnkd.in/gW-Vi9Qt 22) Back of the Envelope Estimation: lnkd.in/gQ6vtM3U 23) Idempotency, Data Latency & Finale: x.com/system_monarch… Let me know what you'd like me to cover, would love to help :)

Puneet Patwari @system_monarch

3 months ago

x.com/i/article/2076…

2 20 166 688K 540

Andrej Karpathy just explained the 5 shifts turning LLMs into agentic systems. 00:00 - Memory turns chat into personal AI 06:41 - Multimodal AI reads the world 16:58 - Thinking models solve harder tasks 24:51 - Search makes LLMs live 30:58 - Tools turn LLMs into workers Most people are still treating LLMs like chatbots. Karpathy is showing the full stack: Memory → Vision → Reasoning → Search → Tools Prompting is the old workflow. Agentic systems are the new one. This 40-minute talk is worth more than most paid AI agent courses. Bookmark and watch it before everyone catches up. Then read how to turn LLMs into self-improving agent loops below

Lunar @LunarResearcher

3 months ago

x.com/i/article/2082…

18 171 875 2.4M 3K

Codex users: here's an amazing way to take advantage of the efficiency and capability of GPT-5.6 Luna Max. It's called 'sol-advisor'. 1. GPT-5.6 Sol High as orchestrator 2. GPT-5.6 Luna Max as implementer for routine tasks 3. GPT-5.6 Terra Max as implementer for complex tasks 4. Fresh GPT-5.6 Sol instance as reviewer A free and open-source plugin for Codex. Installation instructions below 👇 I dare you to install it, and tell me if you hit your weekly usage limits using 'sol-advisor'. I doubt you can.

Dan McAteer @daniel_mac8

3 months ago

Codex users: do this right *now*. Max reasoning effort is off by default. Luna at Max reasoning is ~ Sol Medium / Opus 5 Medium level at 1/6th the cost. Change your life today.

180 212 4K 1.1M 3K

More opensource goodness. We have just released a CLI and TypeScript SDK for finding, validating, and fixing security vulnerabilities in your code. Scan repositories, review changes, track findings over time, and run security checks in CI. github.com/openai/codex-s…

This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser. I've also had some success asking the LLM to present its output as slideshows, etc. More generally, imo audio is the human-preferred input to AIs but vision (images/animations/video) is the preferred output from them. Around a ~third of our brains are a massively parallel processor dedicated to vision, it is the 10-lane superhighway of information into brain. As AI improves, I think we'll see a progression that takes advantage: 1) raw text (hard/effortful to read) 2) markdown (bold, italic, headings, tables, a bit easier on the eyes) <-- current default 3) HTML (still procedural with underlying code, but a lot more flexibility on the graphics, layout, even interactivity) <-- early but forming new good default ...4,5,6,... n) interactive neural videos/simulations Imo the extrapolation (though the technology doesn't exist just yet) ends in some kind of interactive videos generated directly by a diffusion neural net. Many open questions as to how exact/procedural "Software 1.0" artifacts (e.g. interactive simulations) may be woven together with neural artifacts (diffusion grids), but generally something in the direction of the recently viral x.com/zan2434/status… There are also improvements necessary and pending at the input. Audio nor text nor video alone are not enough, e.g. I feel a need to point/gesture to things on the screen, similar to all the things you would do with a person physically next to you and your computer screen. TLDR The input/output mind meld between humans and AIs is ongoing and there is a lot of work to do and significant progress to be made, way before jumping all the way into neuralink-esque BCIs and all that. For what's worth exploring at the current stage, hot tip try ask for HTML.

Thariq @trq212

6 months ago

x.com/i/article/2052…

1K 2K 18K 14.8M 35K

Anthropic engineer just released a 2-hour workshop on "Graph Engineering" for agentic systems: “80% of our engineers are using self-improving loops. Now everyone is building agentic graphs.” • 00:00 - Introduction to RAG & Graphs • 06:39 - Core of "Graph Engineering" (state, nodes) • 14:29 - 3 feedback loops of Graph agents • 23:06 - Agent evaluation with Graphs • 36:29 - Agent cycles in graphs • 1:15:22 - Agentic RAG & agent context • 1:41:20 - Evaluation datasets based on Graphs This 2-hour workshop will replace 10 paid courses on agentic engineering. Watch it today, then learn how to become a Graph Engineer in the article below.

Codez @0xCodez

3 months ago

x.com/i/article/2079…

127 798 5K 7.7M 15K

Andrej Karpathy just said what no one wants to admit: "Your agent loop is generating garbage at scale and you can't see it." 1 output looks fine. 1,000 outputs = dead entropy, dead diversity. 90% of AI companies are building demos, not products. In a 20-minute conversation, Karpathy reveals the full reasoning stack

Ricker @0xRicker

3 months ago

Anthropic CEO Dario Amodei just reveal that should scare every software engineer: "90% of code on key teams at Anthropic is already written by Claude." They didn't fire the engineers. They made them 10x more productive instead. In this 17-minute talk Dario lays out the full

7 4 60 240K 79

Andrej Karpathy spent 8 years at OpenAI and Tesla. Last week he put everything he knows into one free 2-hour lecture. People pay $15k for bootcamps that teach half of this. You probably don't have 2 hours right now. Don't lose this in the feed. Watch it. Then read the guide below and build your first loop.

Raytar @Raytar

4 months ago

x.com/i/article/2052…

37 271 2K 6.8M 6K

el fundador de una empresa china de IA valorada en más de $20,000,000,000 acaba de dar una clase de 40 minutos sobre enjambres de agentes la explicación más clara que he visto sobre sistemas de IA a gran escala cámbiala por tus 2 horas de Netflix de esta noche

Google engineer explained how to fine-tune a tiny LLM from 46% to 90% accuracy on your phone in 21 minutes - better than $1500 on-device AI bootcamps. pick Gemma 270M -> generate synthetic task data -> fine-tune with LoRA -> quantize to int4 -> deploy to Pixel and hit 2000 tokens per second. That loop is how a 270M model beats a 70B one on your task, running fully offline in your pocket. Gemma 270M + synthetic data + LoRA + int4 quantization + on-device runtime - that's the stack. Watch and save it, then fine-tune your own tiny agent tonight.

h100envy @h100envy

3 months ago

x.com/i/article/2077…

20 58 457 1.0M 1K

Anthropic and Andrew Ng built an agent that uses 90% fewer tokens from scratch: they dropped the entire book of Frankenstein into a prompt - 108,000 tokens asked one question the input dropped from 108,000 tokens to 11 here's how: step 1 → put everything that never changes at the top - tools, then system, then docs step 2 → mark where the static part ends. everything above it gets cached step 3 → one stray space breaks it - and you pay full price again step 4 → the cache dies in 5 min - every read resets the clock step 5 → cached tokens don't count against your rate limits. free headroom most people never touch this - it pays for itself on day one watch & bookmark - this 1-hour brilliant course ↓

codila @0xCodila

4 months ago

x.com/i/article/2069…

59 306 2K 5.1M 8K

Ex-NVIDIA engineer who built Unsloth explained RL, kernels, reasoning, quantization, and agents in 2 hours 42 minutes - better than $5000 fine-tuning bootcamps. pick the base model -> write triton kernels for 2x faster fine-tune -> quantize to 4-bit -> run GRPO/DPO -> ship a reasoning model on your single GPU. That loop is why Unsloth is the default way to fine-tune Llama, Qwen, Gemma, and Phi on hardware you already own. Unsloth + Triton kernels + 4-bit quantization + GRPO/DPO + single-GPU fine-tuning - that's the stack. Watch and save it, then fine-tune your first model tonight.

h100envy @h100envy

4 months ago

x.com/i/article/2068…

14 75 582 1.0M 1K

Ex-Google engineer explained AI agent loops, harness, evals in 20 minutes - better than 500$ courses. trace every run → judge it with an LLM → diagnose → fix → ship. That loop is how agents self-improve over time. Agent loops + memory + harness + evals - thats the stack. Watch it, then save the framework below.

Codez @0xCodez

5 months ago

x.com/i/article/2064…

154 1K 8K 9.9M 24K

50 WEBSITES GOOGLE DOESN'T WANT YOU TO KNOW 1. 12ft. io — bypass any paywall 2. libgen. is — millions of free textbooks 3. sci-hub. se — free research papers 4. alternativeto. net — find free app alternatives 5. justwatch. com — find where to stream anything 6. archive. org — access any old webpage ever 7. gutenberg. org — 70K free classic books 8. pdfdrive. com — free PDF downloads 9. openculture. com — free courses from top unis 10. wolframalpha. com — solve any math instantly 11. photopea. com — free Photoshop in browser 12. squoosh. app — compress any image free 13. remove. bg — remove image backgrounds free 14. cleanup. pictures — erase objects from photos 15. unscreen. com — remove video backgrounds 16. carbon. now. sh — turn code into art 17. ray. so — beautiful code screenshots 18. shots. so — free product mockups 19. smartmockups. com — mockups without Photoshop 20. haveibeenpwned. com — check if you were hacked 21. virustotal. com — scan any file for malware 22. privnote. com — send self destructing messages 23. temp-mail. org — disposable email instantly 24. file. io — share files that auto delete 25. archive. ph — save any webpage forever 26. similarsites. com — find any site alternatives 27. radio. garden — listen to any radio worldwide 28. everynoise. com — explore every music genre 29. tunefind. com — find songs from any show 30. musicforprogramming. net — music to focus with 31. mynoise. net — custom focus soundscapes 32. coffitivity. com — cafe sounds for productivity 33. elicit. org — AI research paper assistant 34. consensus. app — search what science agrees on 35. connectedpapers. com — map research visually 36. semanticscholar. org — free academic search 37. scispace. com — understand any research paper 38. summarize. tech — summarize any YouTube video 39. phind. com — AI search for developers 40. regex101. com — test any regex instantly 41. codebeautify. org — format any code cleanly 42. jsonformatter. org — read JSON like a human 43. explainshell. com — understand terminal commands 44. raindrop. io — bookmark manager that works 45. downdetector. com — check if any site is down 46. tineye. com — reverse image search 47. fast. com — check your internet speed 48. smallpdf. com — edit PDFs free 49. ilovepdf. com — merge and split PDFs 50. 10minutemail. com — temp email in seconds The internet is bigger than Google shows you. Most people never leave the first page.

keyboard_arrow_left Previous keyboard_arrow_right Next

A senior Google engineer just dropped a 19-page PDF on "Loop Engineering" for LLM and agentic systems. Act → Observe → Learn → Repeat • Act: the LLM proposes a code transformation (tile this loop, parallelize that one). • Observe: a compiler runs it and reports back - is it valid? faster? slower? by how much? • Learn: the LLM reads that feedback and adjusts its next move. • Repeat until it stops finding improvements. The agent gets smarter purely from grounded feedback inside its own context window. This 19-page PDF totally changed the way I’m building agentic systems today. Read it now, then explore the article below.

Codez @0xCodez

5 months ago

x.com/i/article/2064…

154 1K 8K 9.9M 24K

Loop Engineering 精华文章汇总! 2026 年 Agent 开始聚焦在长任务后，重点慢慢变成了： 如何设计一个能够持续思考、执行、观察、验证和演进的循环系统？ 从 Codex 到 Claude Code，从 OpenHands 到各种 Coding Agent。 业余项目和生产级系统之间最大的差距是Harness 工程，包括 Loop。 Agent 能不能持续工作几十分钟甚至几个小时？ 能不能在失败后恢复？ 能不能控制成本？ 能不能知道什么时候停下来？ 这些问题，最终都落到了 Loop 设计上。 📚 推荐阅读 1. Loop Engineering — Addy Osmani addyosmani.com/blog/loop-engi… 2. Loop Engineering — Firecrawl firecrawl.dev/blog/loop-engi… 3. What Is the AI Agent Loop? — Oracle blogs.oracle.com/developers/wha… 4. Harness Engineering — OpenAI openai.com/index/harness-… 5. Harness Engineering for Coding Agent Users — Martin Fowler martinfowler.com/articles/harne… 6. Agentic Loops: From ReAct to Loop Engineering datasciencedojo.com/blog/agentic-l… 7. Loop Engineering for AI Agents (Memory-First) — Mem0 mem0.ai/blog/loop-engi… 📄 推荐论文 1. Agentic Harness Engineering arxiv.org/abs/2604.25850 2. From Agent Loops to Structured Graphs arxiv.org/abs/2604.11378 🛠 推荐研究的开源项目 Codex CLI github.com/openai/codex OpenHands github.com/All-Hands-AI/O… PydanticAI github.com/pydantic/pydan… OpenAI Agents SDK github.com/openai/openai-… 重点研究： Loop 如何运行 Loop 如何停止 Loop 如何验证 Loop 如何恢复 Loop 如何调试 Prompt 决定 Agent 如何开始。 Context 决定 Agent 能看到什么。 Loop 决定 Agent 最终能走多远。 Loop Engineering： Think ↓ Act ↓ Observe ↓ Verify ↓ Evolve ↓ Repeat 你设计循环。 Agent 在循环中持续改进。 每完成一次循环，系统都会比上一次更接近目标。 Agent 从来不缺 Loop。 缺的是 Loop 的工程学。