Mr. Buzzoni
@Mr_Buzzoni • 28,032 subscribers
• building a second brain with Сlaude & Obsidian • • @cele8ex member • co-founder @polynternet •
Shorts
Videos
holy f*ck, Karpathy put an entire AI degree on GitHub for free 7 repos, 306,741 stars, and in the right order they take you from backprop to agents that run research overnight each repo is the next layer of the one before 1. micrograd -> how a neural net actually learns backprop in about 100 lines of Python ↳ 2. nn-zero-to-hero -> the video course his lecture notebooks, from one neuron to GPT ↳ 3. minbpe -> how text turns into tokens the tokenizer every LLM starts with, in clean minimal code ↳ 4. nanoGPT -> your own GPT the simplest repo to train and fine-tune one ↳ 5. llm.c -> the same training with nothing hidden raw C and CUDA, no PyTorch underneath ↳ 6. nanochat -> a full chat model end to end GPT-2 level training cost ~$43,000 in 2019 here it's about $48 ↳ 7. autoresearch -> agents that do the research an agent edits the training code, trains for 5 minutes, keeps what improved and repeats all night ↳ train one GPT from scratch and you become the person who gets handed the AI project, along with the raise that comes with it the classroom half of the same degree, what Stanford actually teaches about AI, is below ↓
Mr. Buzzoni176,190 görüntüleme • 2 gün önce
Atlassian's revenue: $1.79 billion last quarter Atlassian's move: fire the engineer who built their infrastructure his move: post a 38-minute breakdown of every system he built, free for anyone to copy what he revealed: > Envoy proxy instead of enterprise load balancers > sidecar architecture for auth, logging, rate limits > DynamoDB + SQS for async provisioning > Packer + SaltStack for automated VM deployments at scale Atlassian charges per employee across 350,000 customers this guy just handed you the enterprise playbook for free save this
Mr. Buzzoni21,792,134 görüntüleme • 5 ay önce
Claude Code tip: once Opus 5.5 is your main model, stop letting your Fable 5.1 quota go to waste put it on call with /advisor run /advisor fable Opus 5.5 keeps doing the work Fable 5.1 sits on the sidelines, reads the whole session, and steps in at three moments: → before a plan: is this right? → when the same error comes back: am I going the wrong way? → before "done": did I miss anything? Fable 5.1 advises. Opus 5.5 writes the code the same idea sits under Jev engineering: the expensive model stops weighing in on every step and only gets called at the moments that change the outcome • the full setup > Opus 5.5 on high runs the main session > subagent one reads code > subagent two edits and runs tests > subagent three looks up docs > all three on medium > Fable 5.1 on call hand the tree and this prompt to Claude Code 👇 "Set up my Claude Code to match this tree: 1. Reuse fitting subagents from ~/.claude/agents and .claude/agents. > Propose new ones only for missing roles > Set each to model: opus, effort: medium > Leave any that set a different model alone and list them 2. Set main session effort to high via effortLevel in ~/.claude/settings.json 3. Check for env vars that disable the advisor (CLAUDE_CODE_DISABLE_ADVISOR_TOOL, DISABLE_TELEMETRY, anything that stops flag fetching) and CLAUDE_CODE_EFFORT_LEVEL, which overrides subagent effort. Report them, don't change them 4. Add a rule to ~/.claude/CLAUDE.md: ask the advisor before a big plan, when an error repeats, and before calling a long task done Show me the changes first. Don't edit files yet." ↳
Mr. Buzzoni326,872 görüntüleme • 1 ay önce
this one page turns Obsidian + Opus 5.5 + Jev into a Karpathy-grade second brain that thinks with you 24/7 built on Karpathy's second brain stack and classic Zettelkasten the loop: > Jev reads every note in your Obsidian vault before Opus 5.5 does > Opus writes only the lines that actually matter > each note links itself into the Zettelkasten graph on save > your second brain keeps working while you don't the result: > 80% off your weekly chat time > hours back for the work your manager actually pays for > a Karpathy-style vault that compounds instead of rotting the people wiring Obsidian, Jev and loop engineering into one second brain this week are the ones the next promotion finds first
Mr. Buzzoni47,606 görüntüleme • 6 gün önce
this is pure f*cking gold for anyone running coding agents Jev founder Diogo Amogo wrote a PDF on building a Jev harness the promise: > 200x faster > 400x cheaper the model hasn't been the slow part for a while the speed and the cost sit in the harness wrapped around it • how to use it > drop this PDF and the article below into Claude Code or Codex > let it rebuild its own setup one evening of setup and your agent starts next week on a harness most teams still don't have
Mr. Buzzoni78,793 görüntüleme • 10 gün önce
Karpathy sketched the second brain. these 7 open-source repos already built it for Obsidian + Claude Code 250,000+ GitHub stars between them, and every one is free to clone sorted by stars: 1. graphify - 134k stars turns your codebase, docs, SQL schemas and PDFs into one knowledge graph you can query ↳ 2. Understand-Anything - 88k stars maps any codebase into an interactive knowledge graph you can click through ↳ 3. claude-obsidian - 15.8k stars a second brain that organizes itself: drop in a source, Claude Code files it ↳ 4. obsidian-second-brain - 4.9k stars long-term memory for Claude Code, stored as plain markdown in your vault ↳ 5. mcp-obsidian - 4.5k stars one MCP server and Claude can read, search and write your whole vault ↳ 6. claudesidian - 2.6k stars a starter vault designed around Claude Code from the first note ↳ 7. second-brain - 756 stars an Obsidian wiki the model keeps up to date, built on Karpathy's LLM Wiki pattern ↳ where to start: > your knowledge lives in code -> graphify > your knowledge lives in Obsidian -> claude-obsidian > you just want Claude to remember -> obsidian-second-brain > you live in Claude Desktop -> mcp-obsidian the skills that sit on top of all seven are in the full map below ↓
Mr. Buzzoni23,447 görüntüleme • 3 gün önce
Claude Code tip, and it's absolute free f*cking gold: run Opus 5.5, Sonnet 5.5 and Fable 5.1 as one team and stop burning Opus tokens on routine work the setup in one line: plan on high, delegate on medium, keep Fable on call • who does what > Opus 5.5 on high - plans and ships the code > Sonnet 5.5 on medium - explorer reads code, worker edits and runs tests, researcher pulls docs > Fable 5.1 via /advisor fable - reads the whole session and speaks up only when it matters • when Fable 5.1 steps in -> before a plan: is this the right approach? -> when an error repeats: am I digging in the wrong place? -> before "done": what did I miss? Jev engineering takes it one layer lower: which file, which tool, retry or stop all go to Jev in under half a second, so the big models only see the real forks paste this into Claude Code ↓ "Rebuild my Claude Code setup: 1. Find subagents in ~/.claude/agents and .claude/agents that fit explorer, worker and researcher. Draft only the missing ones. Set each to model: sonnet, effort: medium. List any that pin a different model and leave them 2. In ~/.claude/settings.json set effortLevel to high and advisorModel to fable. 3. Report anything that disables the advisor (CLAUDE_CODE_DISABLE_ADVISOR_TOOL, DISABLE_TELEMETRY, flag-fetching blockers) and CLAUDE_CODE_EFFORT_LEVEL. Change nothing. 4. Add to ~/.claude/CLAUDE.md: consult the advisor before a large plan, when an error repeats, and before calling a long task done. Show every change as a diff. No edits until I say go." ↳
Mr. Buzzoni144,764 görüntüleme • 27 gün önce
this is pure f*cking gold for anyone running coding agents Jev founder Diogo Amogo wrote a PDF on building a Jev harness the promise: > 200x faster > 400x cheaper the model hasn't been the slow part for a while the speed and the cost sit in the harness wrapped around it • how to use it > drop this PDF and the article below into Claude Code or Codex > let it rebuild its own setup one evening of setup and your agent starts next week on a harness most teams still don't have 👇
Mr. Buzzoni128,589 görüntüleme • 1 ay önce
KARPATHY WAS RIGHT. THIS 40-MINUTE Y COMBINATOR LECTURE PROVES IT Karpathy said we're in the 1960s of AI - most people using Claude Opus 4.8 are still acting like it's just a search engine > software 3.0 - LLMs as operating systems, not chatbots > autonomous agents that run entire workflows without you watching the 32 skills in this article are how you actually cross that line bookmark this 👇
Mr. Buzzoni568,465 görüntüleme • 5 ay önce
I'm sharing with you one of the best papers I've read on Harness Engineering it takes apart Claude Code (Opus 5.5), Codex, Gemini CLI and more, and boils all of them down to one line: agent = model + harness what you walk away with: > what a harness actually is > how today's coding agents are built under the hood > the patterns every one of them repeats > where the whole category is heading > their checklist for building your own the model gets the headlines. the harness decides whether the agent works loop and graph engineering on top of it - in my article below
Mr. Buzzoni84,231 görüntüleme • 22 gün önce
this one page turns Obsidian + Opus 5.5 + Jev into a second brain that thinks with you around the clock built on Karpathy's ideas Jev reads every note before Opus does Opus writes only what matters your vault keeps working while you don't cuts 80% off your weekly chat time frees you up for the work your manager actually pays for the people who set this up early are the ones the next raise finds first
Mr. Buzzoni47,753 görüntüleme • 14 gün önce
I STOPPED LETTING CLAUDE MAKE DECISIONS THE DAY I BUILT THIS JEV AGENT FOLDER I used to let Claude decide, write and act on every single step -> now Claude only writes. Jev makes the calls, code does the acting, and every step leaves a receipt here's what's inside the folder: • the input > AGENTS.md - when to call Jev, when to skip it > state/build_state.py - goal, workers, done, missing, constraint. evidence, never a summary • Jev decides (questions/) > - which model tier gets the task > next_worker.py - which worker acts next > - keep or drop each tool output > - is the goal actually met > - does this send, pay or delete? • code acts (rules/) > hard_rules.py - stop after ten actions, never publish unapproved > thresholds.yaml - act only on confident answers. fraud needs 0.95 • Claude writes (workers/) > - sources and notes > - drafts and briefings > the only place in the whole folder where text gets generated • the proof > receipts/decisions.jsonl - options offered, chosen id, re-check, fallback > evals/ - dozens of my own labelled traces, thresholds tuned on them • the guards (hooks/) > pre_tool_use.sh - every command checked before it runs > - checks "all done" before the agent is allowed to stop 10,000 decisions for $0.42. median 300 ms an LLM writes, Jev decides, code acts
Mr. Buzzoni80,463 görüntüleme • 1 ay önce
REA + Opus 5.5 might be the most dangerous combo on the internet right now and 99% of builders have no idea what to do with it you're paying every month for apps whose features are already sitting on your laptop send this one prompt to Opus 5.5 it reverse-engineers any paid closed-source app for you the engineer who brings a working internal copy of a $500/seat tool to Monday standup is the one who ends the quarter with a new title full guide on becoming a REA engineer from scratch below ↓
Mr. Buzzoni34,412 görüntüleme • 13 gün önce
this is free f*cking gold for anyone building a second brain his second brain article pulled 8 million views, so he packed the whole system into one repo the repo, the guide, the tools, the learning path, all free • the guide > 10 sections, 65 pages, from the core idea to troubleshooting > 5 tracks on top, 44 pages, 15 of them build guides with working code • the machine (.claude/) > 18 agent skills, one per workflow > 72 slash commands - /ingest-pdf, /ingest-youtube, /ingest-voice, /backfill > 6 subagents - curator, linker, researcher, reviewer, ingestor, graph-analyst > 4 are read-only, so nothing rewrites your vault behind your back • the scripts (plain Python, zero dependencies) > graph export, link checker, vault stats, chat converter, site builder • the starter vault > its own CLAUDE.md with page contracts and linking rules > raw/ stays untouched once it lands, wiki/ is what the agent keeps up > log.md - one line per run, so every change is traceable • 87 vetted resources > 28 tools, 26 Obsidian plugins, 15 repos, 12 skills, plus papers and articles five tracks to choose from: > knowledge graphs > Jev engineering > agent harnesses > loop engineering > eval engineering new to this? start with the second brain guide. already deep in it? jump straight to the tracks the Jev engineering and loop engineering tracks alone are worth the clone a consultant bills four figures to build you a research system like this. here it's a public repo under MIT ↳
Mr. Buzzoni56,795 görüntüleme • 23 gün önce
I can't believe this is f*cking free how to build your first AI agent (full guide, with a Jev decision layer) a year ago I burned two weeks on my first agent. this guide gets you there in an afternoon in the right hands it makes one builder ship like a small team:
Mr. Buzzoni61,929 görüntüleme • 26 gün önce
SpaceXAI put out a 12-page PDF on the best GrokBot use cases - read it and you'll use GrokBot better than 95% of people what the bot is, which use cases work today, and how to deploy it without burning trust: step 1 -> every working use case has one shape: read, check, apply a rule, write a draft. if your task fits, the bot can take it step 2 -> five checks before handoff: repeatable, cross-system, rule-based, draft-first, cheap to get wrong. fails one? add a gate or wait step 3 -> start with sales. 36 LinkedIn drafts queued overnight, none sent without review. no team has more agent-ready work step 4 -> engineering gets bug repro, CI/CD watch and dependency audits. the bot picks up the ticket, hits staging and hands back a repro pack step 5 -> operations hands over inbox, expenses and travel. the bot logs into portals that have no API, and finance gets 8-12 hours back at month-end step 6 -> marketing owns 13 of the 56 roles: paid media, content calendars, competitor watch. all overnight, all draft-first step 7 -> the approval line: every documented use case stops before external action. nothing gets sent, booked or paid without you step 8 -> roll out over four weeks: watch, review everything, review flags, then a scheduled routine. skip a week and the risk stacks up step 9 -> the fleet: a Chief of Staff bot on top, specialists below, bots messaging each other. you stop being the middleman step 10 -> rank tasks by the cost of a mistake. the bot nails 49 in a row, then does something strange on the 50th with full confidence. keep a human at every exit the result: 56 roles, one playbook. the bot takes the 80% that repeats, you keep the 20% that matters then read how to build a 24/7 GrokBot agent team in the article below
Mr. Buzzoni36,827 görüntüleme • 16 gün önce
I HAVEN'T OPENED CLAUDE AT MIDNIGHT SINCE I BUILT THIS FOLDER I used to wake up and find out what broke overnight -> now the receipts are already there. dated, graded, waiting on my review what sits inside the folder that took that shift: • the contract > CONTRACT.md - the shift rules, committed > contract.local.md - my overrides, gitignored • the harness (.claude/loops/) > settings.json - spend caps and timeouts. set once > schedule.yml - 05:00, pr-hunter. cron fires it, not me > rubrics/ - code.md reviews merges, writing.md checks voice, safety.md says what not to touch > pr-hunter/ - plan.md runs wake, read, act, verify > does the work. is the gate to merge > tools.allow - what it may call, and nothing else • the state > receipts/ - one folder per shift. 17 shifts today > 5,382 receipts kept. none of them edited > trace.log - what happened, line by line > checkpoint.json - resumes exactly where it stopped > budget.json - $6.10 spent of a $40 day • the edges > - the panic file. never used > .mcp.json - the tools it's allowed near > alerts.yml - who gets paged when a loop stops 2 loops. 3 rubrics. 8 months in service. 0 incidents nothing here is clever. every file exists because a run went sideways once I wrote the rule down instead of trying to remember it $6 a day for a shift I used to work at 1am the model doesn't wake. the folder does
Mr. Buzzoni71,089 görüntüleme • 1 ay önce
this 18-page SpaceXAI PDF on Grok Bot is pure f*cking gold a senior engineer there wrote it it covers how to build a Grok Bot agent system that runs 100% of your tasks 24/7 the shift: you stop managing agents one by one and build a single CEO bot that holds the org chart and distributes the work the playbook in 6 steps: 1. map one workflow as source -> owner -> artifact -> evidence -> finish if the finish can't be proven, it can't be automated 2. give the CEO bot a durable ledger: task id, owner, state, artifact pointer, deadline, retry count memory lives in the system now, not inside a chat window 3. add a specialist only when the job needs its own tools, permissions or eval harness everything else is just a reusable skill the manager calls 4. attach skills to routines with triggers, idempotency keys, retry budgets and a kill switch one key per task kills every duplicate run before it starts 5. parallelize only stable, independent tasks through typed handoffs scale the verifier before the workers, or the fleet eats its own output 6. demand independent evidence (a diff, a log, a rendered artifact) before anything is marked done humans sign off only on irreversible or ambiguous moves the result: a chatbot turns into a 24/7 operating system every task owned, every handoff typed, every result proven read this before you build another Grok Bot it has the architecture most agent tutorials skip the full system is in the article below ↓
Mr. Buzzoni23,183 görüntüleme • 11 gün önce
CLAUDE + OBSIDIAN + LOOP ENGINEERING = A VAULT THAT RUNS ITSELF the core idea: the vault is the loop's state, not the chat window everything Claude knows lives in a .md file the loop: > capture - a thought lands in 00-inbox > context - claude pulls links, tags, and neighbouring notes > draft - edits happen inside a git worktree, never the live vault > review - a critic agent checks the diff before anything ships > commit - appended to the vault, nothing gets rewritten the key insight: frontmatter fields like supports, contradicts, and supersedes are graph edges, not metadata - the note format is the write API start with a plain loop, it runs about 2-4x the cost of one direct call > only move to a full graph once state has to outlive the session, several agents need to coordinate, or you have to explain what changed - that jump can run 10-50x one review assistant climbed from 55% to 72% to 84% just by moving through these shapes in order
Mr. Buzzoni137,517 görüntüleme • 2 ay önce
