Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Jev’s founder just dropped the 12-page harness that makes coding agents 400× cheaper the first thing it shows: your agent spends under 10% of its tokens actually writing code → reading and searching take 56.2% of tool turns and 46.5% of tokens. that’s where the money goes → routing... Opus → Sonnet → Opus to save costs 6.19. pure Opus costs 4.15. every handoff reprocesses the whole context → compress after the question, never before. every chunk gets scored: hide, short summary, long summary or full → instructions load by condition. touch *.tsx and the style guide loads. touch billing/ and its gotchas file loads. compaction can’t erase either → route by trust, not difficulty. secrets and infra stay on first-party frontier models, public docs go to the cheapest one → every command passes an allow / ask / deny policy that reads the script before it runs. not just its name their claim: 200× faster, 400× cheaper drop the PDF into Claude Code or Codex and let it rebuild your harness ↓show more

Carnage

3,193 subscribers

61,361 views • 5 days ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

7 Comments

magsimich5 days ago

Most of the cost is hiding there

catman5 days ago

This makes search, reading, and context control cheaper; the new bottleneck is deciding what information deserves full fidelity before compression or routing.

Atlas5 days ago

compaction solves what the agent needs to see now. we can handle what the agent needs to remember later, including previous attempts and the state they left behind

Gursers5 days ago

@heyjevbook can you substantiate these claims?

TeamShared4 days ago

Routing tool use by task and cost is a practical complement to shared context; the harder part is preserving those conventions across agents and sessions.

Brian Hadu5 days ago

56.2% on reading and searching is wild, what are the top tips to optimize that?

Kiell Tampubolon5 days ago

That 56.2% reading number matches what I see locally. The fix that actually moved the needle was scoring every chunk before it entered context: full text for the risky 10%, tight summaries for the rest. Cheap models can do the scoring, the frontier model only sees what survived.

Related Videos

Jev founder, Diogo Almeida, just put out a PDF on building a Jev harness for coding agents the claim: 200x faster, 400x cheaper the model stopped being the bottleneck a while ago. the harness around it is where the cost and the speed actually live five decisions sit on every loop, and only one of them writes code hand this PDF and the article below to your Claude Code or Codex instance and let it rebuild its own setup your agent gets a better harness tonight, and monday looks different than it would have 👇

Jev founder, Diogo Almeida, just put out a PDF on building a Jev harness for coding agents the claim: 200x faster, 400x cheaper the model stopped being the bottleneck a while ago. the harness around it is where the cost and the speed actually live five decisions sit on every loop, and only one of them writes code hand this PDF and the article below to your Claude Code or Codex instance and let it rebuild its own setup your agent gets a better harness tonight, and monday looks different than it would have 👇

Hanako

37,887 views • 3 days ago

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 👇

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. Buzzoni

120,326 views • 1 day ago

Anthropic shipped 125 settings for Claude The official docs cover 40 One developer found the other 85 and his API bill dropped from $340 to $87 - not by using a cheaper model - not by writing shorter prompts just by moving one line in a config file to the right place > memory scoped per project → past clients never bleed into new work > Extended Thinking on Light by default → 18–25% fewer Opus tokens in week one > cache_control moved to the right line → the fix that turned a $340 bill into $87 > plugins and MCP servers toggled off when idle → saved 25–40K tokens per session > per-project model override → Haiku for docs, Sonnet for infra, Opus only where it matters most Claude users are running a $100/month tool at 30% of its actual capability

Anthropic shipped 125 settings for Claude The official docs cover 40 One developer found the other 85 and his API bill dropped from $340 to $87 - not by using a cheaper model - not by writing shorter prompts just by moving one line in a config file to the right place > memory scoped per project → past clients never bleed into new work > Extended Thinking on Light by default → 18–25% fewer Opus tokens in week one > cache_control moved to the right line → the fix that turned a $340 bill into $87 > plugins and MCP servers toggled off when idle → saved 25–40K tokens per session > per-project model override → Haiku for docs, Sonnet for infra, Opus only where it matters most Claude users are running a $100/month tool at 30% of its actual capability

Shadow Nick

380,456 views • 4 months ago

I BUILT BRIER WITH ANTHROPIC AND GROK - AN LLM SYSTEM THAT CALLS 490 OUT OF 500 CORRECTLY AND SEALS EVERY ONE INTO A ROBINHOOD CHAIN BLOCK free, open, and every figure reproduces on your machine in 4 seconds with no API key - $50 to $750 in 2 days ask → panel → seal → wait → settle → grade brier - runs Opus 5 and Sonnet 5 side by side and reads 400 markets a second before it says a single number I ran 200 events in 10 minutes, knew the answer to each one before it landed, and sent it straight to the chain the hash goes into a Robinhood Chain block, and from there it goes to the exchange or the platform it was right 77% of the time, and in 2 days one free GitHub repo brought in $750 all of it without a single line of code - built entirely by the new Anthropic and Grok models the repo is open and free - take it and seal your first forecast before the market opens

I BUILT BRIER WITH ANTHROPIC AND GROK - AN LLM SYSTEM THAT CALLS 490 OUT OF 500 CORRECTLY AND SEALS EVERY ONE INTO A ROBINHOOD CHAIN BLOCK free, open, and every figure reproduces on your machine in 4 seconds with no API key - $50 to $750 in 2 days ask → panel → seal → wait → settle → grade brier - runs Opus 5 and Sonnet 5 side by side and reads 400 markets a second before it says a single number I ran 200 events in 10 minutes, knew the answer to each one before it landed, and sent it straight to the chain the hash goes into a Robinhood Chain block, and from there it goes to the exchange or the platform it was right 77% of the time, and in 2 days one free GitHub repo brought in $750 all of it without a single line of code - built entirely by the new Anthropic and Grok models the repo is open and free - take it and seal your first forecast before the market opens

Noisy

48,731 views • 20 days ago

A screen recording is the best context you can give an AI agent. It's also the worst. A few seconds of video is thousands of images. Feed that to an agent and you instantly blow up its memory. So I built a tool that fixes it. How it works: Hit record and just do your thing. Click around, type, and talk out loud about what you're doing. While you record, it quietly captures the stuff that actually matters: your voice transcription, every click and the exact thing you clicked, what you typed, and the shortcuts you press. When you stop, it doesn't hand the agent the whole video. It writes a short, readable summary: a timeline of what you said and did. The agent reads that tiny summary first. Then, only when it needs to actually see a moment, it pulls that one exact frame from the video, like flipping to a single page instead of reading the whole book. The result: the agent understands everything you did! Full context. Less tokens.

A screen recording is the best context you can give an AI agent. It's also the worst. A few seconds of video is thousands of images. Feed that to an agent and you instantly blow up its memory. So I built a tool that fixes it. How it works: Hit record and just do your thing. Click around, type, and talk out loud about what you're doing. While you record, it quietly captures the stuff that actually matters: your voice transcription, every click and the exact thing you clicked, what you typed, and the shortcuts you press. When you stop, it doesn't hand the agent the whole video. It writes a short, readable summary: a timeline of what you said and did. The agent reads that tiny summary first. Then, only when it needs to actually see a moment, it pulls that one exact frame from the video, like flipping to a single page instead of reading the whole book. The result: the agent understands everything you did! Full context. Less tokens.

Milind S

12,409 views • 2 months ago

Boris Cherny gave Claude Code a memory that has no database, no indexing, no setup, and 80% of Anthropic runs on it a CLAUDE.md "gets automatically read into context at the start of every session" 12:48 - CLAUDE.md loads into memory every session 13:15 - Keep memory short or it burns your context 15:48 - Memory stacks from you to your whole company 17:30 - The memory command shows what Claude remembers 19:33 - Hit pound and Claude remembers it forever the whole memory layer is files → drop a CLAUDE.md in your repo and Claude reads it in on every session → nest more in subfolders and it pulls them only where it works → stack them personal to project to enterprise so one config serves the whole team → hit the pound key mid task and Claude writes the new fact into memory itself the best coding agent's memory was never a database, it was markdown you can open and edit

Boris Cherny gave Claude Code a memory that has no database, no indexing, no setup, and 80% of Anthropic runs on it a CLAUDE.md "gets automatically read into context at the start of every session" 12:48 - CLAUDE.md loads into memory every session 13:15 - Keep memory short or it burns your context 15:48 - Memory stacks from you to your whole company 17:30 - The memory command shows what Claude remembers 19:33 - Hit pound and Claude remembers it forever the whole memory layer is files → drop a CLAUDE.md in your repo and Claude reads it in on every session → nest more in subfolders and it pulls them only where it works → stack them personal to project to enterprise so one config serves the whole team → hit the pound key mid task and Claude writes the new fact into memory itself the best coding agent's memory was never a database, it was markdown you can open and edit

NO1ennn

13,015 views • 1 month ago

Orchestrators vs. Graphs, clearly explained! orchestrators are great, and everyone builds one first. here is the ceiling: an orchestrator sits above the work and routes every message. five agents report to it. it reads all five. it decides what each one does next, and reads all five replies. that is ten trips through one context, and by the fifth agent that context has read four reports, five instructions and its own reasoning about all of them. Graph engineering fixes this by removing the seat: not a better router, but no router at all. you need both, and here is the sentence that resolves the whole confusion: an orchestrator sits above the work and holds all of it. a graph is the shape of the work, and holds none of it. ↳ above the work: one context that has to see everything before anything ships ↳ inside the work: a splitter that hands out and lets go, and a merge that reads nothing Prompts → Context → Harness → Loops → Graphs the coordination did not disappear. it moved into the edges, where it costs nothing and cannot get tired. the trick is noticing what you actually built. if one node has to see every result before the run can finish, you did not remove the bottleneck. you hired it, gave it the longest context in the system, and made it the thing you were counting on to stay sharp. one thing to know before you scale it. an orchestrator degrades in the one way nothing catches. ↳ it does not crash, time out or return an error. it stays up and keeps routing ↳ it just starts routing worse, somewhere around the fifth report, and every downstream agent does exactly what it was told that last one catches careful people. you can have perfect isolation on every worker and still have one window quietly drifting at the top, and the traces will all look clean because each worker did its job. and the one that eats whole nights: the merge is where this shows up first. ranking five findings is not judgment, it is a sort. if a model is doing it, you are paying a model to read five reports so it can put them in an order that three lines of code would have got right, and now that model has read everything too. below i have quoted my full guide on graph engineering. it covers the three topologies, the verifier patterns, and where the gate should actually open. save this and read it below ↓

Orchestrators vs. Graphs, clearly explained! orchestrators are great, and everyone builds one first. here is the ceiling: an orchestrator sits above the work and routes every message. five agents report to it. it reads all five. it decides what each one does next, and reads all five replies. that is ten trips through one context, and by the fifth agent that context has read four reports, five instructions and its own reasoning about all of them. Graph engineering fixes this by removing the seat: not a better router, but no router at all. you need both, and here is the sentence that resolves the whole confusion: an orchestrator sits above the work and holds all of it. a graph is the shape of the work, and holds none of it. ↳ above the work: one context that has to see everything before anything ships ↳ inside the work: a splitter that hands out and lets go, and a merge that reads nothing Prompts → Context → Harness → Loops → Graphs the coordination did not disappear. it moved into the edges, where it costs nothing and cannot get tired. the trick is noticing what you actually built. if one node has to see every result before the run can finish, you did not remove the bottleneck. you hired it, gave it the longest context in the system, and made it the thing you were counting on to stay sharp. one thing to know before you scale it. an orchestrator degrades in the one way nothing catches. ↳ it does not crash, time out or return an error. it stays up and keeps routing ↳ it just starts routing worse, somewhere around the fifth report, and every downstream agent does exactly what it was told that last one catches careful people. you can have perfect isolation on every worker and still have one window quietly drifting at the top, and the traces will all look clean because each worker did its job. and the one that eats whole nights: the merge is where this shows up first. ranking five findings is not judgment, it is a sort. if a model is doing it, you are paying a model to read five reports so it can put them in an order that three lines of code would have got right, and now that model has read everything too. below i have quoted my full guide on graph engineering. it covers the three topologies, the verifier patterns, and where the gate should actually open. save this and read it below ↓

Hanako

43,781 views • 13 days ago

I cant believe this guy just made a permanent solution to context bloat and open sourced it all! when we tested this tool (Context+) for solving an issue on the OpenCode repository, the agent using this tool used ~6.5k fewer tokens, found the code and fixed it in half the time! the results were surprising: 6 to 10k tokens saved per prompt, completed task in ~2 minutes while the agent running without the tool took ~4 mins for the same and got stuck in loops bro built an entire beast by using all the modern tools that we could think of: undo trees, semantic search by meaning (by haskellforall), advanced refactoring, blast radius, advanced file context trees, restore points... i can keep going on semantic code search and context trees are the future of agentic coding and this tool proves it the feature i loved the most is semantic search and how it gets things done 2x faster with least possible tokens it makes an agent that actually knows what it’s doing and not just guessing, it makes meaning from your code similar to RAG. if you aren't optimizing your context, you are just burning money the developer says this tool is still under development, it can have unexpected behavior and the docs need updates but the video shows the reality of how fast it can be github: get here:

I cant believe this guy just made a permanent solution to context bloat and open sourced it all! when we tested this tool (Context+) for solving an issue on the OpenCode repository, the agent using this tool used ~6.5k fewer tokens, found the code and fixed it in half the time! the results were surprising: 6 to 10k tokens saved per prompt, completed task in ~2 minutes while the agent running without the tool took ~4 mins for the same and got stuck in loops bro built an entire beast by using all the modern tools that we could think of: undo trees, semantic search by meaning (by haskellforall), advanced refactoring, blast radius, advanced file context trees, restore points... i can keep going on semantic code search and context trees are the future of agentic coding and this tool proves it the feature i loved the most is semantic search and how it gets things done 2x faster with least possible tokens it makes an agent that actually knows what it’s doing and not just guessing, it makes meaning from your code similar to RAG. if you aren't optimizing your context, you are just burning money the developer says this tool is still under development, it can have unexpected behavior and the docs need updates but the video shows the reality of how fast it can be github: get here:

forloop

226,491 views • 7 months ago

The new Fable 5.1 built three games and did it 7% cheaper! We gave the same prompts through Atomic Agent: ✦ Sticky ball that rolls up everything in its path ✦ Capybara surfing down a river, Subway Surfers style ✦ Dumpling running an endless conveyor dodge Outputs: ✦ Claude Fable 5, via Atomic Agent: $7.65 ✦ Claude Fable 5.1, via Atomic Agent: $7.08 Run it through an agent and that cache discount is where the money hides. The loop reads the file, watches it break, and pulls the whole context back in to fix it, over and over. Every one of those re-reads costs a quarter on 5.1 of what it cost on 5. Same games on screen, and the newer version just spends less to get there. Across three builds the loop paid for the upgrade on its own.

The new Fable 5.1 built three games and did it 7% cheaper! We gave the same prompts through Atomic Agent: ✦ Sticky ball that rolls up everything in its path ✦ Capybara surfing down a river, Subway Surfers style ✦ Dumpling running an endless conveyor dodge Outputs: ✦ Claude Fable 5, via Atomic Agent: $7.65 ✦ Claude Fable 5.1, via Atomic Agent: $7.08 Run it through an agent and that cache discount is where the money hides. The loop reads the file, watches it break, and pulls the whole context back in to fix it, over and over. Every one of those re-reads costs a quarter on 5.1 of what it cost on 5. Same games on screen, and the newer version just spends less to get there. Across three builds the loop paid for the upgrade on its own.

Atomic Agent

20,255 views • 26 days ago

$THIS MIGHT BE THE #1 OPEN-SOURCE REPO FOR CLAUDE CODE RIGHT NOW. IT GIVES CLAUDE A MEMORY AND SLASHES YOUR TOKEN COST ON EVERY QUESTION The repo is safishamsi/graphify, a free open-source skill that turns any codebase into a knowledge graph Claude Code can read instantly. Instead of grepping through your files every session, Claude gets a map of how everything connects The problem it fixes: Every time you ask Claude Code about a big repo, it does the same thing, greps through dozens of files like a brute-force Ctrl+F, blows through your context window, and sometimes still misses the answer hiding in a file nobody searched. Claude Code has no memory of how your project is structured. Every session starts from zero What it does: It maps your entire codebase into a knowledge graph, capturing not just which files exist, but which functions depend on which, which modules are central, and which files cluster around the same concern. Claude queries the map instead of scanning files How it works, three passes: 1. Code structure, free and local. Tree-sitter parses your files and pulls out classes, functions, imports and call graphs. No LLM, no tokens, just your actual code mapped deterministically 2. Audio and video, if you have them. Transcribed locally and folded into the graph 3. Docs, papers, images. Here an LLM does semantic analysis, figuring out what each document means and where it fits. Only the meaning gets sent up, never your raw source It saves you money: Normally a question about a big repo makes Claude spawn explore agents that scan file after file, eating your context window and your token budget before you get an answer. With the graph already built, Claude queries the map instead of re-reading the codebase every time. Same answer, a fraction of the tokens. The graph only gets built once, then a hook rebuilds it after each commit for free, so you never pay that scanning cost again. The bigger the repo, the bigger the gap The best parts: it's a skill, so once installed Claude knows when to use it without you memorizing commands. It works on non-code folders too, point it at docs or notes and it can spin up an Obsidian vault How to add it to your Claude: 1. Install Claude Code if you haven't: npm install -g Paul Jankura-ai/claude-code 2. Add the skill: claude skill add safishamsi/graphify 3. Open your project folder and run /graphify . to build the graph 4. Optional, make it automatic: graphify hook install so the graph rebuilds after every commit That's it. Ask Claude about your repo and it reads the map instead of burning tokens on a file hunt Bookmark this$

THIS MIGHT BE THE #1 OPEN-SOURCE REPO FOR CLAUDE CODE RIGHT NOW. IT GIVES CLAUDE A MEMORY AND SLASHES YOUR TOKEN COST ON EVERY QUESTION The repo is safishamsi/graphify, a free open-source skill that turns any codebase into a knowledge graph Claude Code can read instantly. Instead of grepping through your files every session, Claude gets a map of how everything connects The problem it fixes: Every time you ask Claude Code about a big repo, it does the same thing, greps through dozens of files like a brute-force Ctrl+F, blows through your context window, and sometimes still misses the answer hiding in a file nobody searched. Claude Code has no memory of how your project is structured. Every session starts from zero What it does: It maps your entire codebase into a knowledge graph, capturing not just which files exist, but which functions depend on which, which modules are central, and which files cluster around the same concern. Claude queries the map instead of scanning files How it works, three passes: 1. Code structure, free and local. Tree-sitter parses your files and pulls out classes, functions, imports and call graphs. No LLM, no tokens, just your actual code mapped deterministically 2. Audio and video, if you have them. Transcribed locally and folded into the graph 3. Docs, papers, images. Here an LLM does semantic analysis, figuring out what each document means and where it fits. Only the meaning gets sent up, never your raw source It saves you money: Normally a question about a big repo makes Claude spawn explore agents that scan file after file, eating your context window and your token budget before you get an answer. With the graph already built, Claude queries the map instead of re-reading the codebase every time. Same answer, a fraction of the tokens. The graph only gets built once, then a hook rebuilds it after each commit for free, so you never pay that scanning cost again. The bigger the repo, the bigger the gap The best parts: it's a skill, so once installed Claude knows when to use it without you memorizing commands. It works on non-code folders too, point it at docs or notes and it can spin up an Obsidian vault How to add it to your Claude: 1. Install Claude Code if you haven't: npm install -g Paul Jankura-ai/claude-code 2. Add the skill: claude skill add safishamsi/graphify 3. Open your project folder and run /graphify . to build the graph 4. Optional, make it automatic: graphify hook install so the graph rebuilds after every commit That's it. Ask Claude about your repo and it reads the map instead of burning tokens on a file hunt Bookmark this

Yarchi

56,502 views • 3 months ago

ANTHROPIC LEAKED A $3.4M FILE THAT BUILDS AN AI BUSINESS AIMED AT $120K A MONTH IN 60 MINUTES AND COSTS $1.55 A DAY team of 5 costs $12,000 a month - this file costs $1.55 a day and never asks for a standup. the whole thing runs on one rule: the cheap model triages, the middle one builds, the expensive one decides. Opus 5 runs exactly once a day and is 13% of the bill - the priciest model is the smallest line on it. Sonnet 5 carries 3 parallel branches, and together with the hunt that's 81% of the entire bill. Haiku 4.5 reads 60 signals for pennies and hands back 12 letters - the block that used to eat your evenings now runs while you sleep. swap the models around and you get 5x overpaid, or a month of selling to nobody. the landing page, payment link, auto-replies and analytics get assembled at once in 12 minutes. 5 dashboards collapse into 3 numbers and one decision - and those numbers get written back into the root. $120K a month is either 400 customers at $300 or 1000 at $120 - and the price decides which of the two machines you built. the hour produces the machine, not the customers - it hunts just as hard even when nobody wants the product. save this and paste it into Claude Code - route by the cost of being wrong, not the price per token ↓

ANTHROPIC LEAKED A $3.4M FILE THAT BUILDS AN AI BUSINESS AIMED AT $120K A MONTH IN 60 MINUTES AND COSTS $1.55 A DAY team of 5 costs $12,000 a month - this file costs $1.55 a day and never asks for a standup. the whole thing runs on one rule: the cheap model triages, the middle one builds, the expensive one decides. Opus 5 runs exactly once a day and is 13% of the bill - the priciest model is the smallest line on it. Sonnet 5 carries 3 parallel branches, and together with the hunt that's 81% of the entire bill. Haiku 4.5 reads 60 signals for pennies and hands back 12 letters - the block that used to eat your evenings now runs while you sleep. swap the models around and you get 5x overpaid, or a month of selling to nobody. the landing page, payment link, auto-replies and analytics get assembled at once in 12 minutes. 5 dashboards collapse into 3 numbers and one decision - and those numbers get written back into the root. $120K a month is either 400 customers at $300 or 1000 at $120 - and the price decides which of the two machines you built. the hour produces the machine, not the customers - it hunts just as hard even when nobody wants the product. save this and paste it into Claude Code - route by the cost of being wrong, not the price per token ↓

Sprytix

29,960 views • 1 month ago

THE ANATOMY OF A SECOND BRAIN THAT WRITES ITSELF one Obsidian vault, three layers, and a loop that files every good answer back into a wiki you never touch the structure: > CLAUDE.md + AGENTS.md - the constitution. same rules whether Claude or Codex runs the loop > raw/ - inbox, articles, assets. frozen and read-only, you feed it and walk away > wiki/ - index.md, entities/, concepts/, synthesis.md. the agent grows this on its own > log.md - append-only, every move timestamped 47 sources in -> 210 pages out the agent typed 96% of it. you typed zero the cost trick: point Opus 5 only at synthesis.md, the one file that needs frontier reasoning. everything else runs fine on cheaper models under the hood it's loop engineering: you ask -> a good answer gets filed back -> the wiki grows start linking entities to each other and you're one step from a full graph your research compounds while you sleep. the full blueprint is below 👇

THE ANATOMY OF A SECOND BRAIN THAT WRITES ITSELF one Obsidian vault, three layers, and a loop that files every good answer back into a wiki you never touch the structure: > CLAUDE.md + AGENTS.md - the constitution. same rules whether Claude or Codex runs the loop > raw/ - inbox, articles, assets. frozen and read-only, you feed it and walk away > wiki/ - index.md, entities/, concepts/, synthesis.md. the agent grows this on its own > log.md - append-only, every move timestamped 47 sources in -> 210 pages out the agent typed 96% of it. you typed zero the cost trick: point Opus 5 only at synthesis.md, the one file that needs frontier reasoning. everything else runs fine on cheaper models under the hood it's loop engineering: you ask -> a good answer gets filed back -> the wiki grows start linking entities to each other and you're one step from a full graph your research compounds while you sleep. the full blueprint is below 👇

Annatar.md

23,649 views • 14 days ago

$THIS GUY USED OPUS 4.8 + KIMI K2.6 TO CUT HIS CODING BILL FROM $4,000 TO $700/MO. WITH KIMI RUNNING A 300-AGENT CODING FLOOR, HE STOPPED PAYING CLAUDE TO DO EVERYTHING kimi does the cheap heavy lifting. it can spin up hundreds of agents, push through thousands of steps, write the rough code, expand files, draft tests and handle the repetitive work that burns the most tokens opus 4.8 only comes in where the money is worth spending. first to plan the spec and define the rules, then again to tear apart the output, catch weak logic and flag the bugs that a fast swarm can miss that is what changed the economics. kimi handled the bulk of the volume for a fraction of the price, while opus stayed in the loop as the architect and the reviewer instead of the full-time builder most people still use one model for every step and wonder why their costs explode. this guy split the jobs properly. kimi runs wide and cheap. opus goes deep and skeptical. one builds fast, the other makes sure it should ship the real edge is not finding one perfect model. it is knowing where expensive intelligence actually matters and where cheap parallel output is already enough. that is how a $4,000 workflow turns into a $700 system$

THIS GUY USED OPUS 4.8 + KIMI K2.6 TO CUT HIS CODING BILL FROM $4,000 TO $700/MO. WITH KIMI RUNNING A 300-AGENT CODING FLOOR, HE STOPPED PAYING CLAUDE TO DO EVERYTHING kimi does the cheap heavy lifting. it can spin up hundreds of agents, push through thousands of steps, write the rough code, expand files, draft tests and handle the repetitive work that burns the most tokens opus 4.8 only comes in where the money is worth spending. first to plan the spec and define the rules, then again to tear apart the output, catch weak logic and flag the bugs that a fast swarm can miss that is what changed the economics. kimi handled the bulk of the volume for a fraction of the price, while opus stayed in the loop as the architect and the reviewer instead of the full-time builder most people still use one model for every step and wonder why their costs explode. this guy split the jobs properly. kimi runs wide and cheap. opus goes deep and skeptical. one builds fast, the other makes sure it should ship the real edge is not finding one perfect model. it is knowing where expensive intelligence actually matters and where cheap parallel output is already enough. that is how a $4,000 workflow turns into a $700 system

Gipp 🦅

16,989 views • 3 months ago

ANTHROPIC LEAKED A $4.2M FILE WHERE 4 AGENTS BUILD THE JEV SYSTEM IN 60 MINUTES, DOING $100K A MONTH four agents cost $0, and the system they assemble in an hour generates $100K a month scout → builder → critic → operator → JEV every agent is pinned to one exact model version, and none of them runs on "latest" the scout on Haiku 4.5 reads the market, the builder on Sonnet 5 builds everything, the critic on Opus 5 decides what ships the operator runs Haiku through the batch API at half price - the one agent that never has to hurry the builder takes 59% of the whole hour, because it's the one writing, and writing costs 5x more than reading short answers are cheap answers - and the whole economics rests on that one rule pin the versions in ROOT.md and next month JEV behaves exactly the same point it at "latest" and the model underneath updates quietly, no error, and the answers start to drift $100K a month is 334 customers at $299 four versions, one system - pin them and JEV stays JEV save this and paste it into Claude Code - let it build you these 4 agents ↓

ANTHROPIC LEAKED A $4.2M FILE WHERE 4 AGENTS BUILD THE JEV SYSTEM IN 60 MINUTES, DOING $100K A MONTH four agents cost $0, and the system they assemble in an hour generates $100K a month scout → builder → critic → operator → JEV every agent is pinned to one exact model version, and none of them runs on "latest" the scout on Haiku 4.5 reads the market, the builder on Sonnet 5 builds everything, the critic on Opus 5 decides what ships the operator runs Haiku through the batch API at half price - the one agent that never has to hurry the builder takes 59% of the whole hour, because it's the one writing, and writing costs 5x more than reading short answers are cheap answers - and the whole economics rests on that one rule pin the versions in ROOT.md and next month JEV behaves exactly the same point it at "latest" and the model underneath updates quietly, no error, and the answers start to drift $100K a month is 334 customers at $299 four versions, one system - pin them and JEV stays JEV save this and paste it into Claude Code - let it build you these 4 agents ↓

Sprytix

37,601 views • 6 days ago

ANTHROPIC LEAKED A TRADING DESK FILE THAT TAKES YOU FROM $10 TO $10,000 AND COSTS $1.50 A DAY Anthropic spent 5 years building this file on closed tech, it costs $1.50 a day and it can return $1.2M a year rules → reading → code → critic → journal → back to rules $10 to $10,000 is 1000×, which is ten doublings in a row at a coin flip that is 1 chance in 1,024, and a single loss anywhere sends you to zero instead of back one step by week four the balance reads $5,400 and only 0.8% are still alive on that rung the 1,024 who tried are not visible in the feed, only the one who arrived is you write the rules yourself: size, max loss, and what you never touch Haiku 4.5 reads the market and summarises it, and never once passes judgement Opus 5 works as the critic and attacks your own backtest instead of praising it the journal logs every trade and why, and it is what rewrites the rules for tomorrow the desk does not raise your odds on any single flip, it lowers how many flips you have to take save this and paste it into Claude Code - let it build your DESK.md ↓

ANTHROPIC LEAKED A TRADING DESK FILE THAT TAKES YOU FROM $10 TO $10,000 AND COSTS $1.50 A DAY Anthropic spent 5 years building this file on closed tech, it costs $1.50 a day and it can return $1.2M a year rules → reading → code → critic → journal → back to rules $10 to $10,000 is 1000×, which is ten doublings in a row at a coin flip that is 1 chance in 1,024, and a single loss anywhere sends you to zero instead of back one step by week four the balance reads $5,400 and only 0.8% are still alive on that rung the 1,024 who tried are not visible in the feed, only the one who arrived is you write the rules yourself: size, max loss, and what you never touch Haiku 4.5 reads the market and summarises it, and never once passes judgement Opus 5 works as the critic and attacks your own backtest instead of praising it the journal logs every trade and why, and it is what rewrites the rules for tomorrow the desk does not raise your odds on any single flip, it lowers how many flips you have to take save this and paste it into Claude Code - let it build your DESK.md ↓

Sprytix

22,861 views • 24 days ago

ANTHROPIC LEAKED A FILE THAT BUILDS A 12-PERSON TEAM FOR $0 - AND THAT TEAM BUILDS A BUSINESS DOING $1.4M A YEAR a 12-person team would cost you $50,000 a month, this one costs exactly zero and shows a result in 5 days root → contracts → staff → gates → clock → back to root the root never describes the product, it describes who is going to build it the first hour writes the team, the next day that team writes the product Opus 5 sits on the architect and the critic, three Sonnet 5 build, one more tests Haiku 4.5 keeps the records, and the auditor rewrites the team rules, never the product $1.4M is either 400 customers at $292 a month, or 39 clients at $3,000 the staff is real and costs $792 a year, the revenue line above it is arithmetic the cascade builds the machine, it does not sign a single one of those numbers want a different product - you change the task, not rebuild the system demand is decided by the offer, never by the file save this and paste it into Claude Code - let it build your ROOT.md ↓

ANTHROPIC LEAKED A FILE THAT BUILDS A 12-PERSON TEAM FOR $0 - AND THAT TEAM BUILDS A BUSINESS DOING $1.4M A YEAR a 12-person team would cost you $50,000 a month, this one costs exactly zero and shows a result in 5 days root → contracts → staff → gates → clock → back to root the root never describes the product, it describes who is going to build it the first hour writes the team, the next day that team writes the product Opus 5 sits on the architect and the critic, three Sonnet 5 build, one more tests Haiku 4.5 keeps the records, and the auditor rewrites the team rules, never the product $1.4M is either 400 customers at $292 a month, or 39 clients at $3,000 the staff is real and costs $792 a year, the revenue line above it is arithmetic the cascade builds the machine, it does not sign a single one of those numbers want a different product - you change the task, not rebuild the system demand is decided by the offer, never by the file save this and paste it into Claude Code - let it build your ROOT.md ↓

Sprytix

39,041 views • 21 days ago

you wrote a long careful note so it would remember the thing properly the note is the part that does not load what loads at the start of every session is the index line, the one sentence summary, and only the first two hundred lines of it or the first twenty five kilobytes, whichever runs out first the topic files underneath are read on demand, which means only if something in that one line makes it go and open them i checked mine: thirty nine index lines loading every session, thirty nine topic files loading none of the time the detail you were careful about is not context it is a footnote your agent has to choose to open and the sentence you dashed off in three seconds is the whole memory

you wrote a long careful note so it would remember the thing properly the note is the part that does not load what loads at the start of every session is the index line, the one sentence summary, and only the first two hundred lines of it or the first twenty five kilobytes, whichever runs out first the topic files underneath are read on demand, which means only if something in that one line makes it go and open them i checked mine: thirty nine index lines loading every session, thirty nine topic files loading none of the time the detail you were careful about is not context it is a footnote your agent has to choose to open and the sentence you dashed off in three seconds is the whole memory

Lagosaur

28,369 views • 20 days ago

run agent harnesses 100% private & offline. (no token costs, no API keys, 100% open-source) your agent runs locally. the model doesn't. every prompt, every file, and every secret still leaves your machine before the agent does anything with it. Magnitude fixes that. it's an open source inference server that runs models on your own hardware and plugs into the coding agent you already use. setup is one command. it profiles your machine, measures the memory bandwidth that sets your token rate, and hands back complete configurations instead of a list of models. each one names a model, a compression level, a context size, and a speed range you can expect. pick one and start working. it doesn't replace your harness. setup asks which one you want and writes that config for you. Pi, OpenCode, Claude Code, Codex, and Cline all work, and there's a built-in one tuned for local models if you don't have a harness yet. that one uses your shell, edits files, and runs scripts out of the box. add skills and it handles Excel, PowerPoint, PDFs, or Chrome. everyday work it covers: → analyze sensitive data → manage private notes → review code and logs → search and organize files → build docs or slides Apache 2.0. no rate limits, and nothing leaves the machine. 𝗻𝗽𝗺 𝗶 -𝗴 @𝗺𝗮𝗴𝗻𝗶𝘁𝘂𝗱𝗲𝗱𝗲𝘃/𝗰𝗹𝗶 the repo is here: (don't forget to star 🌟) i wrote the full breakdown of why picking the configuration is the hard part. the article is quoted below.

run agent harnesses 100% private & offline. (no token costs, no API keys, 100% open-source) your agent runs locally. the model doesn't. every prompt, every file, and every secret still leaves your machine before the agent does anything with it. Magnitude fixes that. it's an open source inference server that runs models on your own hardware and plugs into the coding agent you already use. setup is one command. it profiles your machine, measures the memory bandwidth that sets your token rate, and hands back complete configurations instead of a list of models. each one names a model, a compression level, a context size, and a speed range you can expect. pick one and start working. it doesn't replace your harness. setup asks which one you want and writes that config for you. Pi, OpenCode, Claude Code, Codex, and Cline all work, and there's a built-in one tuned for local models if you don't have a harness yet. that one uses your shell, edits files, and runs scripts out of the box. add skills and it handles Excel, PowerPoint, PDFs, or Chrome. everyday work it covers: → analyze sensitive data → manage private notes → review code and logs → search and organize files → build docs or slides Apache 2.0. no rate limits, and nothing leaves the machine. 𝗻𝗽𝗺 𝗶 -𝗴 @𝗺𝗮𝗴𝗻𝗶𝘁𝘂𝗱𝗲𝗱𝗲𝘃/𝗰𝗹𝗶 the repo is here: (don't forget to star 🌟) i wrote the full breakdown of why picking the configuration is the hard part. the article is quoted below.

Akshay 🚀

55,693 views • 25 days ago

GROK BOT MADE ME $7,600 IN THE LAST 24 HOURS AND COSTS $1.90 A DAY TO RUN the bot is built straight from their own official docs, I didn't write a line of code - just one file, three lines root → scanner → analysis → filter → trade → back to root it reads every market and every headline at once and filters 400 signals a day out of those 400 exactly 1 makes it to a trade, the rest get killed before I wake up the priciest model on the stack gets called once a day, everything else runs on a $0.10 batch X Search eats 53% of the whole bill, because that is where the news shows up before the chart does 12 minutes of checking against a month of building - that is the whole gap between $7,600 and zero the check is 3,600× shorter than what most people do instead of it a 3-line ROOT.md gets read 30 mornings in a row and the bot never asks what to do next write those three lines badly and the machine automates the wrong business perfectly, 30 times save this and paste it into Grok - let it write your ROOT.md ↓

GROK BOT MADE ME $7,600 IN THE LAST 24 HOURS AND COSTS $1.90 A DAY TO RUN the bot is built straight from their own official docs, I didn't write a line of code - just one file, three lines root → scanner → analysis → filter → trade → back to root it reads every market and every headline at once and filters 400 signals a day out of those 400 exactly 1 makes it to a trade, the rest get killed before I wake up the priciest model on the stack gets called once a day, everything else runs on a $0.10 batch X Search eats 53% of the whole bill, because that is where the news shows up before the chart does 12 minutes of checking against a month of building - that is the whole gap between $7,600 and zero the check is 3,600× shorter than what most people do instead of it a 3-line ROOT.md gets read 30 mornings in a row and the bot never asks what to do next write those three lines badly and the machine automates the wrong business perfectly, 30 times save this and paste it into Grok - let it write your ROOT.md ↓

Sprytix

27,231 views • 27 days ago

Production traffic is not uniform. You get a few requests that need your best model, but most are simple questions and lookups you can solve with cheaper, faster models. The most expensive mistake you can make today is sending every request to your strongest model. You need routing. Period. This is the simplest trick to improve the architecture of whatever you are building. Please, don't implement routing yourself. You don't have to. I'm currently working with TrueFoundry's Auto Routing. It reads each request, classifies it as simple, medium, or complex, and sends it to the model assigned to that tier. You have two choices: 1. Send every request to the free heuristic classifier to score signals such as technical vocabulary, code, prompt length, and multi-step reasoning. 2. Send the request to an LLM classifier when its difficulty requires a more nuanced judgment. The beauty of using routing is that nothing changes in your code. You still call a single endpoint model, but routing works behind the scenes to pair every request with the best possible model. TrueFoundry ran several experiments with two different setups: 1. Send every request to Claude Opus 2. Send every request to a router with Haiku, Sonnet, and Opus The first experiment ran 550 deterministically graded academic prompts through every setup. Auto Routing was 69% cheaper while retaining 98% of the baseline quality. The second experiment ran three production-shaped workloads through every setup, using user chats, developer chats, and long agent tasks. Auto Routing was 80% cheaper. Thanks to the TrueFoundry team for partnering with me on this post.

Production traffic is not uniform. You get a few requests that need your best model, but most are simple questions and lookups you can solve with cheaper, faster models. The most expensive mistake you can make today is sending every request to your strongest model. You need routing. Period. This is the simplest trick to improve the architecture of whatever you are building. Please, don't implement routing yourself. You don't have to. I'm currently working with TrueFoundry's Auto Routing. It reads each request, classifies it as simple, medium, or complex, and sends it to the model assigned to that tier. You have two choices: 1. Send every request to the free heuristic classifier to score signals such as technical vocabulary, code, prompt length, and multi-step reasoning. 2. Send the request to an LLM classifier when its difficulty requires a more nuanced judgment. The beauty of using routing is that nothing changes in your code. You still call a single endpoint model, but routing works behind the scenes to pair every request with the best possible model. TrueFoundry ran several experiments with two different setups: 1. Send every request to Claude Opus 2. Send every request to a router with Haiku, Sonnet, and Opus The first experiment ran 550 deterministically graded academic prompts through every setup. Auto Routing was 69% cheaper while retaining 98% of the baseline quality. The second experiment ran three production-shaped workloads through every setup, using user chats, developer chats, and long agent tasks. Auto Routing was 80% cheaper. Thanks to the TrueFoundry team for partnering with me on this post.

Santiago

15,543 views • 16 days ago