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.

Chinese engineering team builds the robot's chip and its control code as one artifact, and neither one is written by hand Most robotics splits the two, a silicon team ships a processor and a software team spends a year discovering what it's bad at. They generate both from the... same movement spec, so the chip knows what motion it exists to serve and the code knows what the silicon can hold: > Describe: the robot's reach, payload, cycle time and safety envelope go in as text > Co-solve: the control loops and the silicon that runs them are generated together > Burn: the chip layout and the firmware come out matched, nothing adapted later • What co-generating actually eliminates > Control code shaped around a latency the hardware never had to impose > Silicon carrying units the motion profile will never once call > The six-month handoff where each team learns what the other assumed • Why this only works generated > A human can't hold both the floorplan and the control math in one revision > A change to the payload re-solves silicon and software in one pass > The tradeoff between them becomes a value, not a negotiation One robot's chip and control stack regenerate from spec in about two hours. And the part that broke my brain: The usual process takes a year and ends with software apologising for the hardware. This one regenerates both overnight and they were never designed apart. Same robot. One version is two teams guessing about each other. The bottleneck was never the chip or the code. It was the wall between them ↓show more

Codio

2,372 subscribers

26,786 views • 13 days ago •via X (Twitter)

Science & Technology

Anya Rossi• Live Now

Private livecam show

7 Comments

Myttle13 days ago

the six month handoff hurts

Kris Watson - Infinity®13 days ago

#bubbles Roberts got a quick hand. They steal bubbles so when you hear Bob Robert or Bobby they are all thieves of something. Quick card his stuff fast.

wast313 days ago

in what volumes is this produced?

Codio13 days ago

in large-scale settings for major corporations

Meh13 days ago

Really sweet looks! I also get cases and housings in my system. And it simulates the electrics and physics. I tell it what parts I want, it makes some pass forms for prototyping. I guess we will see most awesome stuff The next year's if anybody can build this in his shed

æd13 days ago

„The bottleneck was never the chip or the code. It was the wall between them.“

Layla Ibrahim13 days ago

The six-month handoff point is the key one. In autonomous vehicle research we hit the same wall between silicon and software teams. Mismatched assumptions cost more than any technical limit. This removes that wall entirely.

Related Videos

ELON: TESLA'S AI5 CHIP IS THE SILICON VERSION OF “LESS IS MORE” "The AI5 chip will be 40 times better than the AI4 chip – not 40%, 40 times. We have a detailed understanding of the entire software and hardware stack. We are designing the hardware to address all of the pain points in software. We know what the chip needs to do, and we know, just as importantly, what the chip doesn’t need to do. A long list of deletions that are very important. As a result of these deletions, we can actually fit AI5 in a half reticle with a good margin for the traces from the memory to the Tesla chip accelerators, the ARM CPU cores, and the PCI blocks. This is a beautiful chip." Source: Elon Musk, Tesla Podcast, October 22, 2025

ELON: TESLA'S AI5 CHIP IS THE SILICON VERSION OF “LESS IS MORE” "The AI5 chip will be 40 times better than the AI4 chip – not 40%, 40 times. We have a detailed understanding of the entire software and hardware stack. We are designing the hardware to address all of the pain points in software. We know what the chip needs to do, and we know, just as importantly, what the chip doesn’t need to do. A long list of deletions that are very important. As a result of these deletions, we can actually fit AI5 in a half reticle with a good margin for the traces from the memory to the Tesla chip accelerators, the ARM CPU cores, and the PCI blocks. This is a beautiful chip." Source: Elon Musk, Tesla Podcast, October 22, 2025

Mario Nawfal

363,507 views • 1 year ago

A Chinese fab wired Anthropic into its chip pipeline and cut design costs to almost nothing Most of a chip isn't the clever part, and teams still pay senior engineers to redraw all of it every generation. They split the die by who should own it, the commodity half generated and verified automatically, the differentiated half left entirely to humans: > Classify: every block is tagged as commodity, adapted or genuinely novel > Generate: commodity blocks are produced from the house rules, not redrawn > Verify: those blocks are checked automatically against the same rules > Reserve: the novel blocks never enter the pipeline, engineers own them outright • What counts as commodity and nobody admits it > Power distribution, clock trees, scan chains, reset logic > Standard interfaces that have been the same for fifteen years > Test structures that get copied from the last project anyway • Where the savings actually come from > Roughly 70% of the die stopped consuming engineer hours > The same senior people now spend all their time on the 30% that matters > A new generation reuses the rules, so the cost drops again each time The commodity half of a die regenerates in about two hours. And the part that broke my brain: Their best engineers had been spending most of each project redrawing clock trees. The pipeline took that work away and the headcount stayed the same. Same silicon. One version pays senior salaries for solved problems. The expense was never the hard part of the chip. It was everything around it that nobody wanted to own ↓

A Chinese fab wired Anthropic into its chip pipeline and cut design costs to almost nothing Most of a chip isn't the clever part, and teams still pay senior engineers to redraw all of it every generation. They split the die by who should own it, the commodity half generated and verified automatically, the differentiated half left entirely to humans: > Classify: every block is tagged as commodity, adapted or genuinely novel > Generate: commodity blocks are produced from the house rules, not redrawn > Verify: those blocks are checked automatically against the same rules > Reserve: the novel blocks never enter the pipeline, engineers own them outright • What counts as commodity and nobody admits it > Power distribution, clock trees, scan chains, reset logic > Standard interfaces that have been the same for fifteen years > Test structures that get copied from the last project anyway • Where the savings actually come from > Roughly 70% of the die stopped consuming engineer hours > The same senior people now spend all their time on the 30% that matters > A new generation reuses the rules, so the cost drops again each time The commodity half of a die regenerates in about two hours. And the part that broke my brain: Their best engineers had been spending most of each project redrawing clock trees. The pipeline took that work away and the headcount stayed the same. Same silicon. One version pays senior salaries for solved problems. The expense was never the hard part of the chip. It was everything around it that nobody wanted to own ↓

Codio

21,580 views • 9 days ago

your eval measured one system. production ran another. they start from the same idea and end somewhere else, which is why the number was right and the outcome was not. the code is the same until someone rewrites it for prod, because the eval lived in a notebook and prod is a service. then the data diverges. clean fixtures in the test, late and incomplete data in the run. then execution diverges. the eval assumed every call lands. production has retries, timeouts and rate limits. by the end the two runs share a name and nothing else. > why nobody catches it both sides are internally correct. the eval measured what it ran. production ran what it was given. neither one is wrong, and neither one is responsible for checking they were the same artifact. so the gap widens quietly. someone patches prod without touching the eval. someone adds a fixture without touching prod. no test anywhere fails. > what to actually do ship one artifact. the code that was measured is the code that runs. model the friction. whatever the real environment charges you, charge it in the eval too. easiest way to see the gap on your own idea, without writing the harness: (only first 100 registrated) it runs the tested code live, unchanged. that is the whole point. and treat the eval as a prior, not a result. it tells you what to expect. only the live run tells you what happened. your eval was accurate. it was accurate about something you never shipped.

your eval measured one system. production ran another. they start from the same idea and end somewhere else, which is why the number was right and the outcome was not. the code is the same until someone rewrites it for prod, because the eval lived in a notebook and prod is a service. then the data diverges. clean fixtures in the test, late and incomplete data in the run. then execution diverges. the eval assumed every call lands. production has retries, timeouts and rate limits. by the end the two runs share a name and nothing else. > why nobody catches it both sides are internally correct. the eval measured what it ran. production ran what it was given. neither one is wrong, and neither one is responsible for checking they were the same artifact. so the gap widens quietly. someone patches prod without touching the eval. someone adds a fixture without touching prod. no test anywhere fails. > what to actually do ship one artifact. the code that was measured is the code that runs. model the friction. whatever the real environment charges you, charge it in the eval too. easiest way to see the gap on your own idea, without writing the harness: (only first 100 registrated) it runs the tested code live, unchanged. that is the whole point. and treat the eval as a prior, not a result. it tells you what to expect. only the live run tells you what happened. your eval was accurate. it was accurate about something you never shipped.

Hanako

45,239 views • 3 months ago

since everyone woke up talking about loops, this article talks about domain chips, which i think are super underrated compared to loops.. a chip is a small specialist that's only good at one job QA, trading, content, research whatever ... and not one giant brain trying to do everything.. you wrap it in three things: >a workflow (same shape every time, so it's stable) >the tools that job needs >a benchmark that says what good looks what makes this special is the benchmark; without it, a loop just changes things randomly, when with it, the chip can score every pass and keep only what actually wins.. the problem with loops is they can spin 1000× and still not improve when a domain chip not only focuses on the domain you delegate to them, but they actually improve themselves with every spin. you can start with domain chips here this repo will help you build your first chip, the workflow, the benchmark pack, and the autoloop policy, with guardrails so it can't fool itself: below you can read the article fully + i put a good example of the difference between having a normal loop and a domain chip..

since everyone woke up talking about loops, this article talks about domain chips, which i think are super underrated compared to loops.. a chip is a small specialist that's only good at one job QA, trading, content, research whatever ... and not one giant brain trying to do everything.. you wrap it in three things: >a workflow (same shape every time, so it's stable) >the tools that job needs >a benchmark that says what good looks what makes this special is the benchmark; without it, a loop just changes things randomly, when with it, the chip can score every pass and keep only what actually wins.. the problem with loops is they can spin 1000× and still not improve when a domain chip not only focuses on the domain you delegate to them, but they actually improve themselves with every spin. you can start with domain chips here this repo will help you build your first chip, the workflow, the benchmark pack, and the autoloop policy, with guardrails so it can't fool itself: below you can read the article fully + i put a good example of the difference between having a normal loop and a domain chip..

JUMPERZ

36,704 views • 5 months ago

$Elon Musk just put a number on the flaw at the center of Nvidia’s empire. Wall Street has not done the math yet. Nvidia’s Blackwell is the most sought-after silicon on Earth. Every AI lab wants it. Every sovereign nation is bidding for it. Blackwell runs every model, for every company, in every data center on the planet. That universality built the empire. It is also the fracture point. Musk: “We believe the AI5 chip will be about a third of the power of an Nvidia Blackwell for roughly comparable performance. And much less than 10% of the cost.” One-third the power. Comparable performance. Less than ten percent of the cost. Musk: “This is a chip that is very much optimized for the Tesla AI software stack. It’s not meant to be a general purpose chip.” Nvidia builds silicon that serves a million different customers. Every transistor spent on universal compatibility is a transistor not dedicated to one task. Tesla is building silicon for exactly one customer. Itself. When you strip away every function you will never call, you do not get a lesser chip. You get a weapon. Here is what the market refuses to see. Data centers drink unlimited power from the grid. Robots run on batteries. Musk: “In order to have a functional robot, you have to have a great AI chip. And it needs to be an inexpensive chip and it needs to be very power efficient.” You cannot put a Blackwell inside a walking machine. It would drain the battery before it crossed the room. The entire AI revolution lives inside air-conditioned buildings bolted to the electrical grid. Musk is not competing for that market. He is engineering the silicon that survives outside of it. One-third the power is not a spec sheet footnote. It is the physics threshold that severs intelligence from the wall socket. Without that number, every robot on Earth stays tethered. With it, the algorithm walks. Less than ten percent of the cost is not a pricing strategy. It is the line where a machine brain stops being a capital expenditure and becomes a commodity component. When the chip inside a humanoid costs less than the motors in its legs, you do not manufacture hundreds of robots. You manufacture millions. Wall Street is valuing the AI revolution by who dominates the data center. Musk is building the only silicon designed to leave one. Nvidia built the brain of the cloud. Musk is building the brain of the physical world. No one has priced that in yet.$

Elon Musk just put a number on the flaw at the center of Nvidia’s empire. Wall Street has not done the math yet. Nvidia’s Blackwell is the most sought-after silicon on Earth. Every AI lab wants it. Every sovereign nation is bidding for it. Blackwell runs every model, for every company, in every data center on the planet. That universality built the empire. It is also the fracture point. Musk: “We believe the AI5 chip will be about a third of the power of an Nvidia Blackwell for roughly comparable performance. And much less than 10% of the cost.” One-third the power. Comparable performance. Less than ten percent of the cost. Musk: “This is a chip that is very much optimized for the Tesla AI software stack. It’s not meant to be a general purpose chip.” Nvidia builds silicon that serves a million different customers. Every transistor spent on universal compatibility is a transistor not dedicated to one task. Tesla is building silicon for exactly one customer. Itself. When you strip away every function you will never call, you do not get a lesser chip. You get a weapon. Here is what the market refuses to see. Data centers drink unlimited power from the grid. Robots run on batteries. Musk: “In order to have a functional robot, you have to have a great AI chip. And it needs to be an inexpensive chip and it needs to be very power efficient.” You cannot put a Blackwell inside a walking machine. It would drain the battery before it crossed the room. The entire AI revolution lives inside air-conditioned buildings bolted to the electrical grid. Musk is not competing for that market. He is engineering the silicon that survives outside of it. One-third the power is not a spec sheet footnote. It is the physics threshold that severs intelligence from the wall socket. Without that number, every robot on Earth stays tethered. With it, the algorithm walks. Less than ten percent of the cost is not a pricing strategy. It is the line where a machine brain stops being a capital expenditure and becomes a commodity component. When the chip inside a humanoid costs less than the motors in its legs, you do not manufacture hundreds of robots. You manufacture millions. Wall Street is valuing the AI revolution by who dominates the data center. Musk is building the only silicon designed to leave one. Nvidia built the brain of the cloud. Musk is building the brain of the physical world. No one has priced that in yet.

Dustin

160,794 views • 7 months ago

ANTHROPIC LEAKED A FILE WHERE 4 AGENTS BUILD A BUSINESS DOING $100K A MONTH IN 60 MINUTES four agents run around the clock, cost $0 and work against a $100,000 target scout → architect → 2 builders → critic → back to root what makes it a team isn't the four agents, it's the three contracts written between them the scout hands the architect who already ships this and what they charge the architect hands the builders a spec with the interfaces named the builders hand the critic a diff and the tests that cover it leave any one of those boundaries vague and the work goes missing exactly there Haiku 4.5 triages cheap, Opus 5 decides expensive, Sonnet 5 builds, Opus 5 judges the most expensive model runs twice: once to plan, once to pass verdict two builders in parallel is two ways to disagree - which is exactly what the gate is for $100,000 is 334 customers at $299, $50,000 is 167, $10,000 is 34 and what you pay isn't even visible on that same scale the cost was never the thing standing in your way save this and paste it into Claude Code - let it write your 3 contracts ↓

ANTHROPIC LEAKED A FILE WHERE 4 AGENTS BUILD A BUSINESS DOING $100K A MONTH IN 60 MINUTES four agents run around the clock, cost $0 and work against a $100,000 target scout → architect → 2 builders → critic → back to root what makes it a team isn't the four agents, it's the three contracts written between them the scout hands the architect who already ships this and what they charge the architect hands the builders a spec with the interfaces named the builders hand the critic a diff and the tests that cover it leave any one of those boundaries vague and the work goes missing exactly there Haiku 4.5 triages cheap, Opus 5 decides expensive, Sonnet 5 builds, Opus 5 judges the most expensive model runs twice: once to plan, once to pass verdict two builders in parallel is two ways to disagree - which is exactly what the gate is for $100,000 is 334 customers at $299, $50,000 is 167, $10,000 is 34 and what you pay isn't even visible on that same scale the cost was never the thing standing in your way save this and paste it into Claude Code - let it write your 3 contracts ↓

Sprytix

147,555 views • 1 month ago

This is getting out of hand: Claude Fable 5.2 can now take any gadget apart from one picture and nobody is talking about it. I gave it one picture of a PS5 controller and got back one that splits open part by part in 24 seconds. What made it work is teardown, a free open-source skill for Claude Code (3,000+ stars on GitHub) that makes Claude build the inside first. Why model a controller by hand for a week when Claude can write every part as code? Here is the workflow: from that one picture Claude writes the motion as math before it draws a single part, so every panel reads the same numbers as the model. It renders a still at every beat and fixes what it finds, then records the tour at 60 fps: the touchpad lifts, both sticks come out on their gimbals, the motor and gears behind L2 and R2 slide out and trigger force jumps from 30% to 78% the second push-back switches on. Step-by-step setup: 1. Install teardown with one npx line 2. Open Claude Code and type /teardown 3. Give it one picture of the thing you want to see inside 4. Let it check every still before it films 5. Record the tour and watch it come apart It runs inside Claude Code, so you'll never keyframe a teardown by hand again. The full guide and every prompt are in the article below.

This is getting out of hand: Claude Fable 5.2 can now take any gadget apart from one picture and nobody is talking about it. I gave it one picture of a PS5 controller and got back one that splits open part by part in 24 seconds. What made it work is teardown, a free open-source skill for Claude Code (3,000+ stars on GitHub) that makes Claude build the inside first. Why model a controller by hand for a week when Claude can write every part as code? Here is the workflow: from that one picture Claude writes the motion as math before it draws a single part, so every panel reads the same numbers as the model. It renders a still at every beat and fixes what it finds, then records the tour at 60 fps: the touchpad lifts, both sticks come out on their gimbals, the motor and gears behind L2 and R2 slide out and trigger force jumps from 30% to 78% the second push-back switches on. Step-by-step setup: 1. Install teardown with one npx line 2. Open Claude Code and type /teardown 3. Give it one picture of the thing you want to see inside 4. Let it check every still before it films 5. Record the tour and watch it come apart It runs inside Claude Code, so you'll never keyframe a teardown by hand again. The full guide and every prompt are in the article below.

Nyxera

35,505 views • 2 days ago

Anthropic's graphics team automated Blender so one request builds a structure, and most of the run finishes before a vertex exists. The geometry was never the hard half. Deciding what counts as one object is where the time actually goes. > PLAN - what the structure depends on, ordered > SPLIT - where one object ends and the next starts > BUILD - geometry emitted against a finished plan A structure that arrives as one fused mesh is finished the moment it lands, because nothing in it can move. Split it wrong and it is worse. A door welded into its frame looks correct until somebody tries to open it. So the run spends its first minutes on hierarchy, and the modelling that follows is close to mechanical. That is why a single request holds at the scale of a building instead of a prop or one isolated room. What lands in the viewport is a tree of named parts, each one editable without disturbing anything else. The request and the tree it produced sit in the article below.

Anthropic's graphics team automated Blender so one request builds a structure, and most of the run finishes before a vertex exists. The geometry was never the hard half. Deciding what counts as one object is where the time actually goes. > PLAN - what the structure depends on, ordered > SPLIT - where one object ends and the next starts > BUILD - geometry emitted against a finished plan A structure that arrives as one fused mesh is finished the moment it lands, because nothing in it can move. Split it wrong and it is worse. A door welded into its frame looks correct until somebody tries to open it. So the run spends its first minutes on hierarchy, and the modelling that follows is close to mechanical. That is why a single request holds at the scale of a building instead of a prop or one isolated room. What lands in the viewport is a tree of named parts, each one editable without disturbing anything else. The request and the tree it produced sit in the article below.

wast3

17,714 views • 3 days ago

Loops vs. Graphs, clearly explained! loops are great, but they have a ceiling: a loop makes one unit of work better. it cannot decide which units exist. so you end up with a very good agent running the wrong three steps, in the wrong order, one at a time. Graph engineering fixes this by moving the decision up a layer: what runs, what runs at the same time, and what never runs at all. you need both. here's how it works: a graph splits your system into two kinds of decision. ↳ inside a unit: the loop. produce, check, correct, repeat until green ↳ between units: the graph. split, fan out, merge, gate, send back Prompts → Context → Harness → Loops → Graphs you get parallel work, isolated contexts, and steps that stop running when nothing needs them. the trick is being selective about what becomes a node. only spend a model where judgment lives. merging, ranking, deduping and schema checks are edges, and edges are code. free, instant, and they cannot be argued out of a verdict. a graph where every edge is an agent pays rent on its own wiring. one thing to know before you scale it. a graph has two return paths, and almost everyone builds one. ↳ the correction edge is short. a gate rejects one unit back to the step that produced it, and it fixes the run you are in ↳ the learning edge is long. an accepted result goes back to the splitter as a constraint, and it fixes every run after skip the second and you get a graph that is fast and never gets smarter. next week it starts from the same place with the same blind spots. and a smaller one that eats whole nights: when a unit fails, return that unit, not the batch. send back four slices because one failed and you have just rewritten three correct ones. do it twice in a run and the run never converges. 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 ↓

Loops vs. Graphs, clearly explained! loops are great, but they have a ceiling: a loop makes one unit of work better. it cannot decide which units exist. so you end up with a very good agent running the wrong three steps, in the wrong order, one at a time. Graph engineering fixes this by moving the decision up a layer: what runs, what runs at the same time, and what never runs at all. you need both. here's how it works: a graph splits your system into two kinds of decision. ↳ inside a unit: the loop. produce, check, correct, repeat until green ↳ between units: the graph. split, fan out, merge, gate, send back Prompts → Context → Harness → Loops → Graphs you get parallel work, isolated contexts, and steps that stop running when nothing needs them. the trick is being selective about what becomes a node. only spend a model where judgment lives. merging, ranking, deduping and schema checks are edges, and edges are code. free, instant, and they cannot be argued out of a verdict. a graph where every edge is an agent pays rent on its own wiring. one thing to know before you scale it. a graph has two return paths, and almost everyone builds one. ↳ the correction edge is short. a gate rejects one unit back to the step that produced it, and it fixes the run you are in ↳ the learning edge is long. an accepted result goes back to the splitter as a constraint, and it fixes every run after skip the second and you get a graph that is fast and never gets smarter. next week it starts from the same place with the same blind spots. and a smaller one that eats whole nights: when a unit fails, return that unit, not the batch. send back four slices because one failed and you have just rewritten three correct ones. do it twice in a run and the run never converges. 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

73,867 views • 2 months ago

THIS CHINESE DEVELOPER TOLD A TINY AI DEVICE "I NEED A CAR CONTROL PROGRAM" - AND IT WROTE THE CODE, FLASHED THE HARDWARE AND LIT UP THE LEDS IN ONE SHOT small white screen showing "listening..." - he speaks the command out loud - the device hears it, writes the firmware, flashes the Raspberry Pi Pico and the red LEDs fire up in sequence no keyboard, no IDE, no Stack Overflow - just talking to a device that understands what he wants and builds it the breadboard on the desk is the hardware - the tiny screen in his hand is the AI that programs it - voice in, working electronics out he vibe coded the entire control system by describing it - the AI handled every line of MicroPython, every GPIO pin assignment and every LED pattern what used to require an electronics degree and a week of debugging now takes one sentence spoken out loud the people still manually writing firmware are one voice command behind

THIS CHINESE DEVELOPER TOLD A TINY AI DEVICE "I NEED A CAR CONTROL PROGRAM" - AND IT WROTE THE CODE, FLASHED THE HARDWARE AND LIT UP THE LEDS IN ONE SHOT small white screen showing "listening..." - he speaks the command out loud - the device hears it, writes the firmware, flashes the Raspberry Pi Pico and the red LEDs fire up in sequence no keyboard, no IDE, no Stack Overflow - just talking to a device that understands what he wants and builds it the breadboard on the desk is the hardware - the tiny screen in his hand is the AI that programs it - voice in, working electronics out he vibe coded the entire control system by describing it - the AI handled every line of MicroPython, every GPIO pin assignment and every LED pattern what used to require an electronics degree and a week of debugging now takes one sentence spoken out loud the people still manually writing firmware are one voice command behind

Cortex

11,462 views • 5 months ago

Claude Code, Grok Bot and DeepSeek working together as ONE team is the most insane thing you'll see this year... your business can now run on autopilot, this is what it looks like from the inside: - Grok Bot finds what's going viral and makes the content - Claude Code writes the posts, pages and emails - DeepSeek builds the websites and tools - they all share one brain, so nobody forgets what you sell or how you talk - they pass work to each other without you in the middle your only job left is to approve what goes out the full setup is in the article, steal it before everyone else does

Claude Code, Grok Bot and DeepSeek working together as ONE team is the most insane thing you'll see this year... your business can now run on autopilot, this is what it looks like from the inside: - Grok Bot finds what's going viral and makes the content - Claude Code writes the posts, pages and emails - DeepSeek builds the websites and tools - they all share one brain, so nobody forgets what you sell or how you talk - they pass work to each other without you in the middle your only job left is to approve what goes out the full setup is in the article, steal it before everyone else does

Machina

15,145 views • 13 days ago

ANTHROPIC LEAKED 6 RULES THAT LET ONE FOLDER REPLACE A WHOLE AI ENGINEERING TEAM COSTING $3.2M A YEAR folder with 6 files replaced the AI engineers - and the team that spent 4 years writing them isn't needed anymore. file → commit → description matches → loads → runs → leaves a prompt has no history, a folder has a git log - one commit updates every session at once, and every edit shows up in the diff. two prompts fight over one window - two skills just both apply, and nobody merges anything by hand. 8 skills sit on disk, one enters the context - the other seven cost zero tokens until the description matches. a prompt can only ask - the folder carries build py and template.docx inside it, so the deterministic parts stay deterministic. the same skill works in Claude Code, in Cowork and through the API - a prompt lives only in the window you pasted it into. and the main one: a skill leaves the window when it's done - a prompt never leaves. save this - 6 rules, one folder, zero copy-paste ↓

ANTHROPIC LEAKED 6 RULES THAT LET ONE FOLDER REPLACE A WHOLE AI ENGINEERING TEAM COSTING $3.2M A YEAR folder with 6 files replaced the AI engineers - and the team that spent 4 years writing them isn't needed anymore. file → commit → description matches → loads → runs → leaves a prompt has no history, a folder has a git log - one commit updates every session at once, and every edit shows up in the diff. two prompts fight over one window - two skills just both apply, and nobody merges anything by hand. 8 skills sit on disk, one enters the context - the other seven cost zero tokens until the description matches. a prompt can only ask - the folder carries build py and template.docx inside it, so the deterministic parts stay deterministic. the same skill works in Claude Code, in Cowork and through the API - a prompt lives only in the window you pasted it into. and the main one: a skill leaves the window when it's done - a prompt never leaves. save this - 6 rules, one folder, zero copy-paste ↓

Sprytix

11,389 views • 3 months ago

THE HEAD ENGINEER AT ANTHROPIC LEAKED A SETUP WHERE ONE FILE TURNS INTO A FINISHED PRODUCT IN 24 HOURS you hire nobody - you paste a $3.2M file into Fable 5, and the system unrolls itself from there. root → wiring → 4 branches → tests → release → back into the root in the first hour you write 4 things: boundaries, stack, roles, escalation - everything downstream is a consequence of that file. the first 5 hours the system wires itself into tools and only then decides what to build - a team would spread that across two weeks. 8 straight hours run 4 parallel branches with zero meetings - the same block a team fills with standups. by the time you finally open the diff, tests and the critic have already rejected it twice. the changelog is written by the run, not by a human from memory - so it never falls behind the code. your total involvement across the day is exactly 2 hours: the first one, and the one where you say ship it. the 9 hours you were asleep weren't downtime, they were capacity - a queue you have to launch by hand becomes your monday. and the key part: at 05:00 the report rewrites that same root - what shipped, what got rejected, what it cost. without that last line this is just a schedule - with it, the schedule improves itself. save this and paste it into Claude Code - and tomorrow starts from a better file than today did ↓

THE HEAD ENGINEER AT ANTHROPIC LEAKED A SETUP WHERE ONE FILE TURNS INTO A FINISHED PRODUCT IN 24 HOURS you hire nobody - you paste a $3.2M file into Fable 5, and the system unrolls itself from there. root → wiring → 4 branches → tests → release → back into the root in the first hour you write 4 things: boundaries, stack, roles, escalation - everything downstream is a consequence of that file. the first 5 hours the system wires itself into tools and only then decides what to build - a team would spread that across two weeks. 8 straight hours run 4 parallel branches with zero meetings - the same block a team fills with standups. by the time you finally open the diff, tests and the critic have already rejected it twice. the changelog is written by the run, not by a human from memory - so it never falls behind the code. your total involvement across the day is exactly 2 hours: the first one, and the one where you say ship it. the 9 hours you were asleep weren't downtime, they were capacity - a queue you have to launch by hand becomes your monday. and the key part: at 05:00 the report rewrites that same root - what shipped, what got rejected, what it cost. without that last line this is just a schedule - with it, the schedule improves itself. save this and paste it into Claude Code - and tomorrow starts from a better file than today did ↓

Sprytix

20,755 views • 3 months ago

A DEVELOPER PROVED THE REGEX YOU'VE WRITTEN A THOUSAND TIMES IS SECRETLY A COMPILER AND THAT ALMOST NO ONE WHO USES THEM HAS ANY IDEA WHAT ACTUALLY RUNS 36 minutes from Paul Wankadia, the engineer behind a regex engine that compiles your pattern straight down to raw machine code -- walking through what really happens between the slashes. -> The moment it clicks, regex stops being magic punctuation you paste from Stack Overflow and becomes what it actually is: a tiny machine. Your pattern gets turned into a state machine, and that machine is what runs against every character of your text. That one idea explains everything you never understood. Why one regex returns instantly and a nearly identical one hangs your whole server. Why some patterns are safe and others are a denial-of-service waiting to happen. It was never random -- it's whether the machine underneath is built well or badly. Writing a regex was never the skill -> reading one is. And now that an AI agent hands you dense, clever patterns you'd never write yourself, the person who can see the machine underneath is the one who catches the one that takes down production at 3am. Everyone copies regex and prays. This is the talk that ends the praying. Save it. The next time a pattern "Just works," you'll actually know why ↓

A DEVELOPER PROVED THE REGEX YOU'VE WRITTEN A THOUSAND TIMES IS SECRETLY A COMPILER AND THAT ALMOST NO ONE WHO USES THEM HAS ANY IDEA WHAT ACTUALLY RUNS 36 minutes from Paul Wankadia, the engineer behind a regex engine that compiles your pattern straight down to raw machine code -- walking through what really happens between the slashes. -> The moment it clicks, regex stops being magic punctuation you paste from Stack Overflow and becomes what it actually is: a tiny machine. Your pattern gets turned into a state machine, and that machine is what runs against every character of your text. That one idea explains everything you never understood. Why one regex returns instantly and a nearly identical one hangs your whole server. Why some patterns are safe and others are a denial-of-service waiting to happen. It was never random -- it's whether the machine underneath is built well or badly. Writing a regex was never the skill -> reading one is. And now that an AI agent hands you dense, clever patterns you'd never write yourself, the person who can see the machine underneath is the one who catches the one that takes down production at 3am. Everyone copies regex and prays. This is the talk that ends the praying. Save it. The next time a pattern "Just works," you'll actually know why ↓

slash1s

193,635 views • 5 months ago

Elon Musk on Tesla’s AI5 chip: At Tesla, there are two chip programs: • Dojo → training • AI4 / AI5 → inference in vehicles AI4 is what’s shipping in cars today AI5 is being finalized and it’s a massive leap By some metrics, AI5 is ~40× better than AI4 (Not 40%. 40 times.) Why the jump is so big: "This is because we work so closely at a very fine-grained level on the AI software and the AI hardware. So we know exactly where the limiting factors are. And so effectively the AI hardware and software teams are co-designing the chip" • Tesla fixes actual bottlenecks, not just adds compute • Example: SoftMax – AI4: ~40 emulation steps – AI5: a few steps natively in silicon Raw improvements: • ~8× compute • ~9× memory • ~5× memory bandwidth • Native mixed-precision handling But the real gain comes from eliminating core AI4 limitations at the silicon level, turning those raw gains into a ~40× real-world improvement This isn’t just a faster chip, it’s a fundamentally smarter one

Elon Musk on Tesla’s AI5 chip: At Tesla, there are two chip programs: • Dojo → training • AI4 / AI5 → inference in vehicles AI4 is what’s shipping in cars today AI5 is being finalized and it’s a massive leap By some metrics, AI5 is ~40× better than AI4 (Not 40%. 40 times.) Why the jump is so big: "This is because we work so closely at a very fine-grained level on the AI software and the AI hardware. So we know exactly where the limiting factors are. And so effectively the AI hardware and software teams are co-designing the chip" • Tesla fixes actual bottlenecks, not just adds compute • Example: SoftMax – AI4: ~40 emulation steps – AI5: a few steps natively in silicon Raw improvements: • ~8× compute • ~9× memory • ~5× memory bandwidth • Native mixed-precision handling But the real gain comes from eliminating core AI4 limitations at the silicon level, turning those raw gains into a ~40× real-world improvement This isn’t just a faster chip, it’s a fundamentally smarter one

X Freeze

73,645 views • 9 months ago

THIS IS F*CKING OBSCENE opus 5.5 got Fortnite running inside Minecraft and Minecraft running inside GTA 5, with nobody's permission and nobody's documentation three engines, three completely different ways in, and one of them has no way in at all THREE GAMES, THREE DIFFERENT WAYS IN → Minecraft is Java and ships a loader on purpose. that one is the front door. you write against an interface somebody meant for you to use → GTA 5 has no modding api and no front door. you get a compiled binary and nothing else, and the only route in is to read it → Fortnite is a live service with anti cheat sitting on top. there is no way in at all, and that single fact changes the entire job WHAT READING A BINARY ACTUALLY MEANS → no source, no comments, no function names. you recover what code does from what it is, which is the opposite direction from how it was written → the exact build gets pinned before anything else, because one patch moves every address underneath you, and a different version is a different game → nothing counts until it is confirmed inside the running process. launch it, play it, look at it. opus 5.5 does the reading, the pinning and the checking, and that loop is the part that used to be a team and a year WHAT THIS ACTUALLY CHANGES → reverse engineering was the most expensive skill in modding. one person, one binary, months of evenings, and the knowledge died with the project → now it is a loop that runs while you sleep, and what comes out of it is not a guess, it is something that got checked against the game while the game was running → so the hard question stopped being whether you can get into an engine. it is what you do once you are standing in three of them at once one layer was read out of a binary one came through the front door one was never touched at all, and it still had to land right

THIS IS F*CKING OBSCENE opus 5.5 got Fortnite running inside Minecraft and Minecraft running inside GTA 5, with nobody's permission and nobody's documentation three engines, three completely different ways in, and one of them has no way in at all THREE GAMES, THREE DIFFERENT WAYS IN → Minecraft is Java and ships a loader on purpose. that one is the front door. you write against an interface somebody meant for you to use → GTA 5 has no modding api and no front door. you get a compiled binary and nothing else, and the only route in is to read it → Fortnite is a live service with anti cheat sitting on top. there is no way in at all, and that single fact changes the entire job WHAT READING A BINARY ACTUALLY MEANS → no source, no comments, no function names. you recover what code does from what it is, which is the opposite direction from how it was written → the exact build gets pinned before anything else, because one patch moves every address underneath you, and a different version is a different game → nothing counts until it is confirmed inside the running process. launch it, play it, look at it. opus 5.5 does the reading, the pinning and the checking, and that loop is the part that used to be a team and a year WHAT THIS ACTUALLY CHANGES → reverse engineering was the most expensive skill in modding. one person, one binary, months of evenings, and the knowledge died with the project → now it is a loop that runs while you sleep, and what comes out of it is not a guess, it is something that got checked against the game while the game was running → so the hard question stopped being whether you can get into an engine. it is what you do once you are standing in three of them at once one layer was read out of a binary one came through the front door one was never touched at all, and it still had to land right

Roxx

195,977 views • 25 days ago

How to build an AI team in Claude Code — every step One AI employee is Claude Code plus three things on its desk: a job description it reads every time it wakes up, playbooks for the jobs it owns, and a notebook where its lessons go. Build that five times and you have five strangers. The team is everything you add on top. Here is the whole build, in order. 1. A desk for every employee. Each one is its own folder: job description, playbooks, notebook, working files. Nothing leaks between desks, so one glance tells you what each employee knows and owns. 2. The company folder. Three pages every employee reads the second it wakes up: one about you (how you work, what always waits for your yes), one on how the company talks, and the roster with a one-line job for each employee. Write your brand voice once and the hire you make on Friday inherits it on Friday. 3. The shared notebook. Corrections that belong to everyone (a brand rule, a platform policy, a tool nobody touches without asking) go in one notebook the whole team reads. Correct one employee Tuesday, all of them follow it Wednesday. 4. Handoffs. When a job shows up, Claude Code reads the roster one-liners and passes it to the right employee. Employees pass work down the line themselves: spin up a teammate in the background, wait for an answer, or leave a note in a teammate's inbox. You are out of the middle. 5. One board. To do, doing, done. Every employee moves its own cards and submits finished work for your review. One screen tells you who is on what. 6. A clock on every desk. One employee sorts the inbox at nine, another leaves you a trend brief each morning, another checks for approvals every half hour. The team has been at work for two hours before you wake up. 7. Fences. Each employee gets an enforced list of what it can touch (leave writing off the list and it physically cannot write a file) and a short list of what always waits for your yes: sending outside the company, spending money, deleting anything. 8. One employee whose whole job is hiring the others. Say what you need, answer three questions, and a new hire is on the roster and online in your dashboard a few minutes later. The rest of the team knows it exists before you've said a word. How it all moves, for anyone who has never written code: underneath the team sits a set of scripts. A script is a small piece of code that does one chore the same way every time. One builds a new hire's desk and adds the roster line. One carries messages between employees. One wakes each employee on its clock. One files a new playbook onto the right desk. You never open them. They fire when the moment comes. A worker runs a job. A team runs the company. This is how my own operation runs: Quill sorts the inbox, Nova reads the trends, Atlas writes the proposals, Rack keeps the servers alive, all from one shared company folder, one notebook, one board. If context, skills and notebooks are still new words to you, that's who this is for. Seven courses inside my community take you from knowing nothing to building this, in order. The platform itself, the exact code my employees run on, every one of those scripts included, is in there too: download it, open it, start hiring. Join us and start today. The link is in my profile. #aiemployee #aiagent

Professor Glitch

94,800 views • 2 months 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

39,336 views • 2 months ago