The argument in one line.
Claude's five surfaces — Chat, Code, Cowork, Design, and Routines — compound into an autonomous work system only when used together, but most users never leave the default chat window.
Read if. Skip if.
READ IF YOU ARE…
- You use Claude regularly but mostly in the default chat window and suspect you are leaving its most powerful features untouched.
- You have heard of Claude Code or Routines but have not yet set up a working project with either.
- You manage recurring workflows (meeting prep, weekly reviews, content scheduling) that currently require manual effort every time.
- You build or ship software and want to understand how an AI coding agent differs from copy-pasting code from a chat conversation.
SKIP IF…
- You are already running Claude Code with a mature CLAUDE.md, active Routines, and shared Cowork projects — this is an orientation video, not an advanced techniques tutorial.
- You are looking for model-level benchmarks or technical comparisons between Claude versions — this covers the product surfaces, not the underlying models.
The full version, fast.
Claude ships five distinct surfaces — Chat, Code, Cowork, Design, and Routines — and each handles a different class of work. Chat is for thinking and drafting; Code is for building and editing files; Cowork puts shared context in one place for teams; Design generates visual assets without a Figma round-trip; Routines run automated tasks on a schedule or event trigger, 24/7 from Claude's cloud. The video argues that users who treat these as one connected system — and who spend 20 minutes writing a CLAUDE.md and a tight project instructions field — get dramatically better output than users who repeatedly re-explain themselves in fresh chat windows.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 00:25
01 · Five major areas
Overview of Chat, Code, Cowork, Design, and Routines as one ecosystem.
00:25 – 01:09
02 · Projects: folders with memory
Projects persist context across chats — job hunt, fitness, apartment renovation examples. Drop docs in once, stop re-explaining.
01:09 – 01:49
03 · Custom Instructions: constraint sheet
The instructions field is not a wish list. The tighter and more specific, the better every output lands.
01:49 – 02:36
04 · Research vs Deep Research
Basic web toggle returns in 30s. Deep Research reads dozens of sources and returns a structured brief in 5-10 minutes.
02:36 – 03:34
05 · MCP: connecting your tools
Model Context Protocol lets Claude read and write to Notion, GitHub, Google Calendar, Slack inside a conversation. Set up once, use everywhere.
03:34 – 05:20
06 · Lovart sponsor segment
Sponsored Lovart demo: brand kit from one product photo generates full campaign — hero render, social assets, packaging.
05:20 – 07:42
07 · Claude Code: building the gym timer
Builds a rest-timer HTML app from a plain-English brief. Agent creates files, wires preset buttons, adds localStorage — no editor opened.
07:42 – 07:38
08 · Code vs Chat for multi-file work
In Code, renaming a function across an entire codebase happens in one pass. In chat, you hunt for every file manually.
07:38 – 09:14
09 · CLAUDE.md: the biggest unlock
Four sections: what the project is, stack/conventions, test/build commands, protected files. Spend 20 minutes once, save hours weekly.
09:14 – 11:34
10 · Three rules + AI Master plug
Ask before destructive ops. Run tests after every change. Summarize the diff in plain English. Mid-video AI Master course promotion.
11:34 – 12:55
11 · Cowork: shared context
Shared workspace puts pinned docs, constraints, and prior decisions in one place. Japan trip with three people, different travel dates.
12:55 – 13:54
12 · Cowork: concurrent editing
Two people can edit the same conversation simultaneously, like Google Docs. Use when more than one human touches the same body of work.
13:54 – 15:41
13 · Claude Design: no Figma round-trip
Generates slides, landing pages, social posts, motion graphics inside Claude. Token cost is twice chat — plan first, then build.
15:41 – 17:05
14 · Design system from a screenshot
One screenshot of a site you like. Claude extracts full palette, typography, component styles. Every output after that lands on-brand.
17:05 – 18:25
15 · Landing page and pitch deck
Landing page in one generation. Five-slide pitch deck in 2.5 minutes, no font cleanup or color correction needed.
18:25 – 18:51
16 · Routines: local vs remote
Local routines only run when the laptop is awake. Remote routines run in Claude's cloud 24/7. For anything you rely on, use remote.
18:51 – 20:32
17 · Meeting prep automation
Daily 6AM routine: pulls Google Calendar, buckets events, researches external attendees, posts one Slack brief per real meeting.
17:18 – 17:44
18 · Weekly gym recap
Sunday 8PM routine: reads Google Sheets training log, groups volume by push/pull/legs, compares to prior week, emails next-week plan.
17:18 – 17:44
19 · Event-based triggers
Trigger on Slack message, form submission, CRM lead — the trigger is the event, not the calendar. Most real automation work lives here.
17:44 – 19:20
20 · Five biggest mistakes
One per feature: default chat no project, no CLAUDE.md, treating Cowork like Slack, using local routines, vibe designing without a system.
19:20 – 19:52
21 · Wrap + final CTA
Five features, one workflow. Think in chat, build in Code, collaborate in Cowork, produce assets in Design, automate in Routines.
Lines worth screenshotting.
- Custom Instructions in a Claude Project are a constraint sheet, not a wish list — the tighter the constraints, the less work the first draft needs.
- Deep Research takes 5-10 minutes instead of 30 seconds because it reads dozens of sources and cross-references them before it answers.
- Claude Code operates at the file level: it reads and edits files on your machine while you watch, which is why a multi-file refactor takes one instruction instead of dozens of copy-paste cycles.
- A CLAUDE.md at your project root ends the guessing — four sections: what the project is, the stack and conventions, the test/build commands, and the files the agent must never touch without asking.
- Design burns tokens roughly twice as fast as chat; writing and tightening the brief in a regular chat conversation before opening a Design canvas cuts token use by more than half.
- Importing one screenshot into Claude Design's system builder reverse-engineers the full palette, typography, and component styles in one pass.
- Remote Routines run in Claude's cloud around the clock; local routines only fire when your laptop is awake, which is the exact moment you do not need them.
- Event-based triggers — new CRM lead, Slack message, form submission — handle the bulk of real automation work; time-based schedules handle the recurring morning jobs.
- Projects eliminate the daily re-explanation tax: drop the context in once, and every chat inside that project starts from it.
- Three rules belong in every CLAUDE.md: ask before any destructive operation, run the test suite after every change, summarize the diff in plain English before applying it.
Five Claude surfaces work as one system.
WHAT TO LEARN
The gap between a Claude user stuck copy-pasting and one with an autonomous workflow comes down to which surfaces they have actually wired together.
- Custom Instructions in a Project function as a constraint sheet — the tighter the constraints, the more the output arrives in the right shape on the first attempt, with no editing required.
- Claude Code operates at the file level; the agent reads and modifies files on your machine while you watch, which means a multi-file refactor that takes dozens of copy-paste cycles in chat takes one instruction in Code.
- A CLAUDE.md at the project root ends the guessing — four sections: what the project is, the stack and conventions, the test/build commands, and the files the agent must never touch without asking.
- Three lines belong in every CLAUDE.md: ask before any destructive operation, run the test suite after every change, and summarize the diff in plain English before applying it.
- Cowork's value is the shared project, not the shared chat. Pinned documents and shared custom instructions mean every teammate's conversation reads from the same source of truth.
- Design burns tokens roughly twice as fast as chat — write and tighten the brief in a regular chat before opening a Design canvas to cut token use by more than half.
- Importing one screenshot into Claude Design's system builder reverse-engineers the full palette, typography, and component styles in one pass; every output after that lands on-brand instead of on the default gradient.
- Remote Routines run in Claude's cloud around the clock; local routines only fire when your laptop is awake, which is the exact moment you do not need them.
- Event-based triggers — new CRM lead, Slack message, form submission — handle the bulk of real automation work; time-based schedules handle the recurring morning jobs.
Terms worth knowing.
- CLAUDE.md
- A plain-text file placed at the root of a project folder that an AI coding agent reads before every task. It tells the agent what the project is, what conventions to follow, what commands to use for tests and builds, and which files are off-limits without explicit permission.
- MCP (Model Context Protocol)
- A protocol that lets Claude connect to and write to external apps — Notion, GitHub, Google Calendar, Slack — directly inside a chat conversation, without copy-pasting between tools.
- Project (Claude)
- A persistent folder inside Claude that holds uploaded documents, a custom instructions field, and a chat history. Every new chat opened inside a project automatically reads from those files and constraints.
- Remote Routine
- A Claude automation that runs on a time schedule or event trigger in Claude's cloud, independent of whether the user's machine is on or awake.
- Design System (Claude)
- A reusable visual specification — color palette, typography, component styles — that Claude extracts from an uploaded screenshot or Figma file and applies consistently to every Design generation.
- Cowork
- A shared Claude workspace where multiple people contribute to the same project context, files, and custom instructions, so every teammate's chat reads from the same source of truth.
- Deep Research
- A Claude mode that reads dozens of sources, cross-references them, and produces a structured research brief rather than a single chat reply. Takes 5-10 minutes versus 30 seconds for a basic web search.
Things they pointed at.
Lines you could clip.
02:25
“The trick most people miss is that this field is not a wish list. It's a constraint sheet. The tighter and more specific you make it, the better the output gets.”
07:42
“That is what code actually does. It reads and edits files on your machine while you watch every step.”
12:11
“That file is the difference between an agent that guesses and an agent that behaves.”
14:38
“I plan in chat, I build in Design. That single habit cuts my design token use by more than half.”
17:43
“A local routine only fires when your laptop is on and awake, which is exactly when you do not need it.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
00:00Claude has five major areas, and together they cover a huge portion of everyday work. From information analysis and coding to design, team collaboration, and full task automation. In this video, I'll walk you through each one so you can see what Claude is truly capable of and which features you might be missing out on.
00:14We start with chat because it's the layer everything else runs on. I want to lock in three building blocks here and then show you when to hand the work off to the other features. Projects are folders with memory.
00:27I keep one project per ongoing thing in my life. One for my job hunt, one for my fitness plan, one for the apartment renovation I've been dragging through for a year. Inside each project, I drop the documents Claude actually needs, like my resume, a few job descriptions I'm targeting, or the contractor quotes I'm comparing.
00:46From that moment on, every chat I start inside that project already knows what I'm working on. The shift here is subtle but huge. You stop re explaining yourself every morning.
00:57I open the job hunt project, I paste a new job description, and I just ask for a tailored cover letter. Claude already has my resume and the four roles I've been targeting, so the answer lands in the right shape on the first try. Inside each project, there's a custom instructions field, and this is where I install a persona.
01:16For my job hunt project, I tell Claude exactly what level of role I'm targeting, what industries I want to avoid, and that I want cover letters that sound like a human wrote them. For my recipe project, I tell Claude the first cook for two people. I hate cilantro and I want grocery lists grouped by aisle.
01:33The trick most people miss is that this field is not a wish list. It's a constraint sheet. The tighter and more specific you make it, the better the output gets.
01:41Mine runs about a 150 words per project, and that paragraph does more work than any prompt I write afterwards. Claude has two research modes, and most people only notice one of them.
01:54The basic research toggle sits right above the prompt bar. You flip it on, and Claude searches the web in real time before it answers. Useful for quick competitive checks, like what are people saying about Claude versus GPT four o right now.
02:07Results land in about thirty seconds. Deep research goes further. It takes five to ten minutes instead of thirty seconds, but it reads dozens of sources, cross references them, and builds something closer to a real research brief.
02:21I use it when I actually need to understand the space before I commit to a video, competitor landscape, pricing shifts, what a new model actually benchmarks on. The output is a structured doc you can save reference, not just a chat reply. Next up is MCP.
02:36It stands for model context protocol. It is how Claude connects to the tools you already use, Notion, GitHub, Google Calendar, Slack.
02:45So it can read and write to them directly inside a chat conversation. You set it up once under settings, then customize, then connectors, and you pick which apps to authorize.
02:55Here's how that looks in practice. I open a regular chat conversation. I type add this topic idea to my content calendar for next Tuesday, and Claude writes directly into my Google content calendar without me leading the conversation.
03:10No copy paste, no tab switching, Claude just makes the entry. That one connector alone saves me about 10 context switches a day. Use chat for thinking out loud, for drafting, and for any one shot task that fits in a single conversation.
03:25The moment the task needs to touch files on your machine, you move to code. The moment it needs to run on its own without you, you move to routines.
03:34That hand off is the whole gang. Quick break. If you're launching anything visual, you usually pick between two ugly options.
03:41Either you shoot it yourself on a kitchen table at midnight or you ask Claude or GPT and get that plastic AI slot. Stop wasting your time.
03:51Love art sits between those two, and I'll show you on a real case. My friend Maya is launching Echoform, a premium over ear headphone brand this month.
04:00Her starting point is just one product shot on a clean studio background and a screenshot of a brand she loves. I drop both into Lovard's brand kit and it instantly pulls the color palette, typography, and visual direction into one coherent system.
04:15Then I prompt the full launch campaign and Love Art builds the entire brand ecosystem on the canvas.
04:22Instead of generating just a few images, it creates a complete set of launch assets, a hero product render, a lifestyle campaign shot, a website banner, packaging design, a product feature sheet, and other brand materials all following the same visual direction.
04:39Once the core campaign is ready, I expand it into marketing content. I ask Love Art to turn the same brand system into social media assets including Instagram creatives, carousel posts, story formats, and paid ad variations.
04:55Every piece feels connected because they all come from the same campaign foundation, not separate AI generated images. From the hero image, I use Love Art's multi angles feature to control the camera perspective and generate multiple product views from the same shot.
05:12I can adjust the camera position, create a side profile, close-up detail shot while keeping the product, lighting, and brand identity consistent. This gives Maya all the visual variations she needs for her website, ads, and social content without recreating the product from scratch.
05:29Link is in the description with free credits attached. Sign up and run the same flow on your own brand tonight. Now let's get back to Claude.
05:37Now let's move to Claude Code. Let's say I want a tiny one page web app I can actually use at the gym. It's a rest timer between sets.
05:44There are three preset buttons for sixty, ninety, and one hundred twenty seconds. A big countdown sits in the middle and the page beeps when the timer hits zero. I want everything in a single HTML file with plain java drift and CSS so I can just double click it on my laptop and it runs.
06:02I open Claude code in the project folder I just made and I paste that brief. Claude reads my claude.md and asks one short question about the visual theme, then it gets to work.
06:13It creates index HTML and writes a small styles block at the top. It drops in the timer logic and wires up the three preset buttons.
06:22Finally, it hooks up a short beep through the web audio API. The whole build takes about ninety seconds, and I watch every file as it lands. Then I do the thing that makes code feel different from Chad.
06:34I tell it to remember the last preset I used and select it again on the next load. Claude opens the file again and adds two lines that touch Lowther's 10. The page now remembers ninety seconds between visits, and I never had to open a text editor or paste a single snippet.
06:51That is what code actually does. It reads and edits files on your machine while you watch every step. In regular chat, you would have copied a snippet, pasted into a text editor, refresh the browser, and then ask Claude why it broke.
07:04In code, you stay in the chair as the reviewer, and the agent runs the loop. There is one more thing worth calling out here.
07:11Claude code can update dozens of files in a single pass. That is actually one of the biggest reasons people switch from chat to code. You describe a refactor once, say, rename a function or change an API endpoint across the code base, and Cloth finds every file that references it and updates them all.
07:30In chat, you would have been copying, copying, pasting, and hunting for the third place you forgot to change. In code, that whole sweep happens while you watch. The single biggest unlock in Claude code is a file called Claude dot m d.
07:43You drop it at the root of your project, and the agent reads it before every task. Mine is split into four short sections. It says what the project is and what it does.
07:53It lists the stack, the libraries, and the conventions I follow. It tells Claude which commands to run for tests and builds.
08:00It also names the files Claude should never touch without asking first. That file is the difference between an agent that guesses and an agent that behaves. Spend twenty minutes writing yours once and you save hours every week after that.
08:14Three rules I put in every claude.md and you should too. Ask me before any destructive operation like deleting files or force pushing.
08:22Always run the test suite after a change. Summarize every death in plain English before applying it. Those three lines keep the agent useful instead of scary.
08:31Quick note before we move on. AI master is the platform my team built for ourselves, and we teach the exact workflow we use every single day. Inside, you get a full library of courses on the tools that actually matter right now.
08:44There's a deep track on Claude covering chat, cow work, and design. There's an image generation track with lessons on mid journey, nano banana, and g p t image two. And there's a video track covering c dense two, VO, and Cling.
08:57On top of the courses, you get a sandbox to practice on the strongest models the day they ship. Right now, there's 30% off the annual plan, and that's the version most of our students actually use. Head to a master if you want the same setup without piecing it together yourself.
09:13Now let's keep moving through the stack. Cowork is the part most solo creators skip and most teams under use. It's a shared workspace where multiple people share the same project context, the same memory, and the same custom instructions.
09:25Here's the problem Cowork solves. Say you and your partner are planning a two week trip to Japan. You both open claw on your own laptops.
09:32You add a bunch of context about Kyoto temples you partner adds notes about ramen spots and hotels in Osaka. Neither of you sees what the other one added.
09:41You get two half baked itineraries that don't talk to each other. Catwork fixes that by putting the context in one place.
09:48You both see the same notes, the same constraints, and the same prior decisions. Claude builds one itinerary that actually reflects what both of you want. Let me show you a real one.
09:58This project is called Japan trip October, and there are three of us in it. Me, my partner, and her sister who's joining for the second week.
10:06We each have different priorities and different travel dates. Pinned at the top of the project, I have our flight confirmations, a rough budget doc, and a shared list of must do versus would be nice activities.
10:19Every chat any of us starts in here already reads from those three files. My partner can ask for a Kyoto Day plan that hits three temples and a tea house. Her sister can ask for a Tokyo evening plan inside our budget.
10:33I can ask Claude to merge them into one calendar. The other thing co work added recently is concurrent editing. Two people can work in the same conversation at the same time, the way you would in a Google Doc.
10:45In our trip planning, my partner is editing the Kyoto day while I'm tweaking the Tokyo one, and Claude keeps up with both of us in real time. Use co moment more than one human is touching the same body of work.
10:58A couple planning a trip, a small team writing a proposal, a founder and a virtual assistant managing a calendar. Anywhere the context needs to live in one place instead of in everyone's head. Next one is where it gets fun.
11:12Design is the feature that surprised me the most. It produces slide boards, landing pages, social posts, and even short motion mafics, and it does it inside Claude. No Figma round trip.
11:22No separate tool. Before I show you anything pretty, I wanna give you the most practical tip in this whole video. Design burns tokens about twice as fast as regular chat.
11:33Every generation renders a full visual, and that is expensive under the hood. If you walk in and start vibe designing, you can chew through your monthly allowance in a week. So here is the rule I follow.
11:45I plan in chat, I build and design, I write and refine the prompt in a regular Claude conversation first until the brief is tight.
11:54Only then, I bring it into design and generate. That single habit cuts my design token use by more than half. The biggest mistake people make in design is vibe designing without a design system.
12:07You type make me a landing page for my AI tool, modern and clean, and Claude gives you something that looks like every other AI landing page on the Internet. Generic gradient, generic font, generic hero. It's fine.
12:20It's also forgettable. The fix takes about fifteen minutes and it changes every output you ever generate after that. Here's the move that changed how I use design.
12:29Inside the design systems tab, you can import a screenshot of any site you love. Claude reverse engineers the system straight from that image. You get the full palette and the typography back in one pass with the component styles already mapped.
12:42I picked a landing page I actually like the look of. One screenshot, nothing else.
12:48I tell Claude, hey, build me a design system based on this site. That is the entire prompt.
12:53Now I want to use this system on something real. So I'm going to build a landing page for AI master with it. I open a new design canvas, attach the system I just made, and I prompt for a hero section, a feature breakdown, a pricing block, and a CTA.
13:09One generation and it lands close enough that I'd only tweak copy from here. No restyling, no fighting the model on colors.
13:17That is the whole point of front loading the system. Same design system, different output. I asked for a five slide pitch deck for AI master.
13:25Title, problem, solution, traction, call to action. About two and a half minutes later, I have a deck that already matches our brand.
13:33No font cleanup, no color correction, no realigning boxes. This is the part that converted me. Decks used to eat a full afternoon.
13:41Now they take the time it takes to write the bullets. Next one is Claude routines. It is the reason this stack stops being a set of tools and starts being a system.
13:51A routine is Claude running on its own, on a schedule, or on a trigger without you in the chair. First thing to know, there are two flavors. Local routines only run when your laptop is on and awake.
14:02Remote routines run-in Claw's cloud twenty four seven whether your machine is open or closed. For anything you actually want to rely on, use remote. Local is fine for tinkering.
14:11You open the routines tab, you click new routine, you pick remote, You give it a name. You write the prompt, which is just what you want Claude to do every time it runs. You pick the model.
14:21You toggle adaptive thinking on if the task needs reasoning, off if it's a simple fetch. That is the whole setup screen.
14:28Before a routine can do real work, it needs access to your tools. Under more customized connectors, you connect what you use.
14:36Slack, Google Calendar, Gmail, Notion. The connector list is long enough to cover most stacks. You authorize once and every routine in your account can use those connections.
14:46First demo, I want a routine that runs every morning at six and tells me what my day looks like. With research already done on anyone I'm meeting, I call it meeting prep buddy.
14:57Schedule triggered daily 6AM. The prompt does five things in plain English. Pull my Google Calendar for the next twenty four hours.
15:05Bucket events into personal, internal, and external. Skip the personal and solo ones. For external meetings, research the company and the attendees.
15:14Post one Slack message per event to my prep channel with the meeting time, the attendees, and a short brief. I hit save. Next morning at six, before I open my laptop, my Slack has one tidy message per real meeting with the background already done.
15:30That used to be forty five minutes of manual work. Now it costs me zero. And if you wanna go further, you can wire your CRM in as another connector, and the brief gets even richer with deal stage, less contact, and open issues.
15:44Second demo, and this one ties back to the gym thread from earlier. I want a routine that looks at my training week, tells me how I actually did, and suggests what next week should look like. I call it weekly gym recap.
15:55Schedule trigger every Sunday at 8PM. The setup is almost embarrassingly simple.
16:00I log my sessions in a plain Google Sheet, one row per set with date, exercise, weight, reps, and a notes column for how it felt. Nothing fancy, no app.
16:11The Google Sheets connector is already authorized in my Clot account, so the routine can read that sheet directly. The prompt does four things in plain English. Pull every row from the last seven days.
16:22Group the volume by muscle category, so push, pull, and legs. Compare the working weights against the previous week, and flag anything that moved up or stalled. Then suggest a loose plan for next week, respecting that my left shoulder gets cranky on overhead press.
16:38Last step in the prompt is to email me the recap. Subject line is the week number, the body is a short summary on top, then the volume breakdown, then next week suggested loads. I hit save and pick Sunday 8PM as the schedule.
16:51Sunday at 8PM, the email lands in my inbox before I even think about training. I open it on the couch, push volume is up 9% this week, pull is flat. Legs are down because I skipped a session.
17:03Bench moved from 100 to 102 and a half. Overhead press is flagged as stalled with a note to drop the weight 10% and add a set.
17:12Next week's plan is right there, ready to copy into the sheet. Twenty minutes of Sunday night planning, gone, and the routine ran while I was on the couch.
17:22The same shape shows up in a dozen different stacks. A client posts in your shared Slack channel and the routine spins up a task in linear with the message attached. New lead fills out your type form and the routine drafts a welcome email plus a CRM row.
17:37The trigger is the event, not the calendar. That shift is what makes the rest of the stack feel autonomous.
17:44Schedules handle the every morning jobs. Event triggers handle everything else which is most of the work. It's before I wrap, one mistake per feature.
17:53These are the ones I see the most and they are the easiest wins on the way out. Most people live in the default chat and repaste the same context every morning. Spin up a project for anything you touch more than twice.
18:06Drop the docs in once. Write a tight custom instructions paragraph, and every chat after that starts smarter than the last one. If you open Claude code in a repo without a claude dot m d, you are asking the agent to guess your stack, your conventions, and your guardrails.
18:21Spend twenty minutes writing one, four short sections, the three rules from earlier, and the agent stops improvising on you. Teams open co work and treat it like a shared Slack thread.
18:32That misses the whole point. The value lives in the shared project. Pinned docs, shared memory, shared custom instructions.
18:39Put the context in the project once and every teammate's chat reads from the same source of truth. A local routine only fires when your laptop is on and awake, which is exactly when you do not need it. If the routine matters, flip it to remote so it runs in Claude's cloud around the clock.
18:55Local is for tinkering. Remote is for anything you actually rely on. You type modern modern and clean and you get every other AI landing page on the Internet.
19:05Spend the fifteen minutes on a design system first, import a screenshot if you have to, and every output after that lands on brand instead of on the default gradient. Fix Fix any one of these this week and the stack starts paying you back immediately. Five features, one workflow.
19:24You think and draft in chat. You build and ship code in code. You collaborate collaborate with humans in cow work.
19:29You produce visual assets in design. You automate the repetitive parts in routines. That is the system.
19:35And if you want the same workflow without piecing it together yourself, AI master is where we teach all of it with a sandbox on the top models so you can practice the second you finish a lesson. Link is in the description. See you in the next one.
The Hook
The bait, then the rug-pull.
Five surfaces, one system. Most Claude users are running the equivalent of a car with the engine in neutral — the chat window is just the dashboard; the real horsepower is in Code, Cowork, Design, and Routines. This 20-minute walkthrough shows all of them working, and names the single mistake that keeps most people stuck in each one.
Named ideas worth stealing.
07:38list
The Four-Section CLAUDE.md
- What the project is and what it does
- The stack, libraries, and conventions
- Commands to run for tests and builds
- Files Claude must never touch without asking
A four-section markdown file at the project root that the agent reads before every task, replacing the need to re-explain the project on every session.
Steal forAny software project where an AI coding agent is doing the editing
08:16list
Three Non-Negotiable CLAUDE.md Rules
- Ask before any destructive operation (delete files, force push)
- Always run the test suite after a change
- Summarize every diff in plain English before applying it
Three guard-rail lines that keep an AI coding agent useful rather than scary. They prevent silent damage, catch regressions, and give the human one more checkpoint.
Steal forAny CLAUDE.md or system prompt for a coding agent
11:40concept
Plan in Chat, Build in Design
Write and refine the design brief in a regular Claude conversation until it is tight, then open Design and generate. Design burns tokens twice as fast as chat, so this habit cuts token use by more than half.
Steal forAny workflow using Claude Design or other token-heavy generative design tools
19:20model
Five-Surface Claude Workflow
- Think and draft in Chat
- Build and ship in Code
- Collaborate with humans in Cowork
- Produce visual assets in Design
- Automate the repetitive parts in Routines
A one-sentence-per-surface mental model for routing work to the right Claude surface instead of defaulting to chat for everything.
Steal forTeam onboarding, personal AI workflow documentation
How they asked for the click.
VERBAL ASK
19:20product
“If you want the same workflow without piecing it together yourself, AI Master is where we teach all of it with a sandbox on the top models.”
Second plug for AI Master. First plug at approximately 8:15 with 30% off annual. Both plugs are tightly integrated with the content — the gym timer and calendar automation demos directly reflect AI Master course topics.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
