The argument in one line.
Claude Code is the best-value AI coding assistant available because its $100/month plan provides unlimited usage, superior context awareness, and granular control through prompting that outperforms competitors despite lacking native Windows support.
Read if. Skip if.
READ IF YOU ARE…
- A developer who codes primarily on Mac or Linux and wants to evaluate Claude Code's real-world productivity gains through a power user's month of data.
- A Windows developer using WSL who's curious whether Claude Code's pricing justifies switching from free or cheaper coding assistants.
- A solo developer or small team lead comparing coding assistant costs across Claude, Cursor, Codeium, and Augment Code for budget planning.
SKIP IF…
- You code exclusively on Windows without WSL — the video confirms Claude Code has no native Windows support and the speaker notes this as a genuine blocker.
- You're already deep into open-source coding assistants or have philosophical objections to vendor lock-in — the speaker shares this concern but ultimately endorses the proprietary tool.
- You're a hobbyist looking for free coding assistance — the cheapest viable option discussed costs $20/month with severe limits, and the speaker's actual workflow costs significantly more.
The full version, fast.
Claude Code earns its keep as a daily coding workhorse because prompting style, not raw model power, dictates the quality of its output. The creator's method treats it like a controllable agent: vague exploratory prompts trigger its REPL-style code search, while keywords like "ultrathink," "architect," or "plan" force deeper reasoning, and the /init command builds a CLAUDE.md map that compounds context across sessions. The practical playbook is to start on the $20 Pro tier, run /init on every repo, clear context aggressively rather than grinding bad threads, revert instead of patching tangled output, and orchestrate parallel work through git worktrees. Windows users should expect to operate inside WSL Ubuntu for full compatibility.
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 – 01:06
01 · Cold open — 318 commits in May
Concrete number hook. Establishes he's a heavy user, mostly Sonnet 4. Frames Claude Code as 'phenomenal' but flags the Anthropic lock-in as his one hesitation.
01:06 – 02:37
02 · OS support — Windows + WSL reality check
Linux/Mac out of the box; he runs Ubuntu via WSL because he came up in the game industry on Windows. Self-aware that this isn't 'normal' for devs.
02:37 – 04:00
03 · Pricing and plans
Just-announced: $20 Pro now includes Claude Code (limited). $100 Max 5x is his recommendation. $200 Max 20x he's leaving because he never hits limits.
04:00 – 11:02
04 · Prompting that matters
Trigger words that change behavior: 'ultrathink', 'architect/plan/implement', 'don't write any code — ideate with me'. REPL behavior + /init to generate claude.md. Don't be afraid to start over.
11:02 – 15:56
05 · Key commands
/init (codebase map), /clear (new context), /model (Opus vs Sonnet 4 — Sonnet daily, Opus rarely), /compact (auto), /cost, /status. Everything beyond is advanced.
15:56 – 18:52
06 · Orchestrating tasks — git worktrees
His preferred way to run multiple Claude Code instances in parallel: git worktree add path branch, cd in, run claude. Notes WSL still hurts when tools assume Linux-native.
18:52 – 22:11
07 · Impressions — positive plus caveats
Negatives: tabs-vs-spaces merge deadlocks, weird spacing states — just nuke and retry. Positives: it's fast, it's fun, prompt-as-control surface gives you real agency.
22:11 – 29:02
08 · Live demo — /init plus diagram Roo Code
Opens VS Code, runs /init on a repo. 26 tool uses, 90.6k tokens, 2.5 min, 149-line claude.md. Then asks it to ultrathink + mermaid-diagram Roo Code's orchestrator/subtask system.
29:02 – 30:09
09 · Agent ranking — his subjective top 3
References his previous ranking video: Claude #1, Augment #2, Roo Code #3. Notes he's burned through Cursor, Klein, Augment as 'main tool' over the last year.
30:09 – 32:33
10 · Next steps — his actual monthly stack
Reads his stack line by line: $100 Claude Max 5x, $30 grandfathered Augment, $200 Codex (may downgrade), $200-1000 API for evals on Roo Code. Google has 'fallen off.'
32:33 – 33:11
11 · Outro — invitation, not CTA
'Anything you want me to go deeper on? Let me know your thoughts below.' Notes he's never hit the $200 limit but has heard about $100 limits. Peace out.
Lines worth screenshotting.
- 318 commits in a single month using Claude Code is operational proof that the tool supports serious production volume, not just demos and experiments.
- Claude Code's flat monthly pricing converted AI-assisted coding from a usage-anxiety experience (burning $20-30 per day on the API) into a fixed predictable expense.
- WSL on Windows is the correct workaround for Claude Code's Linux/Mac-native design — Ubuntu on WSL covers 80-90% of use cases with adequate RAM.
- Typing 'ultra think about this and implement the following' or 'architect, plan, and implement' produces qualitatively different (more thorough) outputs than a plain implementation request.
- Claude Code's REPL mode — telling it to ideate without writing any code — creates a deliberate planning loop that separates architecture decisions from implementation.
- The /init command creates a CLAUDE.md by scanning the existing codebase, including other AI tool configuration files like Cursor rules and GitHub Copilot instructions.
- Using the @ symbol to reference a specific file and tab-complete the path is the fastest way to constrain Claude Code to a targeted change rather than a codebase-wide intervention.
- Running multiple Claude Code instances simultaneously is supported on the $200/month Max plan — enabling parallel work streams without hitting usage limits.
- Vague prompts work better than expected in Claude Code because its REPL-style context awareness compensates for ambiguity in ways a one-shot chat interface cannot.
- The Pro plan ($20/month) now includes limited Claude Code access — enough to evaluate whether the tool fits your workflow before committing to a higher tier.
- Locking into the Anthropic ecosystem is a real tradeoff that GosuCoder names explicitly — the performance advantage must be weighed against the vendor dependency risk.
- The $100/month Max 5x plan is the correct intermediate tier for developers who want to test multiple coding tools without feeling forced to use Claude Code to justify $200/month.
Steal the outline-canvas format.
JoeFlow / Mod Boss monthly recap playbook
One hand-drawn canvas + one webcam + one month of real usage numbers = a 30-minute video that watches like a friend explaining their stack.
- Open with a concrete past-30-days number. '318 commits' beats 'In this video we'll cover...' every time.
- Use a single outline canvas as your visual spine — section headers in red marker, scroll as you go. Replaces a slide deck, teleprompter, and B-roll in one move.
- Demo the thing on screen at least once. /init costs 26 tool uses + 90.6k tokens. That single live demo earns the entire 'impressions' section.
- End on a transparent stack reveal with real dollar amounts. Don't hide your costs — it's the most screenshotted slide in the whole video.
- Flag your own bias out loud ('I know Windows is not normal, but it's what I use'). Immunizes you against the obvious counter-comment and reads as honesty.
- Replace 'smash like' with a content-request question ('what should I go deeper on?'). Doubles as comment-volume driver.
Terms worth knowing.
- WSL (Windows Subsystem for Linux)
- A compatibility layer built into Windows that lets users run a Linux environment natively without a virtual machine — required to run Claude Code on Windows because the tool only officially supports Linux and macOS.
- REPL (Read-Eval-Print Loop)
- An interactive programming environment that reads a user's input, evaluates it, prints the result, and repeats — used here to describe Claude Code's mode of answering questions about a codebase without writing or changing any files.
- CLAUDE.md file
- A markdown file in a project's root directory that Claude Code reads to understand the codebase structure, tech stack, conventions, and instructions — generated by the /init command and used to improve context-aware responses.
- git worktrees
- A Git feature that allows multiple working directories to be checked out from the same repository simultaneously, each on a different branch — used here to run multiple Claude Code instances in parallel on separate feature branches.
- compaction
- Claude Code's automatic process of summarizing and condensing earlier conversation history when a session's context window fills up, allowing work to continue without losing important context.
- Mermaid diagram
- A text-based diagramming format where code is written in a simple syntax and rendered into flowcharts, sequence diagrams, or other visuals — Claude Code can generate Mermaid syntax that a renderer then displays as a graphic.
- evals (evaluations)
- Automated test suites that measure AI model output quality across a defined set of prompts and expected responses — used by developers to benchmark different models and track performance regressions.
- RootCode
- An open-source AI coding agent that can orchestrate sub-tasks and call external tools including Claude Code, positioned here as a complement to Claude Code for multi-agent workflows.
- AugmentCode
- An AI coding assistant with strong codebase indexing and retrieval capabilities, described here as particularly useful for planning and ideation before implementation tasks are handed to Claude Code.
- API costs
- Per-token charges incurred when calling an AI model directly through its API rather than through a flat-rate subscription plan — relevant for evals, testing pipelines, and tools that don't bundle model access into a fixed monthly fee.
Things they pointed at.
02:40linkClaude Code on Pro plan announcement
09:50toolAugment Code
11:40toolGitHub Copilot
15:00toolCursor
15:50toolKlein
29:10toolWindsurf
29:15toolTrae
31:10productOpenAI Codex
33:05toolGoogle Gemini
Lines you could clip.
00:00
“In the month of May, I did 318 commits with Cloud Code.”
18:52
“It is just fun. Everything about it just exudes fun.”
21:20
“You can just reword a problem that other agents have trouble with, and you word it in such a way with Claude Code, and it'll freaking nail it.”
29:02
“Claude four Sonnet, very specifically Sonnet, does really good in Cloud Code. And it should — it is their tool that they're building.”
31:20
“There's something really special about Codex being able to be on a bike ride, get something popped into your head, pull over, kick something off, bring it back, and work on it in Claude Code.”
33:00
“Google has kind of fallen way off.”
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.
metaphorstory
00:00In the month of May, I did 318 commits with Cloud Code.
00:05Now I did have some with some of the other ones, like, used a little bit of codecs, root code, and augment code, but the vast majority of my commits and the vast majority of the number of lines of code that I produced or refactored came from Cloud Code. It is phenomenal.
00:21When I first started using Cloud Code back when it came out, I actually liked it back then too. It was early, but it had cost way too much money for me to be able to justify using this on a daily basis.
00:34Now Cloud Code, it's a CLI based coding assistant that is really locked into the Anthropic ecosystem.
00:41So it's important to note, I'm a little bit nervous about being locked into ecosystems, which is why I still really use and really love RootCode. And I think open source really is where I wish things like this would actually be.
00:54But I think there's something to argue with how good CloudCode actually works and the fact that they can give you this monthly fee to allow you basically unlimited unlimited use.
01:06So for OS support, this is what's a big problem for me and still is today. So they out of the gate support Linux and Mac. Now the reason that this is a problem for me is I am primarily a Windows user, and I know a lot of you would be like, why?
01:23Why are you using Windows? And part of the problem with that is when I started programming years ago, I started in the game industry.
01:31And everything you do in the game industry with Windows based, that's just become kind of my native operating system. And on top of that, I actually do games, so there is a dual purpose there. So I I do have a Mac.
01:45I do dual boot Linux. I have a Linux computer to my right, but I prefer working on a Windows computer.
01:51I know that is not normal or popular, but that is what I prefer. So what I end up having to run for Windows in particular is I actually have to run WSL. Now WSL, it's actually gotten a lot better, And I'm actually using Ubuntu with WSL.
02:10I know some people use Debian or some other flavors of it, but Ubuntu has been really good to me. And I can do 80 to 90% of everything that I need to with WSL, and I have enough RAM and system resources to where that doesn't actually matter.
02:25So I'm just calling that out. If you are a Windows user, you are going to probably have issues unless you are willing to set up WSL, and I know some people are resistant to that.
02:35So that's just worth noting. Now on the pricing side, as I was putting together this video, I actually saw that Kat Wu posted this.
02:43You can now actually use Clogcode a little bit on the pro plan. So what does that actually mean? We've got the pro plan at $20 a month.
02:50Now this plan is what I was on for the longest time. And this is typically for users that are just working in the web UI, but now you can actually come in and at least try out Cloud Code.
03:02I'd recommend doing this. This is a massive, massive thing for them. What kind of limits you get?
03:08I bet they're gonna be very tight, to be honest with you. So I wouldn't expect to be able to do a lot, but you can at least see how it works for you. Then they have a $100 a month plan, and this is the max five x plan.
03:21I've actually just switched to this for this upcoming month because I really do wanna be able to use some of my other coding assistance a bit more and not feel like I'm paying $200 a month and I have to use ClawCode.
03:36I do love it, though. And I if I wasn't in the position that I'm at where I wanna test a bunch of different coding tools and things like that, the $200 max, I think it what is it?
03:4720 x or 10 x? I can't remember exactly. But it's some x in here.
03:52I never hit a single limit on it, and I literally was running multiple instances of clogged code. So just something to note here, these plans are phenomenally priced.
04:03Because when I was early on using Cloud Code, I would burn $20.30 dollars in a day, and that wasn't even me, like, aggressively using it, just using the API.
04:13So just noting that that it's probably worthwhile considering jumping up to the $100 a month plan if you find yourself liking CloudCode. But try it out on the $20 a month plan because this is an exciting exciting update for the people that haven't had a chance to test it out yet.
04:32Alright. So jumping over to the next topic here that I wanna talk about, and this is one that kind of threw me off a bit when I first were getting going, but I've come to really appreciate it.
04:44And that's really that prompting matters a ton. So you can actually guide Claude to I d ideate with you.
04:54You can guide Claude to think very hard Claude code, to think very hard about something. I work very hard about it or make a checklist. And it's just all about how you prompt it.
05:04So a few things that really changed it for me, and I'm not lot of you probably know this already, but I literally will type in ultra think about this thing, whatever it is, and implement the following.
05:17You can also do things like architect, plan, and implement the following.
05:23Things like that make a huge difference in, like, getting it to do what you want it to do. And you don't have to use UltraThing because sometimes you may just want something basic to be done.
05:34So you just say, you know, do this thing. You can also say, don't write any code. I want to ideate with you.
05:40And this is important because, uh, what you wanna do in certain cases is think through a plan and kinda figure things out. I typically use augment code for this, just to be totally honest.
05:52But I have found myself doing this more and more with Cloud Code because I find that that is just it's actually pretty good at it at times. The other thing that's really interesting is REPL. And so, basically, what you can do is you can do something like, um, tell me tell me about my code base.
06:12And it will create, like, a read, evaluate print loop. It's what REPL stands for.
06:18And it does a really, really good job answering kinda context specific questions.
06:25Try it out, especially if you're using the pro plan. I was very impressed with its ability to pick up context of your code base.
06:34And it gets better if you run the init command and create a cloud dot m d file. And this is something that you can keep up to date.
06:43But, basically, what claud will do is go through your code base and build it so sort of, like, information, kinda like a map of your code base that it can then use. What's also interesting about this is it will search for other, uh, AI tools and see how you prompted those.
07:00So it'll say, like, do you have GitHub Copilot instructor? Because I found mine when I was doing it, and it kinda built, uh, used some of that stuff. It searches for, like, cursor rules and so on.
07:10So it knows about the other files that it's looking for there when it's when it's kinda building out that rule set. So I would I would learn to talk to it in the way that works for you. I I started out thinking that I needed to be, like, very specific.
07:25So I was like, you know, you basically you can use at to go, like, chat.ts.vu, and then you can tab into to kinda complete that so that it picks the reference to that file.
07:37And I would say change this specific thing. And it's really, really good when you do this.
07:45And if you got a high enough level or a strong enough level understanding of your code base, this is the quickest way to get things done in your code. But you can also be very vague.
07:58And this is the thing that surprised me the most because its REPL loop does pretty dang good at figuring out what it needs to do.
08:08So I can say something like, you know, in my orchestrator mode, I want to add a new tool, blah blah blah.
08:18So, basically, I'm not giving it files. I'm just saying go do it.
08:22And, know, 80% of the time, it'll find what it needs to and make some good, uh, good changes there.
08:31The other thing that I would say is it's important, uh, on the prompting side to not be afraid to, uh, to kind of start over. Because sometimes you prompt it, and you realize when you reread it that you were too vague.
08:47Or, actually, it answered the question, but you had meant something different. Because a lot of times I get busy, and I'm like, oh, no. No.
08:53No. No. And I looked back over there.
08:56I'm like, oh god. I woulda had no idea. I'm surprised it actually got as far as it did.
09:00So don't be afraid to start over and just revert and do it again. Uh, because a lot of times, I find that people feel like they have to keep grinding in the same direction, but don't do that. Like, get rid of it and and start over.
09:14So now what I wanna do is I wanna go through some tips that I've learned that have actually really helped me a lot with, uh, using Cloud Code. The first thing is if I want to do something like I kick off a task let me do that right now.
09:28So in this one, I'm actually going to ask it to ultra think through refactoring a file. I can also type in stuff while it's syncing. Please make sure all functionality remains intact.
09:42Also, make sure that you break out components when possible.
09:48So you can send it messages while it's thinking. This is something that's very it's kind of an awesome feature that I didn't know about until a few days after they release it.
09:58It's a fairly new feature. But it's worthwhile knowing that this exists because it will allow you to think about, oh, I should have told it this.
10:08And now you can actually send it, and it'll do a good job kind of pushing that in. The other thing that I would say is don't immediately just go to yes and don't ask again in the session.
10:19Make sure you're comfortable with the direction that it's going before you kick that off.
10:26Um, and then use this no and tell Claude what to do differently a lot. This has saved me so many times when it is unclear or going away that I don't like. I'll say something like, no.
10:39I'd rather you focus on a different part. You know, this this is all too vague right now, but you get what I'm saying here. It's like you can actually redirect it if it's going going awry there.
10:51Then I just wanna be like, alright. I'm gonna clear this. Let's get get rid of that.
10:54We're starting over with a new context here. So you can also use at to get a list of files. So I can see if you hit at here and I start typing chat.ts.view, I can hit enter, then the path will be be filled in there.
11:09So in this in this query here, I'm telling it in a particular file. I'm telling it the problem.
11:16In this case, I don't have an idea what the prob like, how to actually fix it, so I'm asking it to give me ideas on how to fix it. And because I've directed it into that file, it's going to be a lot better at helping track down that problem there.
11:30So that's a it's something that highly, highly recommend that, and it's really, really helpful if you know the area of the code that you wanna work in. But you don't have to be that specific on it.
11:41You can be very vague, but you should also be kind of specific too if possible. So you can be you can target functions.
11:49You can target files. You use things that are greppable or searchable. It's really good at that, very similar to the way Codex works.
11:57And so in this case here, uh, it's adding some debugging logs because it doesn't really know it's trying to get some information on it. I think this is fine.
12:05I'd probably accept this and kind of continue on there. But if I if I want to be super vague on it so here, I'm gonna ask it, can you walk me through writing code about how to add a new tool like multi image? But this time, I want it to be search and replace for brief chat.
12:21Now what this is going to do, I didn't give it a file to start in. I did give it some things that are greppable, but I wouldn't even have to do that.
12:30It usually will do a pretty good job finding it. And I've actually tested this prompt a few times, and this one actually will return really, really good results.
12:39So if I take this same prompt here, and I'm going to open up AugmentCode over here on the right, and you can see some of the stuff that I do in there.
12:49I'm gonna put it in chat mode, and I'm just gonna say, basically the same question. And we're gonna take a look at kind of what it comes up.
12:59So this one at said to create a new search replace tool handler, which is awesome. And, yep, that's exactly what it's talking about here.
13:07It's a very, very similar similar response. It's very, very snappy for each.
13:13So for this longest time, I actually just kind of the first couple weeks I used Cloud Code, I kind of ignored its ability to pick up context of my code base.
13:24Don't do that. Like, if you're not paying for Augment code, Cloud Code does a really, really good job, uh, getting context of your code base.
13:35And whatever they're doing under the hood here is incredible. So that that's kind of just some of the tips that I picked up kind of running through this. Now on the command side, there's only a few that I think are, like, very important.
13:48The omit one we talked about, I'd recommend trying this. It's impressive the what it puts together, and the CloudMD file does help, uh, in my experience.
13:59On an existing code base, it does help. Now what this will do is this will actually generate a Cloud dot MD file, and I can show you some examples of that here in a minute.
14:08I'll actually pull over to to Versus Code and just run this in my one of my repos. I've cleared it out already. I'll show you what that looks like.
14:16Now the other command is clear. This is one that's very important because there's sometimes you're just done with the you're done with the context that you're in.
14:26So, like, wipe it out and start a new context. Don't keep it going forever. Don't be like one of the people that keep a chat window going forever and don't realize that you should just clear your context.
14:37So clear it When you're done, there's also models.
14:42So you can actually switch switch what models are active. This can be things like Opus or Sonnet four. Sonnet four is what I run on a daily basis.
14:50I very rarely switch to Opus because that eats into my limits very, very highly. The other thing that they have is they have a compact command.
15:00This is one where it will it will auto run for you. You don't really need to worry about this a lot of times, so you can just let this kinda do its thing.
15:09Right? If you are running via the API, there is, like there's a cost command and there's a status command, and then all the normal ones, like log in and log out.
15:20These are the ones, if you're just getting started, that you should just get comfortable with, specifically, a net clear, checking your or changing your model, compacting.
15:30Beyond that, everything else becomes kind of advanced and awesome. Right? Like, things that you can run a cloud code via a deploy pipeline.
15:40So you can or you could have it set up so that it runs on where with no input. So just auto runs for you.
15:47So there's a lot of, like, really advanced stuff you can do, which I've been playing around with. And I think it's important just to keep that all that in mind. Now orchestrating tasks, this is something that I actually really am excited about.
16:01Because we have a CLI tool, if you think about it, it becomes very easy to specifically take and send out bite sized jobs.
16:11And so in root code, what I actually did is I created a, uh, a clog code mode that literally would break out tasks, and it would send CLI command to clog.
16:25And this is great, and it works great some of the time. I honestly found it more kind of a pain than anything trying to it go between root code and, uh, cloud code.
16:36But there are tools starting starting to pop up where you can manage multiple cloud code instances. One of the problems I run into a lot of times, a lot of them don't support WSL very well.
16:48So, again, when I have to go over to Linux to try them out. But I do locally the way I do a lot of orchestrating of tasks. I do that with Git Worktrees.
16:57And git work trees is something that's actually really, really, really powerful. Or, I mean, simply, you can actually just, uh, run multiple instances. I've never ran into an issue having two or three instances running at a time, and I typically run them in the terminal window of my Versus Code.
17:16So it's just all right there. I know some people might prefer to run that outside of Versus Code, but I I'm traditional IDE guy. I like it being a Versus Code.
17:24It just it's so nice having it integrated in there very well. On the Git Worktree side, what it typically does is you can actually set up sort of, like, feature branches in sort of this tree structure.
17:39So there's these things like Git Worktree add. Right?
17:43So you can do git work tree. You can list them. You can do git work tree.
17:48I think it's remove. There's some other commands too. I think prune is one.
17:53I'd have to go through and and look at all of them. So what you do here is you go git work tree add. Here's your path to where you wanna add it to have the the repository sort of synced to, and then this would actually be your branch.
18:10And now you can just c d into that path and run plot. And I can run that and have multiple instances working on different feature branches very, very easily, and then just switch over to the run them.
18:27I still run into the same issues that I have locally with my other type of multi agent systems, which is I have a I have a server that needs to be booted up, and this is kind of, uh, probably kind of whining more than anything.
18:44But, basically, I have to stop and start the server when I switch between those instances. So it's not the most ideal thing, uh, where, ideally, I have that could be a little bit quicker for me to actually be able go in and test it.
18:58It's again, it's just me kinda whining about it more than anything. So on the impression side, I just wanna say it's very positive. I have very few negative things that happen.
19:08What I will say on the negative side is there are times that you end up with situations where, um, spacing or merging or weird states you get into it where you can't get clogged code to fix it, you're better off just nuking it or going into manually fixing it.
19:32Trust me, I've spent enough time with clogged code now to basically be able to pick up instantly when it's in a position where it's not gonna get out of it. Sometimes it's it's just simple with, I know this just gets on the but tabs versus spaces, it will, uh, get into a state where it just can't merge things because it's trying to do one or the other, and it just is really weird.
19:55It's very rare that these things happen. But, again, it's just you'll you'll pick it up as you use it that if as long as you're committing and you're keeping the progress of things you're working on well enough, you're not gonna feel bad reverting whatever change it's got or go in and help it sometimes.
20:13Uh, but on the positive side, it is just fun. Like, I'm not kidding you about how much fun it is to use. I it's everything about it just .
20:26And I know it's it's coding, and a lot of people would be like, uh, why is that fun? But it really is.
20:33I I feel like it's so good at what it does, and there's so much nuance to how you kind of guide it that you feel like you have control of it.
20:48Whereas some of the other AI agents, you have some control, but not to the point that clog code is.
20:54I don't need to change modes. I basically just guide it based on how I prompt it to behave from this ultra hardcore, like, really deep thinker, task list creator, all the way down to, I just wanna talk to you about an idea that I've got.
21:11And there's all the variance between there that just makes you feel like you have control over this agent unlike other agents. And you may find that you can just reword a problem that other agents have trouble with, and you word it in such a way with ClawCode, and it'll freaking nail it.
21:31The other thing that I wanna say from the impression side is that it's fast. Like, it it is never it never feels to me like it's a slog.
21:43So a lot of times, you know, with some of the other agents, especially GitHub Copilot, one of my big complaints about that is how slow it is. It it literally takes, like, triple the time some of these other agents.
21:54But Cloud Code is fast. You can make it faster, and you can make it slower by how you prompt it, which is also what I love about it, which goes back to sort of the control that you have over it. Um, until you use it, you really would have a hard time understanding what I'm saying about that.
22:09But it is it is a real thing that I feel. I'm jumping over into Versus Code because I wanted to show you what this looks like in practice here. So the first thing that I would do is you can see here that I've been working with ClawCode quite a bit today.
22:22Now I have nuked my particular claud dot m d file.
22:29And as you can see here, there's actually an update ready. So I'm gonna go ahead and restart claud. That was weird.
22:35I actually got a weird error here at first, which I've never got before, so that was kinda interesting. Regardless, I am in Clog.
22:43The first thing that you see when you come in is this what's new.
22:48It's such a such a nice little touch. I know it's not much, um, but you've got this Clog code can now be used with Clog Pro subscriptions.
22:57You also have this upgrade, which is pretty dang sweet.
23:03That is a cool feature that they've added, uh, where before I logged out, I had to log back in to kinda link my max plan. And I can tell you right now the to do list stuff has actually been handled very well after compaction.
23:17I their compaction feature is state of the art. Like, it's incredible. So now let's just run a net here, and I'll kinda show you what this looks like and kinda talk through what it's doing.
23:27Now this actually may take a while, and I may have to speed up some of this. It's looking at the entire code base structure.
23:35It's reading some lines of code. You could see sort of the style that it adds. Like, we've got this actualizing.
23:43It keeps track of the number of tokens. There it's really informative that things are happening. This usually takes a few minutes.
23:51So what I'm gonna do here is I'm just going to periodically talk about what's happening, but mostly skip over a lot of this. So now we're at the point where it has done the analysis of my code base, and I'm gonna go ahead and say, yes.
24:04You can create that file. You can kinda see what it asks you. The second option is yes, and don't ask me again during this session.
24:11That's also a really nice thing because there are some commands when I'm working with it that I always want to approve. There are other commands that I never want to auto run, so I I like having that that control there.
24:24And you can see here that it did 15 tool users, 61,300 tokens, one minute thirty six seconds, and it completed the claw dot m d file.
24:34And just to show you what that looks like here and how it actually comes together, uh, this is what it does.
24:42It kinda builds its own guide that it uses to help better, uh, kind of understand the structure of my code base.
24:51It even picked up my old folder that I'm trying to phase out of that. It's kinda migrated up into a new version of, uh, Vue.
25:01It's got all the technology that it's used. We've got all the domain models. Does a really good job with local development.
25:07Like, all of this is 100% correct. All of this is correct.
25:12So it's it's really it's really useful. And I found that this kind just having this kinda does speed up the development efforts a bit.
25:22In particular, for me, we care a lot about our design stuff. So having this in here is actually really nice as well because it doesn't need to go look in other code to try to figure out that design.
25:35You can see our integrations that it's using, so on. Right? So now I am in the RU code repos.
25:41This a public code base that you can pull down from GitHub. And I basically asked it to diagram out how the Orchestrator subtask version system works.
25:53So you can see here that it actually generates some mermaid diagram sequences. We need to load those into mermaid renderer to kinda see how that actually works. But let's not focus on that right now.
26:03Let me focus on the flow that this actually went through here. So the first thing that I ended up doing here is I basically asked it to tell me about this code base.
26:13So this is one of those generic queries that I was talking about. It read the read me.
26:19It read the package dot JSON, and it got enough information from that to give me a definition of what the code base is.
26:28It even picked up the the main components on it. This was relatively fast. Let me see if it actually tells me the time there.
26:33I don't think it does when we're this part in. Now, basically, what I I did is I ran a net on it.
26:39So let me show you what the, uh, cloud dot m d file looks like here. And you could see that it actually did 26 tool users, 90.6 k token, and ran for two and a half minutes to generate that. And I knew a good amount about the root code code base, enough to say that the majority of this I can confirm is correct, but I I'm not as familiar with every part of it.
27:02So I don't have the same ability that I do with my own code base to be able to say, yeah, that's 100% all correct. But it did generate a 149 lines of information here based on, you know, 90,000 tokens that it ran through.
27:18And you can kinda see some of the summarization there. Then I basically ask it about diagramming it out. I actually told it to ultra think through this, and you can see very particular.
27:30Like, you can feel the difference that it needs to go through. Like, look. It triggers this thinking portion in here.
27:37So you can start seeing thinking where if you go up through these, because I didn't tell it to ultrathink, you're not actually getting the thinking as part of the output here. So you could see that just carry through through all of this.
27:49Now it does slow down the process, but sometimes you really do want it to take its time and generate, you know, a more thoughtful response.
28:00Four minutes and, uh, nine seconds. So there we go. 32 tool uses, 88,000 tokens, four minutes and nine seconds.
28:07So now that I've had that got sort of this high level understanding of the Orchestrator subtask system, what I would typically do now is this is a this is a feature I've been wanting in RootCode for a while, where I want to be able to represent in Root in RootCode, basically, the orchestrator and all the subtask, how much each of those you use, how much do they cost, how many tokens they use, and be able to visualize that and ideally click through them.
28:35I don't actually have an idea for I haven't set that. I figured out the design on that. But, basically, I could go and kick this off now and say, hey.
28:43Now that you have that, go figure out a way to do that and get some ideas on it. Most likely, it's not gonna be something that, um, we're going to to actually be able to keep, but that's just an idea of kinda how you can work through this.
28:56So here, you can see that it actually created its own to do list, and it's gonna work through it one by one. So if you watch my previous video where what I talked about a lot here so let me move my video over a little bit, is basically the rankings of each agent.
29:12And Cloud four does really good. Cloud four Sonnet, very specifically Sonnet, does really good in Cloud Code.
29:21And it should because it is their tool that they're building. They understand the inner working of it very well. And it kind of goes to say, like, my personal ranking is this, and I don't see this changing drastically over the next month.
29:37It has historically changed a lot. Like, one time, I was using cursor 100% of the time, and then I was using, uh, I used Klein for quite a while, and then I used Augment for quite a while.
29:50And I still use Augment now. I think they're phenomenal for certain things. And that it just you know, you kind of have felt so much sort of turbulence in this space because people are leapfrogging each other.
30:04But I think Clog Code is just this really sweet spot for me. Just it's . So I think Cloudcode is just a sweet spot for me.
30:11I'm gonna be paying for the foreseeable future, the $100 a month until that ends up being, like I if if they start nerfing it, I'm gonna, uh, probably upgrade to the $200 a month plan. But I I really do need to try to get my spin down.
30:26So my current stack that I'm gonna be running is going to be $30 a month for AugmentCode because I'm on the grandfathered plan.
30:37I'm go I currently have $200 a month for OpenAI because of codec.
30:46There's something really special about codec being able to be on a bike ride, get something popped into your head, pull over, get kick something off, bring it back, and work on it in Clog code. Now I think there is probably a way that I could probably downgrade this now that they've given you access on, I think, the pro plan.
31:03So I'm gonna look into that because I don't know what the limits are on that. But if the limits are where I can run five, six tasks a day on Pro Plan, there's no reason for me to be on the $200 a month plan there. And then the other thing would be, like, API cost.
31:18Now my API costs are kind of crazy because eval, typically, they've started to creep up in in pricing. So, you know, I'll spend, you know, $200 at eval run easy, and that's me trying to use models that are built into these systems as well.
31:34So this can actually run 200,000 200 to $1,000 a month just on trying to run, um, evals and testing all the different models and stuff. So this is for clogged code.
31:46I'm going to be having my $30 augment code. I'm gonna be keeping codecs for foreseeable future. And then this is going to be a combination of root code and any other API costs that I have.
31:59Of course, there's just going to be miscellaneous costs with with testing other things. So one thing that I'll just note in here is .
32:10There is always a chance. There's some upcoming rumors about them launching something that may bring me back into using Gemini or, uh, Google models again.
32:22But it's going to it's going to be need to be, like, something pretty massively good because Cloud Code right now is in an incredible spot.
32:31Now I think that's gonna about wrap it up. There's so much I could talk about about ClockCode. Is there anything you want me to go deeper into on ClockCode?
32:39Let me know. I've been getting so much feedback on all the stuff that I've been doing. Trying my best to take that in and just continue doing better and better.
32:46Anyway, I appreciate all of you. Let me know your thoughts below.
32:51If you've had a chance to try Clogcode, if you're paying for the 100 or $200 a month plan, if you hit limits because I've never hit a limit on the $200 a month plan, but I've heard people hitting it on the $100 a month plan. So I'm curious what is gonna happen as I switch to that.
33:05Anyway, till next time, everyone. Have a wonderful, wonderful day. Peace out.
The Hook
The bait, then the rug-pull.
Adam opens with a number — 318 commits in May — and never breaks character from 'I just used the thing, here's what happened.' No 'hey guys,' no promise theater. The whole 33 minutes is a single outline canvas, one webcam, and a creator transparent enough to read his own $200-a-month OpenAI bill out loud.
Named ideas worth stealing.
04:45list
Prompting trigger words
- ultrathink
- architect/plan/implement
- don't write code — ideate
- REPL: tell me about this codebase
- @filename to reference files
- don't be afraid to start over
Specific phrases that change how Claude Code behaves. 'ultrathink' triggers deep reasoning; 'ideate with me' suppresses code; @ tab-completes file references; revert + restart beats grinding when the prompt is wrong.
Steal forany creator who teaches AI tools — phrase-as-API lists are shareable and screenshot-friendly
11:20list
Core Claude Code commands
- /init (generate claude.md codebase map)
- /clear (wipe context, start fresh)
- /model (switch Opus vs Sonnet 4)
- /compact (auto-summarize)
- /cost + /status (API mode)
Minimum command set to be productive. Sonnet 4 is the daily driver; Opus burns limits fast.
Steal fortutorial videos on any CLI tool — 'just these 5 commands' is a watchable beat
16:50concept
Git worktrees for parallel Claude Code instances
- git worktree add <path> <branch>
- git worktree list
- git worktree remove
- cd into path, run claude
Run multiple Claude Code agents on different branches simultaneously without context-switching. Each worktree is an isolated working copy.
Steal formulti-agent orchestration content — pairs naturally with batch/cluster narratives Joe is already running
How they asked for the click.
VERBAL ASK
32:33newsletter
“Is there anything you want me to go deeper into on Claude Code? Let me know your thoughts below.”
Invitation, not demand. No 'smash like, subscribe.' Asks for content suggestions — doubles as engagement signal. Pairs with 'I appreciate all of you' — warm sign-off, not transactional.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
