All The Claude Code Features You Dont Know About

0:00In this video, we're going to be talking about everything that Anthropic has shipped within the last week and stick with me because there is a massive announcement here at the end. So, Anthropic just shipped a ton of stuff and when I say just about everything, I mean everything like voice mode for Claude code, a brand new product called co-work, interactive visuals, a $100 million partner fund, doubling the uses limits over spring break for everybody. Oh, and by the way, Claude code alone just crossed a billion dollars in annualized revenue. So, let's let that sink in for a bit because we need to catch our breath after this Nemo claw announcement. Anthropic is striking back and shipping features. There was a major upgrade announced today with co-work and

0:46I'll get into that a little bit later on how it affects you. So, if you've been sleeping on Claude code and maybe you haven't tried it out and then maybe you've just been on the fence about it, this video is going to change your mind, I guarantee you. Because what they shipped this month is not just an update, it's a statement. They're trying to keep pace with the open claw hype and Anthropic is striking back and the numbers back it up. Time magazine just reported that Claude code hit a billion dollars in revenue and it doubled to 2.5 billion just this last month and they are on track to surpass OpenAI in revenue by the end of this year. But don't tell that to Sam Altman. So, let me break it down starting with Claude code because this is what I use every single day. First up is voice mode, finally. You type a {slash} voice

1:32command and now you can talk to Claude code with your actual voice. I've used whisper flow for the last year, so but if you don't, now you can use voice command and you can talk to your agent. Amazing. Hold down the space bar, speak, let go and it sends the message. Push to talk, which is pretty cool because it's not just sitting there listening to everything in your room, but you get to control when it gets to hear you and they They 20 languages now. Amazing. Russian, Polish, Turkish, Dutch, Español, como te llama, whole bunch more. Whatever. You can even customize your key bindings. If space bar is not your thing, control delete, whatever you want to do. Now, here's what really got me. It understands the technical terms. You can say variable names, repository names, function names, your dog's name,

2:17and it actually gets it in the right context. So, most voice tools butcher anything technical, this does not. And I have been testing it out today, and it's been pretty close to WhisperFlow, but thing is, I've been using WhisperFlow for the last year, so it knows a lot of my cadence and tone and how I talk, but it did a pretty good job. And plus, WhisperFlow, I'm paying $19 a month for. So, there are some moments where you get into deep part of your project, and you just want to say, "Hey, check this file." instead of typing it out. So, yeah, it has its great moments of, you know, saving your time and stuff. But, the next step that really got me more curious was the slash loop command. So, basically, you can now set up reoccurring tasks right inside of Claude Code. So, you just type slash loop, then

3:04a time interval, then what you want it to do, like, slash loop 5 minutes, check the deploy. Or, slash loop 30 seconds, run some tests. And it just does it. So, every interval, as long as your session is open. Why is this a big deal? And why am I even talking about it? Because before this, if you wanted to monitor something, you had to set up a whole separate cron job, or like write a script, or use some external tool. Now, you can just tell Claude Code to keep checking on something while you go work on something else, brew some coffee, water the plants, go surfing, whatever you want to do. I have been using it to run some tests today, and every 30 seconds while I'm building, and it's done a pretty good job. So, I make a change, and I don't even have to think about running tests. It just runs

3:49autonomously. It just does it and tells me if something broke and that does end up saving time. Again, you're going to hear that more and more in the channel. Saving time, I'm all about saving time. I don't have a lot of time to waste. So, for monitoring deployments, come back, guys. You deploy something, you tell Claude code {slash} loop 2 minutes, check the site is responding, and you could go eat lunch while it watches your deployment. That's pretty nice. That is the dream right there. I do run already 17 cron jobs in my Open Claude setup right now, and the {slash} loop command could replace half of them for quick monitoring tasks, which is pretty cool. If you're struggling to keep up with content, well, I'm about to save you about 40 days worth of work. I built [music] something called Content Machine. It's 10 AI agents that run on the Open Claude orchestration, and they

4:36handle everything. Scripts, thumbnails, X posts, blogs, outreach, clips, newsletters, all of it. So, I went from 1,000 subscribers to 4,000 subscribers on YouTube in 7 days using this exact system. Every single morning, I wake up and the content's already done. I spend maybe 15, 20 minutes reviewing and approving them, and I move on with my day. It works for any niche, fitness, [music] finance, real estate, marketing, whatever you are building, and it is 100% completely customizable to your use case. So, you get the mission control dashboard, all of the cron jobs, everything I've built over the last 40 days helping me gain more and more people to subscribe and join the community. So, you plug in your own

5:22thing and it molds it to you. It learns how you talk and it writes [music] so it doesn't sound like AI slop. $97 one time, it's not a subscription. I'll put the link down below and you'll thank me later. Now, the big one, the million token context window, and if you're new to the scene, this is absolutely huge for Claude models because they were previously just on $200,000 context window. And if you're new to AI, context window means everything. It means how much memory it holds with every single query you're talking to these LLMs. So, it has five times the amount, 750,000 spoken words, which is pretty much like 10 Bibles over and over again. So, it does have much more context when you're now speaking to any Claude model. So,

6:08this is open to people who are on all plans. So, to put that into perspective, the old limit was just 200,000 tokens. And a lot of these other LLMs have leveled up since then. So, we went from 200,000 to a million. That's pretty amazing. That's a nice jump. What does that actually mean like realistically? It means you can load an entire code base into the context without it forgetting stuff halfway through. Before, if you had a big project, Claude code would start compacting like 2 seconds in, and it would lose all the details. So, this is basically summarizing older projects of the conversation to make room. That's what the compact event means, and you lose a ton of details. So, this is where things get weird. It would forget a decision you made 20 minutes ago, and I know

6:54you've experienced this. Now, with a million tokens, it won't forget. You can have those really long sessions where you're building something complex, and it just remembers everything. This changed the entire project for me because I'm working on Mission Control, which is a Next.js app with a bunch of different pages and API routes. And being able to keep all that in context is huge. For example, when I was working in my Mission Control, I would go through three to five compaction events every single day. We do have backups for memory and and things in place, so it does remember. But not having to go through a compaction event since this upgrade has been amazing and a much more clarity clarity when it comes to talking back and forth with your open Claude agent. Speaking of Opus 4.6, which

7:39coordinates all of my models of all 12 agents, It is now the default model in Claude. So, Opus 4 and 4.1 are gone. They got removed entirely and they brought back the Ultra Think keyword where you can actually tell it to think harder on a specific prompt. This is super underrated if you do not know and you're using Claude code, type in think, think harder, and ultra think and they will light up. So, it basically does an extended thinking about what you're telling it and what you're asking it to do. Don't forget about plan mode, too. Plan mode is super underrated. I don't feel like a lot of people talk about it enough. Before you ship a feature and execute something, toggle cut uh plan mode so you can at least see the plan that Claude and Opus 4.6 creates so you

8:24know what's going on. So, basically for normal stuff, it uses medium effort right out the box, but when you actually need to really dig deep on a hard problem or debugging, you can say ultra think and it cranks everything up to maximum. It does use a ton of tokens, so be aware of that. They also added a slash effort command where you can set low, medium, or high analysis depth. So, low for quick questions, high for deep architecture questions. Um so, it's pretty nice, pretty nice. So, let me hit you with some of the smaller updates that are actually useful because it's in those small updates that things get lost in the sauce. Slash color lets you color code your terminal sessions. Kind of cool. So, if you're running multiple Claude code sessions at once and you have 50,000 terminals like you're in the

9:09Matrix, you can actually tell them apart visually. I do this all the time because I do have one session for building and for one for debugging and maybe one for working on a specific feature that we're working on. So, you can now start Claude code with dash n and give your session a name right at startup, which helps keep things a little bit more organized. So, instead of just random sessions sitting on all over your monitors, you can actually name them by what you're working on. The slash context command got better, too. It now identifies which tools are consuming the most contacts and flags memory bloat. So, if your MCP servers are eating up the context window, now you know exactly which ones. And they added the auto memory directory where you can set a custom directory for

9:54automatic memory files with timestamps. That's a pretty big one for anyone running complex setups. So, for the big repo people, they added a work tree sparse path. So, if you're working on a massive mono repo, you can actually tell Claude Code to only check out the directories you need. No more pulling down the entire thing when you only need one folder. HTTP hooks are new, too. Before, hooks were just shell scripts. Now, you can post JSON to external URLs. So, you can trigger webhooks, set data on your own APIs, whatever you need. And MCP is wild. MCP servers can now request structured input from you mid-task. Like, it can pop up a form asking you for specific information or open up a URL for you to authorize something.

10:40Interactive forms right in your term. Yay! Performance-wise, they fixed the memory leak issues, reduced base memory by 16 MB, cut the bundle size by 510 KB, reduced prompt re-renders by 74%, and cut remote control polling server load by 300 times. So, it's much faster, lighter, cheaper, all that. That's amazing. Why does this matter? Because when you're running this thing all day, every day, like I am, you're going to save on token spend. If you're going through the API or through OAuth, it's just going to be better on the wallet. So, that is Claude Code in a nutshell, but Anthropic didn't just stop there. Let me talk about co-work, because this is going to blow your mind. Co-work launched with a major update today. And basically, this is a no-code agent that

11:25rivals Open Claw. So, 38 plus connectors right out of the box, Gmail, Google Drive, Microsoft, Slack, Notion, a whole bunch more, Doom, Quake, if that's your thing. They launched a plugin marketplace. You can create custom skills and sub agents, scheduled reoccurring tasks, contacts files, and they just announced a Microsoft Copilot partnership. Yay! Plus, it now has Windows support and direct local file access and sandbox VM. So, here's my honest take on Co-work versus what I already use. Co-work is basically trying to do the exact same thing as Open Claw, but less technical. So, it's kind of good for people that are just getting into the space and want to get their feet wet with a few automations. More plug and play. That's awesome. And for a

12:10lot of people, that's exactly what they need. If you're, you know, not technical and you just want to connect your Gmail to your Notion and automate some workflows, Co-work's probably going to be great for that. But, here's why I still run things on Open Claw. Real quick, before we keep going, if you're watching this and you want to actually build with some of these tools than just watch a videos about them, then you're going to want to check out our community down below, Shipping School. We have a full Claude code course, a full Open Claw course, and four live bootcamps every single week where we actually help you get set up from scratch. Like, actually set this thing up, not just watch a tutorial and figure it out by yourself. And we also provide one-on-one coaching so you can book a call with me, we can share screens, and I can help you get Claude code or Open Claw running on

12:56your machine. That's it, no fluff. I built this community because watching YouTube's only gets you so far. We launched it just 3 days ago and we have over 55 members. You need people around you who are actually building, people who hold you accountable, and coaches who can help you when you get stuck. I'll put the link in the description down below. Get in now before the price goes up. I have 12 agents running on the Mac Mini, and it's the entire Wu-Tang Clan running everything. The RZA writes the scripts, Inspectah Deck handles the X content, Ghostface does the research. We have Solomon coding. He's not even part of the Wu-Tang Clan, but he's just sitting there waiting for coding projects. I have 17 cron jobs firing off every day. They all have custom skills. I have MCP servers doing things that no

13:41plug-and-play platform can do because it's more specialized for my use case. The level of customization and control you get with open claw is just different. You know, code work is going to nail the 80% of use cases for most people and I love that. But for builders who want to go deeper and be more specific with their AI agents and who want to actually run and own their own setup, who want to run everything locally on their own hardware, open claw is still the move, plain and simple. Real quick, two more things from this week and I don't want to bloat this video out too much, but these were key with Anthropic releases. They launched what they're calling image with Claude. So, interactive visuals. Claude can now generate charts, diagrams, and illustrations right inside its responses. No code needed from the user.

14:26You just ask it to visualize something and it creates an interactive visual right there, which is pretty cool for anyone doing like data analysis or presentations or just trying to understand more complex concepts visually. I'm a visual learner. And they dropped a hundred million dollar partner network, the Claude partner network. Pretty cool. Enterprise deployment support, partner training, technical support. They are going all in on getting Claude into big organizations. So, which just tells you where their head is at. They're not just building tools for individual developers, they are building an ecosystem. And when you combine that with the revenue numbers I mentioned earlier, you can see the trajectory and this company's moving pretty quick. So, one more thing, from March 13th through the 28th, all Claude

15:12users get double usage limits outside peak hours. Now's the time to ship, guys. Get your item, get your idea, get your project, get it shipped out. Now there's no excuse. So, that applies to Claude Web, desktop, mobile, Claude Code, co-work, Excel, PowerPoint, everything on Earth. So, if you have been hitting your limits, try using it during off-peak hours, and you get twice as much. That is a little nice bonus, and I know what they're trying to do there. That's just a marketing strategy. So, here's my take. Anthropic is not slowing down, they are shipping updates almost every single day. Voice mode, the loop command, a million tokens, co-work, interactive visuals, a hundred million dollar partner fund, and a billion dollar product all in one month. Amazing. That is not normal. That is a company that knows exactly where they

15:57are going. And for builders like us, this is pretty incredible. I use Claude Code, I love Claude models. I might as well just change my name to Claude at this point. Every update, the tools are becoming more powerful, which means we could build fatter faster, or fatter. And things could just get better for us as they keep upgrading these models. So, we could ship more, we could grow more, and if you're not building with these tools right now, I have no idea why you'd even be watching this type of video. And I don't know what you're waiting for, because I went from not knowing what a bash command was to literally running a full AI operation on my Mac mini that has generated $3,300 in monthly recurring revenue in eight days. And tools like Claude Code are a huge part of why. If you want to learn how to set this up and get it all going, we have builders inside of Shipping School,

16:4379 members in eight days, and right now we're doing exactly that. Four boot camp calls every single week where we hold your hand and we get your item built. Real people building real things, shipping real products, helping each other out. You could check it out. I will put the link down below. Get in there before the price goes up. And if you haven't subscribed already, what are you waiting for? We release four to six to 10 videos every single day. If you want to stay up to date with AI news, make sure you hit subscribe, hit the little bell icon, and we'll see you in like 10 minutes. Have a good day.