Qwen 3.8 Agent OS Just Changed AI Agents Forever! artwork

Qwen 3.8 Agent OS Just Changed AI Agents Forever!

AI News Today | Julian Goldie Podcast

August 5, 2026

Qwen 3.8 Agent OS Just Changed AI Agents Forever!
Speakers: Julian Goldie

Topics: Marketing, Business

**Julian Goldie** (0:00)
And 3.8 Max is one of the most powerful models in the world, and you can now plug it into an agentic operating system to build and automate anything that you want. Now, the reason that you do this is, number one, it's 2.4 trillion parameters. Number two, it's actually beating Fable 5 on many of the benchmarks. And number three, it's really powerful, one I've tested out. So today, what I'm going to walk you through is the open-source Qwen machine for agentic operating systems, so you can build and automate anything you want using Qwen, and you'll discover how to plug it in into your existing workflows and systems, so you can get the most out of it and have it working and orchestrated directly alongside your AI agents. Now, you might say, why would you use this? Again, it's a 2.4 trillion parameter giant that you can plug into your dashboard. This is an open-source agent. Bear in mind, it's self-improving as a model, so it's actually getting better and better all the time. And so you can plug this into your chat, your memory, your tickets, your builds, and you've got your agentic operating system.
Then you can have all of your workflows in between, and you can plug in Qwen 3.8 Max into this. So let me show you an example here. So we have, for example, Hermes agent Qwen 3.8 plugged into our Hermes agentic operating system. And so what that means essentially is like, we can switch between models when we want to do different things. We can also have Hermes directly orchestrate all of our other agents or other agent profiles. We can spawn sub-agents of teams. And then also we can, for example, get our agents to automate anything. So for example, here we said forward slash learn, and then we've given it a guide. And it's actually created a custom guide and skill locally using this system. Now, one of the other benefits of Qwen 3.8 is actually pretty fast, and it's designed for agentic models. And also you can't use, for example, Claude CLI with this system. But if you have a coding plan with Qwen, you actually can plug that into Hermes agent. Now, also what's really good about Qwen specifically with agentic operating systems is that it's designed to code for hours autonomously alone. So you can plug it into a goal system, as you can see right here, and then you can walk away, leave your AI agents building for you, and you can come back to your history of the goals here. Now, one of the reasons you would do this is because if you have a big task that needs to operate autonomously, and you don't want to babysit it, you can just have it looping over in the background, and that way you don't need to prompt your AI anymore. The other thing that I like about Qwen is that it's very good at call tooling. So, for example, if we use something like Obsidian inside Qwen 3.8 Max's Agenetic Operating System, then it's very good at looking at our context, understanding our memory systems, and then updating that with new things it actually does. So, for example, here you can see Agenetic OS Memories, and it's pulling in the latest information, updating our Agenetic Memory System, and from there, that gets plugged into Obsidian. You can also use this for lead outreach. So, what we have over inside the outreach section here, is a system where we've plugged in the Hunter API key, we can use Qwen 3.8 Max as a brain, and we can find new leads around anything that we're searching for. So, if I type in, for example, like SEO agencies that are interested in white labeling backlinks, you'll find me a list of relevant leads, enrich the leads automatically, and then we can actually draft an e-mail campaign, send out the e-mail campaign and schedule it, and also manage your inboxes using Qwen 3.8 with Hermes directly there too. So, the great thing about that is, the even lead generation can be handled inside a custom workflow, where it feels like a SaaS tool, but actually what it is, is Qwen 3.8 Max with Hermes agent, just with a nice UI.
Now additionally, you could use this inside other agentic harnesses. So, let's say for example, you're not a big fan of Hermes agent. Instead, what you could do is use Claude code, and then you could switch the brain over to Qwen 3.8 Max, and start coding with Qwen 3.8 Max as the brain, Claude code as the agentic harness, and then you can see and view all of the builds you've created over here too. So we've even created like full games using this system, full websites, et cetera, and it looks absolutely amazing, and it's very nice to create different things like this. Then what I also like to do here is I will plug in open source projects into this system so that I can basically have custom workflows ready to go if I find something useful. So let me give you an example. We found Open Design, which was a GitHub project, open source, and so we plug that into our agentic operating system so that we can create beautiful designs whenever we want to. You can see some examples of what we've created over here. So we can create these amazing looking websites. We can custom design them. We avoid the AI slop of normal design. If you have a look at these websites, they look super nice, but they're easy to build because we've got it inside a custom workflow using Open Design. We could even plug in free models. So what you could have, for example, is you could have Paperclip with Hermes and Qwen 3.8 as the CEO inside your team, which we can have over here. And then we can have local models actually doing the work as sub-agents underneath. And so the great thing about that is the Qwen isn't used as a brain, as the orchestrator, as a frontier model at the top of the chain, but it can use free APIs or local models to orchestrate the actual work, to do the actual hard stuff underneath. And then also what we can have over here inside our agentic operating system is we have a group chat. So what this means is we can have Hermes with Qwen 3.8 plugged in, but we have all of our other agents inside a group chat too. So we can tag, for example, or we can ask Hermes to build something out. So it's a really powerful system. And bear in mind, like, the output and the quality of Qwen 3.8 Max is just very nice in general. So you can see an example of a build that we've created here, like the UI is nice, the design is nice. It creates fun stuff and it's easy to use. Here's another example of like a flight simulator style game. We even created, for example, Minecraft as you can see right here.

7 more minutes of transcript below

Thousands of transcripts fetched by people building searchable podcast archives

Feed this to your agent

Try it now — copy, paste, done:

curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/YOUR_EPISODE_ID