Claude: NEW AI Operating System is INSANE! artwork

Claude: NEW AI Operating System is INSANE!

AI News Today | Julian Goldie Podcast

June 17, 2026

I Built an Agent Operating System (No Code) That Unifies All My AI Agents in One DashboardThe presenter tours a no-code “agent operating system” they built with Claude to run and manage multiple AI agents in one dashboard, including voice-activated Hermes/Jarvis, Claude with UltraCode sub-agent...
Speakers: Julian Goldie
**Julian Goldie** (0:00)
I built inside my agent operating system, and this is something we built out of Claude. By the way, I don't code, but I literally built this out from scratch. And it's become the most powerful, impressive system that I've ever created or even seen inside AI. And just speaking honestly, like for example, we've got a voice activated version of Hermes over here. We have Claude with UltraCode sub agent teams built over here. We have Fusion, which can achieve Fable 5 level intelligence according to the benchmarks inside here. And every new update that comes out, we build it into this system. And so I want to show you this for inspiration to show you like what's possible and the way that you should be looking at AI and also what I've been building out recently and how we've done this. So for example, you see we've built out like a full RPG open world game using this system. And also anything that comes out, we can just plug into this system as well. So here's another example of like a Skyrim style open world game that we've created using this. And of course, you can use it for like business related stuff as well. So what we've done recently is created this SEO pipeline where it can basically deploy content. Now you might be thinking, okay, at this point, this looks great. What is it? So an agent operating system is basically how you can take all the power of AI and plug it into a system where you can operate all your agents in one place. Because for example, if I'm just going to Claude, super messy, right? If I'm going into, for example, well, how do my agents share memory? They don't. And also, how do I build new cool stuff that helps me get the most out of everything? I don't, right? And also, like, how do you get Obsidian plugged into Chach-EBT? Well, it's very difficult, complicated, and too hard for 99% of people. So how do you make sure that if you're not a coder, you get the most out of everything that AI comes with? And that's why we've created this system. So for example, one thing that I've always wanted to do is have my agents talking together. And that's literally what we've built here. We've got a group chat where AI agents could come together and just chat. We've also got, for example, like Paperclip where we have a team, like a full organization of AI agents, and they just build stuff. And like people say, well, you know, Paperclip looks great, but it's kind of like, it just looks good, but nobody's building anything with it. You can see that we've built out seven different projects using this system. And this is the difference is like, we're actually making something that's actually useful, and we use it every day, and it gets better every day that we use it. So another example of this is like, Fable 5 got banned last week, right? Totally wiped off the market. And there were two things that I learned about that situation. It's number one, you always want to be anti-fragile to the point where if something like this happens, you know, if an amazing model gets taken off the market, no problem, we plug new, better models in, we plug in new systems. So for example, what Fusion does, and this is a new system that's just come out from OpenRouter, is basically it's designed to beat Fable at half the price, because what it actually does is it has five different models, and then one judge that fuses all those answers, and then you get one answer. And that on benchmarks outperforms Fable 65%. So it scored 69% on a Draco score and beat Fable 65%, which is pretty insane when you think about it. And the same, for example, like GroqBuild, this was something interesting that just came out recently.
And I'll show you some examples of what we built with it. So this came out recently, and honestly, like up to this point, GroqBuild has not been very good, but they released a new update recently. It's free with my Twitter subscription, so I might as well use it and test it out. And I can just plug it into my system to create whatever I want with it. So for example, we've created this like beautiful landing page using this system. It looks super nice, super clean. And also the other thing I would say here is like, if you have your terminal open or you have ChatGPT open and you're inside Claude, well, how do you find what you created recently? You can't. So what's better? It's much better to have a system like this where you can plug everything in and just come back to it later. And so that's what we like here. It's like we can just build stuff and we never lose it again. Here's another example. So N2 came out last week as a free API and we already plugged it into free Claude Code, which means we get the power of Claude's agentic harness, but we get it plugged into N2. Now N2 is a free API that you can just build stuff with. So what we can actually do here is voice control free Claude Code and then we can build whatever we want with it, which is pretty insane. And we get a live preview as we actually do it, which is amazing when you think about it. And then anything that we do build, we can come back to later and we've got everything plugged in right here. So here's some examples of what we actually build with N2 just for fun. But it's amazing, right? It's amazing that you can do that. It's amazing that you can come back to this and test it out and see what you've built and keep everything organized in one place.

8 more minutes of transcript below

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.

Using your own key:

curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/1000773186661