pre-fix snapshot: oracle pipeline (2026-07-16)

2026-07-16 04:27:29 +00:00
parent 5224645703
commit af11b0952d
43 changed files with 7141 additions and 1558 deletions
@@ -0,0 +1,55 @@
+# NEW Hermes Agent Cloud Update is INSANE! — AI News Today | Julian Goldie Podcast
+# Transcript (cleaned) — 8:14 episode
+# Source: https://open.spotify.com/episode/5V241PvCyCee5hctvuNLFn
+# Resolved RSS: https://feeds.transistor.fm/ai-news-today-julian-goldie-podcast
+# MP3: https://media.transistor.fm/e9f58e1f/568e5431.mp3
+# ASR: faster-whisper tiny/int8/beam1 (CPU). Proper nouns restored from raw.
+
+[0:00] Hermes Cloud has just dropped, and this is a really easy way to basically set up a Hermes agent in the cloud. Have it running 24-7. It takes a couple clicks, and then 60 seconds later your agent is live. So you can build an AI employee in 60 seconds. In fact, we call this the "60-second AI employee."
+
+[0:19] So these researchers just put Hermes in the cloud. This is an always-on agent that never sleeps. Members have finished learns, and it researchers, and it can reach you on every single platform. I actually set it up as a profile inside the Agent OS. I'm going to show you exactly how to get the most out of this.
+
+[0:36] Now, if we have a look — for example, we've got Hermes agent built inside an Agent OS. And we can chat with it over here. It's very quick and easy to use. And the thing with Hermes right now is, there's a bit of a "closed-lid problem." So imagine hiring the perfect employee — they work brilliantly right up until you close your laptop. Lid down, it stops mid-task, mid-sentence, everything halts. That's basically every agent.
+
+[1:00] And the problem is that it can get stuck, kicked, and break. If it's not in the cloud, then you can't run AI APIs. And also, if you're not behind your computer and you need to fix something, that can be a bit of a problem.
+
+[1:14] Now what you can actually use is this cloud setup. And you might say cloud sounds expensive and technical. Actually, you just pick a model — it's online in seconds. And you can also set it up with free models as well. So there's a doorway to an agent that never clicks, clocks off. And it can basically speak to you on every single channel that you actually use.
+
+[1:38] So the doorway is the new portal. It's one login that serves over 200 models plus a tool gateway — web search, image generation, etc. Then you've got the agent itself, which is a cloud agent. So you can give it a model, pick a model, and it gets spun up in a hardened container, which is also good. Because then it's sandboxed and protected in the cloud. It's got its own persistent workspace, and it runs 24-7 on the new portal servers instead of your Mac. So when it's idle, the meter only ticks when it's actually working. So if you stop it, then it's not going to use anything.
+
+[2:08] And it reaches across every channel. So you have one cloud agent, one memory, reachable from Telegram, from Discord, from your Agent OS and us. And that's how the whole system works.
+
+[2:20] Now you might wonder, what happens when you actually deploy one? So it's pretty simple and easy. Like, you just name it, you pick a model. Then it gets put inside a sandbox, the dashboard goes live. It works whilst you sleep. And for example, we can give it a task over here, and then it can work in the cloud completely in a separate container. So it's always on.
+
+[2:41] So if you're wondering how to deploy it, you can just go to the cloud over here, and then you just sign in and set it off. And at least it takes like two minutes to set up. It's pretty simple and easy. So there is a cloud portal here, and then you can set it off from there. It's like two clicks. I've already set my agent OS up — I'll currently show you. But yeah.
+
+[3:03] So the cool thing about this is like, every time I was running Hermes, you know, sometimes it would stop. If it was running on local models, you could close your computer and it'd stop running. Also, for example, sometimes something would break inside the Hermes setup, or the API would run out. And it's like, okay, well, how do I fix it? I have to go back to my computer. I have to set it up again. Whereas with this setup, we've got our agent running in the cloud, and it's ready to go. It's pretty fast as well. That's a cool thing. And you can get this set up in like 60 seconds, right? So 60 seconds, you set up your AI employee — then it never asks for a day off.
+
+[3:39] Now, if you're wondering how to build the setup here, we can use a Hermes cloud profile. And then we can plug it into your Agent OS. So this is what it looks like when you're setting it up — you can do it yourself inside the cloud portal. And then you can create an isolated profile as well.
+
+[3:57] Now, you might say at this point, okay, I already have Hermes on my Mac. Why bother with the cloud? Because your Mac can break. Sometimes you want a Hermes agent that's sandboxed in a separate container, isolated, so that's safer. And then also, if you close your Agent OS, you can still run in the background as well. So it's pretty powerful. You might also say it doesn't use a lot of tokens — but you can set this up with a free API, or you can use a free model or a cheap model. And that way, it's better.
+
+[4:26] Now bear in mind, like you can have Hermes running in the background. So even if you don't have a cloud setup, it can run in the background. But the thing is, then you have to go back into your actual device to fix it and edit it. And if it stops working, then if you're out and about, you can't go and fix that. Whereas with the cloud model, you can just log into the portal online. It's in the cloud and you can edit it. And also, some people don't want to install Hermes agent on their device, because they've got files they don't want the agent to touch. And so if that's the case, then instead of what you can do, it's just have it in the cloud.
+
+[4:58] So let's test this out now. So this is Hermes Cloud. We're going to say, okay, come up with some cute [good] ideas for your automation. It's going to start building here. The good thing as well, when you're using Hermes — if you use it inside an Agent OS like this, then it's plugged into your memory. So it understands you, your business, your memory, and everything else, everything inside one place, so that it can use after-context and come up with better ideas.
+
+[5:23] Also, one thing I'll say is it's very, very fast to respond. Way faster than a lot of APIs. For example, when I'm using like Roo Code 4.5 — Roo Code 4.5 took like a minute to do something like this. Whereas with the cloud setup, it seems very, really fast and responds quickly, which is awesome as well.
+
+[5:46] So that's basically it. That's the whole set up. That's how to use Hermes Cloud, how to get it set up. Again, if you're wondering how to get it — you just go to [hermes cloud URL]/cloud. Why would you use it? Because it's always on. You can run with over 200 models. It's easy to set up in like 60 seconds. And also, if you're not technical, then you might not want to set up Hermes itself. And so if you just use the cloud instead, it's easier. And also it can work overnight, even whilst your machine is off and everything else. Plus, there's no service that you need to mess around with. And it's a really easy way to just get the most out of [it].
+
+[6:25] So that's basically the whole system. What did you learn today? You've got an agent that never sleeps, you only pay when it works, you can reach it everywhere, you can try it for free first as well. You can learn how to put it inside your Agent OS. It has a memory that compounds and improves. And you now have a 60-second AI employee that you can use with Hermes.
+
+[6:45] So the cool thing about this — well, we've actually built it into our agent operating system. And we've got a whole section for Hermes. So for example, we have Goal mode over here, we have chat-to-talk, we have a Hermes voice agent that we can control with our voice and get it to build things. And we can see over the history everything that we've built over here. They've got Hermes Oracle, it pulls in the latest news on a daily basis and gives us the most trending news. We have Hermes [Astro] that looks at our competitors and the keyword research. We have an outreach tool with Hermes agent that can do outreach and find leads and generate leads and send emails for us. Basically every powerful feature that's inside Hermes, plus some custom workflows we've built into this system.
+
+[7:34] So if you want to get that — since I've got the AI Profit Boardroom — plus we have loads of training and courses on Hermes agent, you can get that in there. I've got a profit link in the comments description, or go to the app on [profit.com]. And inside that community, you can ask questions, get help — so you can post questions inside the community here. And I personally answer them with daily video tutorials. Inside the calendar, you can jump on weekly coaching calls, get help and support. Inside the classroom, you can actually get access to all of our best trainings. If you're a complete beginner, you can go from beginner to expert over here.
+
+[8:01] If you want to get our new daily updates, you can get over here. If you want to get our Agent OS with Hermes and Hermes Cloud built in, you can get that here with a video tutorial, the last update date, a full guide, and a zip file to install. And you can also meet people in your local area who are using AI agents like Hermes and others. So feel free to get that, link in the comments description, or go to the app on [profit.com].
+
+---
+NOTE: Light ASR (tiny) left a few ambiguous terms. Best-guess restorations (flagged):
+- "Meas Cloud" → Hermes Cloud; "HGNOS" → Hermes; "anopsite" → agent OS; "SNF" → OS; "zid file" → zip file.
+- "Roo Code 4.5" — the slow coding agent he contrasts with cloud speed (uncertain; could be another tool).
+- "[hermes cloud URL]/cloud" — the actual domain was garbled in audio ("search.com"); likely hermesagent.com/cloud. Unverified.
+- "Hermes [Astro]" / "Hermes Apollo voice agent" — product names unclear in audio; left as best-effort.
+- "profit.com" / "AI Profit Boardroom" — his community; the exact app domain was garbled ("forborn.com").
+Raw file (verbatim ASR): ep_hermes_cloud.txt (same dir).