Hermes AI DESTROYS OpenAI Codex?

That means just be opening our codecs in speed benchmarks and not by a tiny margin by enough that thousands of people are already publicly switching and whilst that's happening Google DeepMind quietly dropped a brand new tool called anti-gravity CLI and Elon Musk just he's in a few weeks he's going to reveal a detailed design for AI satellites that could change where AI actually runs in the world. Four stories all breaking right now and every single one of them matters to you if you're building a business with AI in 2026. Let's start with the biggest one which is Hermes beating codecs. So OpenAI codecs if you've been using it is an AI coding agent and it writes code it runs tasks it handles workflows it launched with a lot of hype right developers loved it.

Business owners actually started using it to build automation so it was one of the go-to tools for getting things done without needing a developer in your team and then something started happening really over the last 24 hours performance was dropping responses getting slower tasks taking longer and the conversation on x for example started shifting from codecs is great to I'm switching and there was actually over 1700 posts in 24 hours from people reporting the same thing it's not a few complaints that's a pattern and when tools start degrading like that when the speed drops when the reliability wobbles people don't stick around they move and where are they moving clawed code the numbers are clear here the migration is already happening right now today as you're watching this but here's the thing that actually matters more than the codex complaints which is a Hermes agent from this research just beat codecs and key speed benchmarks and this is important because Hermes is self-improving self-evolving AI agent it doesn't just run tasks it can learn and improve as it works it gets better the more you use it codecs doesn't really do that claw code doesn't really do that either at least not in the same way Hermes is doing something structurally different and right now it's doing it faster so what does this mean if you're a business owner watching this well it means the agent you're using right now might not be the best one in six months it might not even be the best one next month right the gap between tools is shifting so quickly and the thing you built your workflow around last quarter might be slower more expensive or just worse than something that just launched right this isn't a criticism by the way this is just what 2026 looks like tools move fast and if you're not watching out for this sort of stuff you fall behind and here's the thing i want you to sit with whilst we go through the rest of this by the end of this video i'm going to show you exactly how to think about switching tools about breaking everything you've already built because the last thing you want to rebuild is your entire workflow every time something new drops there's a smarter way to do this and i'll come back to that later now one thing i want to be honest about here the Hermes benchmarks are early but the signal is real when benchmarks start coming out and Hermes is consistently faster on tasks that matter to real workflows not just developer tests but actual runs that matters especially if you're running AI agents for for example client outreach or lead generation or lead follow-up or data pulling right speed directly affects cost faster agents run more tasks on the same budget slower agents eat up your tokens and give you less output so i'm not saying like drop everything and switch to Hermes today i'm saying watch this one closely because the teams that catch these shifts early are the ones that get a head start on better results before everyone else figures it out and if you want to stay ahead of exactly that kind of shift and you want coaching on how to evaluate new AI agents for your business without wasting weeks figuring out yourself join us in the AI profit boardroom right now we're running weekly coaching calls specifically around AI agents like Hermes like Cloud Code and we've built an agent operating system that plugs them all together and gives you like an image studio an SEO studio video studio and everything else all in one amazing dashboard so you can basically automate whatever you want plus we've got a 30 day roadmap for implementing it and daily tutorials walking you through the actual setups link in the comments description or go to the AIprofitboardroom.com to get access now let's talk about Google's DeepMinds anti-gravity CLI because this one flew under the radar today but it's worth paying attention to anti-gravity is a new tool for terminal based coding agents in plain English it's a command line tool that lets AI agents run more complex coding tasks directly from a terminal the name is new the repo just dropped and it already has thousands of people using it right now before there was always anti-gravity but now this is anti-gravity CLI right this is a powerful way to basically code with agents plus it's free for anyone using it right there are token limits of course but you can test it out you can try it out for free and see how you think right now this is actually replacing Gemini CLI which is another way to use Google AI inside your terminal but basically anti-gravity CLI is just designed to make agents and coding agents particularly more capable in terminal environments you want to think of it like this right a terminal is basically where code actually runs it's the engine room most AI tools give you a nice interface a chat window a button to click anti-gravity CLI goes a bit deeper than that so it's trying to let agents work at the level where the real lifting happens now why does this matter to you because the tools you use the automations the pipelines the AI workflows they all run an infrastructure like this underneath so the better that infrastructure gets the more your tools can actually do and when Google DeepMind is building it you know it's going to end up inside products you're already using for example like Gemini, Notebook Lemp, Google Workspace etc and this is becoming the frontier coding agent for Google right the pace that Google is already moving faster now than it ever has I mean in the last few months they've added AI to pretty much everything right and they've added an agentic layer to everything they've actually announced that Gemini Spark is coming as well so all of this stuff is improving really quickly they already released all the announcements from Google IO last week including for example new updates for Stitch new updates for Gemini Spark and the AI agent on desktop and all this stuff is becoming more and more agentic in the things you do day to day and bear in mind like this is rolling out to billions of people right billions of people use Google's products so the whole world is getting AI pretty much at the same time now let's talk about Elon Musk this one is a few weeks out but it's a story worth setting up so Musk replied to an analyst on X yesterday that he will present a more detailed AI satellite design in a few weeks right and this is about something called TerraFab so TerraFab is a project involving Tesla SpaceX and X AI together the goal is large scale chip production and the level that goes far beyond what's on the ground today right and the design Musk is referencing involves putting AI compute into orbit right space-based data centers AI that runs not on a server farm in Virginia or Texas but on satellites in constant sunlight and here's why this matters on the ground running AI is expensive partly because power right data sensors use enormous amounts of electricity but in orbit you have constant solar energy so the sun doesn't set in certain orbital positions and in space there's no air meaning cooling is different right so engineers have actually figured out how to use the vacuum of space to pull heat away from the chips the technical problems are different right it's not necessarily worse and if you solve them you get AI compute that runs on essentially unlimited solar power around the clock pretty crazy now does this mean you're gonna log into a satellite powered Claude next year probably not not in a direct obvious way anyway but here's what this kind of infrastructure move does to the whole industry when the cost of running AI drops dramatically because the power is free and the cooling is free then every single API gets cheaper right every tool you use gets faster every limit that exists today because it's too expensive to run more compute gets pushed back we've already seen what happens when AI inference gets cheaper models get released that were previously too expensive to deploy publicly tools that cost 100 tokens become tools that cost 10 workflows that weren't viable at scale become viable and the TerraFab satellite compute vision if it works is a multiplier on all of that right and Musk is a few weeks away from showing exactly how he thinks it works I want to be clear here this is a design reveal not a launch we're talking about a presentation of a plan not a product shipping tomorrow but the direction matters here right when you see Tesla SpaceX XI coordinating on a single infrastructure project aimed at putting AI compute in space it's not a side project right it's a long-term bet on where AI runs in the world and here's what I want you to think about now right I said at the start I'd show you how to think about switching tools without breaking everything you've already built here it is the mistake most people make is treating every tool like it's permanent they get locked into one agent one workflow one platform and then when something else comes out they either ignore it because switching is too painful or they jump ship immediately and have to rebuild everything from scratch both are bad ideas the smarter move is this right build your workflow in layers and this is something I've been experimenting lately which is basically creating like an agent operating system where I can plug everything in right so for example anti-gravity gets better boom let's put it inside the dashboard right Hermes gets better great let's plug that in right codex released goal update last week where you can create autonomous long horizon tasks where agents can work for hours great let's plug that into the dashboard right and so you see how this stuff gets better and better the more you use it this is what we're doing right now um if you want to be a part of that you can get the zip file the 30-day template and everything that we use to implement it inside the AI profit boardroom link in the comments description or go to the AI profit board.com you can also connect with me directly inside there you get a community of 3100 members who are also using AI automation like me I create video tutorials for people daily who ask questions and also we have weekly coaching calls plus every time we update the agent operating system we actually give you the new update for free as part of the membership now let me zoom out for a second today okay what happened in AI is a specific kind of day it's not the biggest news day ever but it's a representative day and what it shows you is the rhythm of how this industry moves in 2026 you've got one tool codex showing its first real cracks you've got performance degradation users publicly migrating and that's a tool that was like the big thing even just last month there were huge updates for it then you've got another tool Hermes stepping up and beating on speed right that's the cycle something gets built something better gets built after users moves and that cycle used to take years now it seems to take days and weeks it moves very very quickly and we've seen that with Hermes also overtaking open core as well and then you've got musk setting up a presentation for a vision of AI infrastructure that sounds like science fiction but it's being built by the same people already put reusable rockets into commercial use SpaceX wasn't science fiction it was just very early so this is a pattern lots of big moves right here very exciting stuff thanks for watching I'll see you on the next one cheers bye bye