The number one question I get more than

anything is which AI tools should I be

using right now? I talk about chat GBT

voice the number one comment I get

should I be using Hermes anymore? I talk

about cloud code number one comment

should I be using code X anymore? This

video is going to answer all of those

questions for you. If you've ever felt

overwhelmed by all the AI tools out

there and how quick they're updating how

quick they're releasing this video will

get rid of that anxiety for you. I'm

going to show you every single model

every single tool you should be using

right now in August 2026. These are the

best of the best. There will be no more

questions on should you be using a

Hermes over code X anymore or anything

like that. You will know at the end of

this video exactly what you need to

download and also how to use it. If you

stick with me to the end of the video

you'll be using every best AI tool on

planet Earth. You'll know exactly the

lay of the land in the AI space right

now. So let's lock in and get into it

starting off with the best planning

model on planet Earth. This is the model

you go to when you're doing any sort of

plans, overview, brainstorming, idea

bouncing and that model is Fable 5

medium. Fable 5 medium is the smartest

model on planet Earth right now. Anytime

I need to do planning, anytime I need to

come up with a way to execute on an idea

I go to Fable 5 medium. For instance, I

was having an issue earlier today. I'm

setting up my buzz community. I'm

connecting all my AI lab to it my three

Mac studios, my DJ spark everything and

chat GBT 5.6 really struggled to set

this up. I went over to Fable 5 medium.

I said, "Hey, check out what they're

doing over there. How do we do this

better?" It came up with a plan. It came

with the best practices. Handed it back

to chat GBT. Worked in first go. This is

the absolute smartest model on planet

Earth. Because it is so freaking

expensive, I don't use it for execution.

I don't use it to write code. I'll go

over that in a second. But for the

planning, there is no model that comes

close. I'm using this. I also like it on

medium. I find that any higher level

thinking, one is just way too expensive,

burns too many tokens, but two

overthinks, and three, just like speaks

like an alien robot. You just don't like

talking to it or reading what it's

saying. So, medium, I think, is the

sweet spot for getting ton of usage and

getting really readable output. So,

anytime you need to brainstorm, bounce

ideas, come up with a plan, you go to

Fable 5 medium. But, I don't use it for

execution. The model I use to write

code, move things around on my computer,

do actual work for me is Chat GPT 5.6

Soul medium. This is the absolute best

model on planet Earth for execution, for

getting work done, day-to-day tasks. You

go into the Chat GPT app, you switch it

over to Codex if you're doing coding,

Chat GPT if you're doing work. I think

they eventually put this all in one. I

don't think there's any need for Chat

GPT text, work, and Codex. I think

they're all going to be combined soon.

Anyway, besides the point, you come in

here, you turn on full access, you put

it on 5.6 Soul medium, you put it on

fast mode if you got the $200 a month

plan, and you are good to go. This is

the best way to write code. I think from

a code writing perspective, it is right

up there with Fable 5, maybe a

teensy-weensy bit under it, but it is

significantly more cost-effective. You

get significantly more usage. Tibo, the

guy from Chat GPT on Twitter, he does a

usage reset every 5 minutes, so you

basically have unlimited usage. And the

hardness is a whole 'nother story. We're

going to go into the Chat GPT app

harness in just a second, but this model

is the absolute best for execution. If

you need to build documents, if you need

to write code, if you need to do things

on your computer, this is the best model

to do it. I have it powering everything

execution-wise, Hermes, Chat GPT,

everything. 5.6 Soul medium is the way

to go. And much like Fable 5, I I medium

the best. You get way more usage. I

think it's just as smart and it doesn't

overthink things. When you use extra

high and ultra mode, I'll give it like a

simple task and it'll take 10 hours to

do it. It's ridiculous. Medium's the

sweet spot. Medium's the way to go.

Which brings me to my favorite AI tool.

This is just the tool I like to have up

the absolute most. And if you've been

tracking my channel over the last couple

weeks, you've seen me talk about this a

lot. They don't sponsor me. Open AI's

never paid me a cent in their lives. I'm

just truly passionate about a lot of

these tools they're putting out right

now. And that is Chat GPT voice. If you

want a full tutorial on this absolutely

incredible tool, make sure you

subscribe, turn notifications on.

Putting out a lot of videos cuz I do

believe this is like the future

interface of AI. But Chat GPT voice is

incredible. I spend literally hours a

day walking around outside, AirPod in,

talking to Chat GPT voice, getting tons

of work done. You can see how I do it

here. I tell it what to do. I say, "Spin

up a new thread to get this work done."

It spins up new tasks, new threads. It

delegates work to other agents. It's

absolutely amazing. You can get more

work done with voice while not being at

your desk at all. It really is an

incredible workflow. All you need to do

is either in the mobile app or the

desktop app, hit that button right

there. That puts you in voice mode. And

then you just start brain dumping out

what you want to get done. And when

you're ready to do work, you just say,

"Spin up a new thread to do that work."

What that allows your voice agent to do

is create new tasks, create new threads

that delegate that work to those

threads, so you can keep talking to it

and get more work done. It's incredible

for multitasking. And you can do it all

in the mobile app out there in the

beautiful world, in the sun, while you

enjoy your life. Chat GPT voice, check

it out. Which brings me to the AI agent

harness. This is probably the most asked

question I get every time I do something

new. "Oh, you're not using Open Climb

anymore. You're not using Hermes

anymore. You're not using Chat GPT? What

are you Let's put it all to rest. Which

AI agent should you be using in August

2026? Is it Open Claw? Is it Hermes? Is

it Claude Code? Is it chat GPT? Is it

Buzz? Which should you be using? Here's

the answer. And by the way, before I

reveal this, I do bootcamps on each and

every one of these tools in the Vibe

Coding Academy, link down below for

that. It's my private community, I do

live calls, 4 hours of live calls a

week. I've done a live bootcamp on each

and every one of these tools, recordings

in there as well. So, if you join right

now, you get instant access to all those

recordings. Check that out now, link

down below, best decision you'll ever

make in your life. The best AI agent

harness on planet Earth right now is the

chat GPT desktop app. Yes, you heard

that correct. I am using the chat GPT

desktop app right now more than Hermes

Agent, more than Open Claw. That's how

good I believe this harness has become.

I don't know if people realize this, but

the chat GPT desktop app has absolutely

incredible browser use and computer use.

Meaning, when you're talking to chat

GPT, it can control and do anything on

your computer a human being can do, and

I think it does it at a level now that

is better than Claude Code, that is

better than Hermes Agent, that is better

than Open Claw. Let me give you an

example. I text my chat GPT agent a

couple weeks ago, I said, "Hey, I was

scrolling through Twitter, I saw that

Deep Seek V4 Flash was just released." I

took the tweet, I gave it to uh chat

GPT, I said, "Hey, put this on my DGX

Spark and run it." It literally

controlled this Mac Studio I'm on right

now, went on to my network, found my DGX

Spark, searched through all of the

models on Unsloth, found the one that's

the perfect fit for my DGX Spark,

downloaded it, loaded it up, built a

harness around it, tested it for me, and

got it running. Literally no approvals

or anything necessary for me. I said,

"Hey, download and install this." Next

message it sends to "Done. I installed

it. It's good to go. They even loaded it

up and actually ran benchmarks on it for

me, which is absolutely incredible. This

is the best agent harness on planet

Earth. And people go, "Oh, but on Hermes

I can DM and I can text on my phone. I

can do it in Telegram." You don't need

Telegram to be using this. You can just

download the Chat GPT mobile app. They

have really incredible remote

functionality. So, if you go to remote

in the Chat GPT mobile app, you can

control any of the threads from your

computer. So, that means when you're on

the go, you load up the Chat GPT mobile

app, you go in, you start a new remote

chat, and it can control your entire

computer for you. And if all your

devices are on Tailscale, you can

control all your devices from your

computer. There's basically nothing that

the other agents can do that Chat GPT

can't. Chat GPT can do all of the agent

Open Claw and Hermes use cases, and do

it I think better at this point. So, if

you want to know what the number one AI

agent being is right now, I'd say it's

the Chat GPT desktop app. You connect it

to the mobile app. You're good to go. It

does everything you can do with the

other agents, but I think much better,

faster, stronger. Which brings us to

building UI, UX, video games, anything

graphically related. For me, I'm going

with Opus 5 inside a Claude code. I know

Opus 5 has been getting a ton of hate

the last couple of weeks, and I think

for good reason. I think there are

reasons to hate Opus 5. As I talked

about in my review video, it is a big

blabbermouth. It is hard to understand

at times. It does talk weird. Sometimes

it even talks down to you, which is very

strange. But at the end of the day,

whenever I'm doing something UI or UX

related, Opus 5 is the best. It always

consistently comes out with the best

results for UI and UX. Now, does that

mean you got to be using Chat GPT 5 6 to

write some code and Opus 5 to write all

of it and go back and forth between two?

No, that can get very confusing. If

you're working on something where really

tight, amazing UI isn't like the most

important priority in the world, you can

stick with chat GPT 5 6 soul. But if

you're working on something like a

production app you're going to sell to

thousands of people. If you're building

a video game. If you're building

something where having amazing UI is

stone cold critical, I don't think

there's any model out there that's

better than Opus 5 at the moment for

that. Yes, you'll have to kind of get

through it speaking in a kind of weird

way. That's fine. You're going to get

the best results from it though. All the

people on X that are saying Opus 5 is

the worst model I ever used. Oh my god.

Listen, there's a lot of people that are

engaged in farming. They get a lot of

You get a lot of engagement when you

crap on Anthropic. And listen, I think

they deserve some crapping on for some

of the things they do. But at the end of

the day, I'm trying to be objective

here. When I use Opus 5, I get

incredible UX results. I don't need the

extra likes and clicks from saying

Anthropic's the worst. Anthropic's the

worst. I don't care. Opus 5, it's a

great model. Which brings me to the

local model. And listen, even if you

don't have a Mac Studio, even if you

don't have a DGX Spark, this will still

be relevant to you cuz you can still

load local models even if you're only on

a Mac Mini or something like that. Still

important. But right now my go-to model

is Qwen 3 6 27B on the DGX Spark. That's

my go-to for using local at the moment.

I find it is the perfect mix of fast and

intelligent. It seems 40 new AI models

come out every other day. I don't think

you need to be going and downloading

every single one even if it's just a

minor incremental improvement on the

last one. This one has just been my

workhorse, been the kind of the steady

dependable model for me. It doesn't

really mess up. Has tons of support

online. It works really well with the

DGX Spark. I think this is the best

local model to be using. You can run

your Hermes agent off of it if you want.

That's the one downside with chat GPT. I

I think you can actually plug in local

models into the chat GPT app, but I I

it's not something I really want to do.

It feels like Hermes is built for that.

So, if you have local models, I would

recommend setting up a Hermes agent to

run on it. I run this on LM Studio. LM

Studio I find the easiest to kind of

download and load models through. So,

even if you're on like a Mac mini or

some other crappy computer, you can go,

download LM Studio, then go to your

agent and say, "Hey, go on LM Studio and

load the best model for my device." And

it'll load it for you and work. This is

the best local model to be running right

now. I know Qwen 3 8's about to be open

source very soon as well. So, I will

test it out and use it and let you know

how it is. Again, subscribe for that

video when it comes out. Now, let's talk

about other must-haves. These are the

other tools I would recommend

downloading and using and testing out.

For me at the moment, that other tools

you must be using outside the ones I

went through is Buzz. I think Buzz is a

really really really cool concept. For

those that don't know, Buzz was built by

Jack Dorsey. It is an open source

project from him, and it is excellent.

It allows you to work with multiple AI

agents, multiple models all in one. It's

basically a Discord or Slack competitor

built for working with AI agents. And

what's really cool is you can connect it

to your local network, so connect it to

your local compute, and everyone in your

Buzz community can take advantage of

your local compute. It's something I'm

doing right now. I'm connecting my three

Mac Studios, my DGX Spark, my AMD Halo,

and it's all going to power the local

compute in this Buzz community. So, if

you want to take advantage of my local

models, use them, build together with

other people, Vibe Coding Academy is

going to be how you do it. Again, link

for that down below. But, Buzz is really

sick. I think this concept, the

multiplayer AI, is the future. I believe

in the future, it's going to be big for

multiple people to be working together,

working with the same agents, working

with the same context, building projects

together. I'm really betting on that.

So, Buzz, it's free. I download it. You

can plug it the cloud code codex, you

can have your agents all in one place.

It's amazing. I made a tutorial video on

it I think like two weeks ago now. Check

out the channel for that as well. This

was the list. If you are using these

tools for these purposes, you are in the

absolute cutting edge of AI at the

moment. I highly recommend go, download

them all, use them, take advantage of

them, be on the cutting edge. You will

get so much done. I hope this was

helpful. If you want more of these, you

want me to do these monthly, weekly, let

me know down below. I want to know if

it's helpful. I just get so many

questions on what should I be using,

what should I be using? I hope that

answers the question for you. I'm so

grateful you'd watch my videos. It

really means the world, and I'll see you

in the next one.