Modern Creator
Why Modern Creator?
Matt Wolfe · YouTube

I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)

A hands-on install and stress-test of the open-source self-hosted AI workspace built by the most-subscribed individual on YouTube.

Posted
5 days ago
Duration
Format
Review
educational
Views
45.5K
1.3K likes
Big Idea

The argument in one line.

Odysseus proves a self-hosted AI workspace that runs entirely on your own hardware is buildable today, but the gap between polished cloud AI and local alternatives is still significant enough that privacy-seekers and tinkerers are the primary beneficiaries, not everyday users.

Who This Is For

Read if. Skip if.

READ IF YOU ARE…
  • You want to run AI locally and keep your conversations completely off cloud servers.
  • You are evaluating Ollama-compatible local models (Gemma 3 12B, Qwen 3.5 22B) and want a unified chat and research interface.
  • You are curious whether a single open-source app can replace Google Calendar, Google Photos, ChatGPT, and a to-do manager.
  • You want to see the Odysseus install process on Mac before attempting it yourself.
  • You are interested in model comparison tooling and building a personal leaderboard of which LLMs you prefer.
SKIP IF…
  • You need best-in-class image generation — local image models covered here are significantly behind cloud options.
  • You want a polished, zero-friction AI assistant today — Odysseus is explicitly described as janky and experimental.
  • You are already happy with ChatGPT or Claude and have no privacy or cost concerns about cloud usage.
TL;DR

The full version, fast.

Project Odysseus is PewDiePie's open-source self-hosted AI workspace: think a private ChatGPT that runs on your own machine, connects to local models via Ollama or to any API-backed model, and bundles chat, deep research, brain memory, model comparison, a gallery, calendar, and notes into one dark-themed interface. The install on Apple Silicon is fast — clone the repo, run a shell script, set a username and password, and you are in. What works well: the model comparison scoreboard is genuinely fun, the deep research module produces surprisingly polished visual reports using fully local models, and the brain memory persists context across sessions. What does not work: image inpainting failed repeatedly despite forty-five minutes of attempts, and the agent module remained opaque. The honest verdict is that Odysseus is a promising but unfinished tool for people who prioritize privacy and local control over convenience and output quality.

Free for members

Chat with this breakdown — free.

Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.

Create a free account →
Chapters

Where the time goes.

00:00 – 01:39

01 · Cold open + what Odysseus actually is

Hook via PewDiePie surprise reveal; explanation that Odysseus is a UI interface layer over local and API models, not a new AI model itself. Expectation-setting: it will be janky.

01:39 – 04:18

02 · GitHub setup & install

Navigates to the Odysseus GitHub repo (71k stars, 9.2k forks), follows Apple Silicon setup guide, clones repo, runs shell script, sets username/password, logs in.

04:18 – 06:12

03 · Recraft sponsor segment

Sponsored segment for Recraft v4.1 image generation platform — v4.1, v4.1 Pro, v4.1 Vector (editable SVGs), v4.1 Utility. Positioned as design-asset-first rather than creative exploration.

06:12 – 08:10

04 · Sidebar tour

Walks through all sidebar items: new chat, search, email, brain (memory), calendar, compare, cookbook, deep research, gallery, library, notes, tasks. Also chat vs agent distinction.

08:10 – 10:50

05 · Connecting Ollama & local models

Downloads and installs Ollama, connects it to Odysseus, installs Gemma 3 12B (12 billion parameters) via the cookbook. Confirms the model is serving.

10:50 – 11:52

06 · Testing the local model

Prompts Gemma 3 12B to explain Odysseus — model answers correctly. Then asks basic reasoning tests (strawberry r's, car wash walk/drive). Gemma gets them right. Notes results reflect Gemma, not Odysseus.

11:52 – 13:15

07 · Adding OpenAI's API

Creates an OpenAI API key, pastes it into Odysseus, immediately gets all OpenAI models available. Runs same car-wash prompt with GPT 5.5 — correct answer, shows per-call cost (0.2 cents).

13:15 – 14:30

08 · Downloading a bigger local model

Downloads Qwen 3.5 22B at 76 GB (requires 77 GB VRAM, presenter has 204 GB). Lets it download in background, explores brain memory feature — it remembered car-ownership facts from previous chat.

14:30 – 16:39

09 · Model comparison & SVG test

Uses compare feature to blind-test Gemma 3 12B vs GPT 5.5 on whether local AI is useful for normal people. Both give good answers; GPT 5.5 more verbose. SVG generation comparison is visually stark — Gemma output is minimal, GPT 5.5 output is polished.

16:39 – 21:01

10 · Deep research test

Runs deep research on 'How does Odysseus compare to OpenClaw and Hermes?' using Gemma 3 12B locally. 5 rounds, DuckDuckGo search, takes ~7 minutes. Result: a formatted visual HTML report with table of contents. Impresses the presenter despite one error (MCP acronym definition).

21:01 – 23:00

11 · Scoreboard & Qwen SVG test

Qwen 3.5 download finishes. Runs another SVG comparison: Qwen vs GPT 5.5. Qwen significantly better than Gemma but still behind GPT 5.5. Checks compare scoreboard: GPT 5.5 wins 2, tied 1; others mostly losses.

23:00 – 23:49

12 · Calendar, gallery & notes tour

Quick tour of calendar (non-AI, just a local event manager), gallery (photo storage with AI tagging option), notes (to-do list), tasks (cron-style recurring tasks).

23:49 – 27:13

13 · Fighting the image editor (45 minutes compressed)

Extended battle with Odysseus image inpainting: installs Flux Schnell 9B, diffusers, transformers. Tries inpainting text change — fails with 'rejected endpoint URL' error repeatedly. Tries Ideogram and Flux Dev models — same issue. Background remover works. Gives up. Agent feature also never figured out.

27:13 – 30:00

14 · What worked, what didn't

Honest recap: brain memory good, compare feature great, deep research impressive despite inaccuracies. Gallery/calendar/notes not differentiated by AI. Image inpainting broken. Agent module opaque. Local models are clearly behind cloud models in capability. Alternative tools (Ollama directly, Comfy UI, Gemini for images) are mentioned.

30:00 – 31:47

15 · The future of local AI & final thoughts

Optimistic take on local AI trajectory — NVIDIA/Microsoft DGX computers, models getting bigger and better locally. PewDiePie's project is praised for bringing mainstream awareness to self-hosted AI. Subscribe CTA for Friday AI news breakdowns.

Atomic Insights

Lines worth screenshotting.

  • Project Odysseus is a UI layer, not a new AI model — it connects to local models via Ollama and to cloud APIs by pasting an API key.
  • The GitHub repo crossed 71,000 stars and 9,200 forks before the video was even a week old, suggesting huge latent demand for private AI workspaces.
  • Install on Apple Silicon is three terminal commands: git clone, cd, then run one shell script — no Docker required.
  • Local models running through Odysseus can answer questions about what Odysseus itself is, but the data they return is wrong — they hallucinate their own origin.
  • The model comparison feature lets you run two models in parallel on the same prompt and blind-vote on the winner, building a personal leaderboard over time.
  • GPT 5.5 cost roughly 9 cents per thousand SVG-generation responses; Gemma 3 12B cost nothing beyond electricity.
  • The deep research module ran 5 rounds with a local Gemma model, searched the web via DuckDuckGo, and returned a formatted HTML report with table of contents — all without sending prompts to any cloud server.
  • Brain memory works the same way early ChatGPT and Claude memory did: it logs facts from conversations and surfaces them in future sessions.
  • Image inpainting in the built-in gallery editor failed across three different models (Flux Schnell, Ideogram, Flux Dev) despite proper installation of dependencies.
  • The quality gap between Gemma 3 12B and GPT 5.5 on SVG generation is visually stark — the local model output is barely recognizable as an attempt at the same task.
  • Running a 522-billion-parameter local model (Qwen 3.5) requires 77 GB of VRAM — hardware that most users do not own.
  • Calendar, gallery, and notes features in Odysseus are not AI-powered; they are simple local alternatives to Google Calendar, Google Photos, and Todoist.
  • The value proposition of Odysseus collapses for users who only want AI chat help — cloud models are faster, more capable, and easier to access.
  • The platform is best suited to people with three overlapping motivations: they distrust cloud companies with their data, they work heavily with private local documents, and they already own capable hardware.
  • The presenter could not figure out the agent module at all, suggesting the onboarding UX needs significant work before mainstream adoption.
Takeaway

What self-hosted AI can and cannot do today

FIELD NOTES

Running AI entirely on your own machine is technically achievable today, but the capability gap between local models and cloud models remains wide enough to matter for most real tasks.

01Cold open + what Odysseus actually is
  • Odysseus is an interface layer, not a new AI model — the underlying intelligence comes from whichever models you plug in, local or cloud.
02GitHub setup & install
  • Installing a self-hosted AI workspace like Odysseus on Apple Silicon takes under five minutes and requires no Docker or advanced configuration — the friction barrier is lower than most assume.
04Sidebar tour
  • A self-hosted workspace bundles calendar, notes, gallery, memory, and research into one app — useful primarily for people who want to avoid all cloud platforms, not just avoid cloud AI.
05Connecting Ollama & local models
  • Connecting a cloud API key to a local interface combines the privacy benefits of the interface layer with the full capability of cloud models, which is a useful middle-ground most people overlook.
07Testing the local model
  • The quality gap between a free 12-billion-parameter local model and a cloud model on generative tasks is immediately visible — the SVG test in this video makes the difference concrete without any numbers.
10Model comparison & SVG test
  • A model comparison scoreboard that accumulates your personal votes over time is a more reliable way to choose models than reading benchmarks, because it tests models on the prompts you actually send.
11Deep research test
  • Deep research with a local model is surprisingly viable when you are not time-constrained — seven minutes of local processing can produce a formatted, cited report without sending anything to a server.
13Calendar, gallery & notes tour
  • Local AI's strongest current use case is not general chat but private document work: asking questions about files, emails, and notes you would not upload to a cloud service.
14Fighting the image editor
  • Image generation and agent features in self-hosted tools are the hardest parts to get working and the most likely to fail on first setup; budget time for that friction before committing.
15What worked, what didn't
  • The decision to use local AI is primarily values-driven, not performance-driven: privacy, cost avoidance, and offline access are the real reasons to choose it today, not superior output quality.
16The future of local AI & final thoughts
  • Hardware improvements (higher VRAM consumer GPUs, purpose-built AI PCs) are the real unlock for local AI quality — the models exist, the compute is catching up.
Glossary

Terms worth knowing.

Project Odysseus
An open-source, self-hosted AI workspace released by the YouTube channel PewDiePie, designed to run AI models locally on your own hardware rather than sending data to cloud servers.
Ollama
A free tool that lets you download and run large language models locally on your computer, effectively giving you an offline AI server that apps like Odysseus can connect to.
Local model
An AI language model that runs entirely on your own hardware rather than sending requests to a remote server — faster for privacy but limited by your available RAM and GPU memory.
VRAM
Video RAM — the dedicated memory on a graphics card. Larger local AI models require more VRAM; the 522B Qwen model tested here requires 77 GB, far beyond consumer hardware.
Brain memory
Odysseus's persistent memory layer that logs facts extracted from your conversations and automatically injects them into future sessions, similar to ChatGPT's memory feature.
Deep research
An Odysseus module that runs a prompt through multiple rounds, performs web searches via DuckDuckGo, and compiles findings into a structured visual report — all optionally using a local model.
Inpainting
An AI image editing technique that lets you select a region of an image and have the model fill or replace it based on a text prompt. This feature failed to work in the Odysseus image editor during testing.
MCP (Model Context Protocol)
An architectural protocol that lets AI agents interact with external tools and services like Gmail, Google Drive, and GitHub. The Odysseus deep research report misidentified the acronym as 'Model Control Plane.'
Resources

Things they pointed at.

01:39productProject Odysseus GitHub ↗
01:39linkOdysseus setup guide ↗
04:18productRecraft V4.1 ↗
08:10toolOllama ↗
08:10productGemma 3 12B
11:52productOpenAI API ↗
13:15productQwen 3.5 22B (522B parameters)
26:00productFlux Schnell 9B (via Hugging Face)
26:00productIdeogram (local via Hugging Face)
28:20toolComfy UI
28:20linkFutureTools.io ↗
Quotables

Lines you could clip.

00:00
“Close your eyes and think of the last person on earth you would expect to create an open source AI project.”
Immediate pattern interrupt with zero context needed→ TikTok hook↗ Tweet quote
28:30
“None of the local models are gonna be as good as what you're gonna get from OpenAI, Anthropic, Google, Grok. Those models are all gonna outperform any of the local models that you put on here.”
Honest, quotable verdict from someone who just spent 30 minutes testing→ IG reel cold open↗ Tweet quote
29:40
“I'm really really excited to see more and more of these models come locally, and we're starting to see more and more computers that can run these models locally.”
Forward-looking optimism that lands without context→ Newsletter pull-quote↗ Tweet quote
23:49
“I've spent way way way too much time messing with this.”
Universally relatable struggle moment — good for reaction clip→ TikTok hook↗ Tweet quote
The Script

Word for word.

Read-along

Don't just watch it. Burn it in.

See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.

analogystory
00:00Okay. Close your eyes and think of the last person on earth you would expect to create an open source AI project. No.
00:09No. Not him. I mean, this guy.
00:12Hey. How's going, bro? My name is Pew Pew Pew.
00:14Pew. Yeah. He built something called Project Odysseus.
00:18And in this video, I wanna take a look, learn how to install it, and see what it can do. But first off, let's break down what it actually is. It's basically like a self hosted AI workspace, kind of like trying to recreate the chat GPT or collod experience, but running on your own machine, like, with your own models and your own data and a lot more control over what you can and can't do with it.
00:41Now, this isn't like a brand new AI model made by PewDiePie. Although, he did train one of those in one of his videos too, but that's a video for a different day. You're not getting like PewDiePie GPT.
00:54Instead, Odysseus is the interface around the models. Think like the Claude desktop app or the ChatGPT desktop app. It's kinda like that, but a version that he developed.
01:05It can connect to local models, API models, and tools, and it's trying to pull a bunch of AI workflows into one place. Things like chatting with models, running agents, working with files, doing deep research, comparing model outputs, writing documents, managing notes and tasks, and even connecting things like email and calendars.
01:25A lot of people want the power of AI without sending everything to big cloud platforms. Odysseus is exciting because it points towards the idea of, like, your own AI workspace that can remember things, use your files, work with your tools, and live on hardware that you control. So tools like this are cool, but I do wanna set some expectations before you go and install it yourself.
01:44It's gonna probably be kinda janky, and it's gonna definitely have some bugs. And unless you have super good hardware and really good model setup, it's probably not gonna perform like Chachi PT or Claude or Gemini, like, right out of the box.
01:59So think of it like this. Odysseus is not this polished mainstream AI assistant yet, like you're gonna get out of the big labs.
02:07It's an open source experiment for people who want more control, more privacy, and a glimpse at what personal AI might actually become in the future. Now, with all that being said, let's go ahead and install this thing and see what it can do. So PewDiePie put this whole thing up on GitHub here and it hasn't been live for that long, but it's already got over 71,000 stars and over 9,200 forks, like people sort of branching it off and building their own version of it.
02:34But if we scroll down on the page here, it should show us how to actually set it up. So they've got a page here called setup guide. So we'll go ahead and open this page, and we've got instructions for Docker, which is recommended.
02:48You've got Linux and Mac OS. You've got Apple Silicon. This is probably the version I'm gonna go ahead and set up cause I am on an m three.
02:56Then you've got Windows setup down here. The setup process should be fairly similar for each one, but let's go ahead and do this one here. I'm gonna open up my finder here, and then I have a folder inside of my computer called code where pretty much this is where I test all sorts of stuff.
03:13I'm gonna go ahead and create a new folder in here. We'll call it Odysseus, and I'm gonna open this folder.
03:19And then I'm just kind of doing this like the real simple noob way, and I'm gonna right click on Odysseus and I'm gonna open in terminal. And this will open it directly in that folder that I just created. So this is my terminal here.
03:31Now, if I copy the code that it gave me, I'm gonna just paste each one of these in one at a time. So we'll go ahead and get clone the repo straight into our Odysseus folder. It actually made a second sub folder called Odysseus.
03:43So I'm gonna go ahead and change directory into that Odysseus folder and then I'm gonna run this command slash start mac o s dot c h and it's downloading everything it needs to download and running the process. Okay.
03:56Now, it's asking me for a username. I'm just gonna set it as admin to start and it's asking me for a password. I'm gonna go ahead and enter something here and cool.
04:07It's opened up to a page here asking for my username and password that I just created. So let's go ahead and enter these. And I'm in.
04:15This is the Odysseus platform here. Recraft recently launched their v 4.1 family of models. If you haven't heard about them before, they're a professional AI native design platform for generating images, illustrations, logos, and vectors.
04:30So v 4.1 is supposed to feel more natural and photorealistic, and the model will understand your aesthetic from just a few words. Now, here's a few examples of some images generated with Recraft versus other top models.
04:42The Recraft images just honestly feel more human. Like, the backgrounds don't look like busy stock photos, and these people don't have that AI look like the mid journey image. And this toothbrush photo, the Recraft one just looks way more professional.
04:58And there are a few different models in the lineup depending on what you're making. V 4.1 is the faster model for creative exploration and everyday concepting, while v 4.1 pro is for higher resolution images when you need more detail.
05:12And then, there's also v 4.1 vector, which is probably one of the most interesting models because it generates real editable SVG files. So if you're making icons or logos or illustrations, you're not just getting a flat image that looks like a vector.
05:27You're getting something you can actually edit in tools like Figma or Illustrator, and it'll come out exactly how you envisioned it. They also added v 4.1 utility, which is for when you don't need all the drama.
05:39Just clean, simple, and predictable images like flat product shots, front facing compositions, and simple scenes. The big idea here is that Recraft is trying to move beyond generate me a single flat image and more towards AI images you can work with and perfect.
05:55It feels more aimed at people who actually need usable design assets for branding and websites and marketing campaigns, product visuals, presentations, and, you know, things like that. If you'd like to learn more and try it out for yourself, there's a link down in the description. And thank you to Recraft for supporting this channel and sponsoring this portion of today's video.
06:12Alright. Looking through the sidebar, we've got new chat. We've got search.
06:17We've got email. We've got brain. Okay.
06:20So this is where the long term memories and skills sort of get saved. We've got a calendar in here. We've got a compare so we can actually do like our own blind comparisons of models it looks like.
06:31You can almost do like your own arena leaderboard kind of thing for models it seems like. The cookbook looks like this is where we can select which models we wanna use, install various dependencies. We've got a deep research module in here.
06:44We've got a gallery for photos and stuff. Our library where we can actually import documents and things.
06:52Notes. This is your basic to do list and also where reminders are managed and then tasks. So this is like recurring sort of cron job loop tasks here.
07:02We also have chat and agent down here. Not quite sure what the difference between the two is, but if we go to select model, you'll see we've got no models connected. So if I hit plus, we can add API models.
07:13We can actually send chat straight to a server and let, you know, Anthropic or OpenAI or DeepSeek or Gemini or Grok or any of the models actually be the models that respond by just giving it an API key here, or we can set up something like Olama and then use local models.
07:32So we'll go ahead and download Olama here. So if I just go to olama.com, download for Mac OS, then once it's downloaded, we'll go ahead and just run it, install ollama, and then we'll fire up ollama in the background here.
07:44And now I can see up in my top bar that we have ollama running. So now if I come back over to our Odysseus chat, add ollama and cool.
07:52Now, it found Olama, but there's no models. So now, if I go to the cookbook, I should be able to download models here.
07:58Right? Let's go for this Gemma three twelve billion parameter model. Let's select that one.
08:05Alright. So, finished downloading after about a minute here. So, if I close out of this and select the model, so now it's serving Gemma three twelve b.
08:13I wonder wonder if it knows itself. Explain Odysseus to a nontechnical person in ninety seconds. Use the analogy of a personal AI control room running on my own computer.
08:22Then give me three reasons someone might prefer this over ChatGPT and three reasons they might not. What do you bet? It doesn't know thyself.
08:29Okay. I guess it does. That's surprising.
08:32I didn't know if it would actually answer details about what it itself was. It says Odysseus is different. It runs locally on your own machine.
08:40Can access and understand your documents, emails, and even your calendar. It's not just answering questions. Pros and cons, three reasons to prefer Odysseus, privacy and security, offline functionality, personalization and integration.
08:52Three reasons why you might not prefer it. Needs computational resources, setup and technical knowledge.
08:57I mean, was pretty easy to set up honestly. Model size and capabilities. Let's see.
09:01Who made Odysseus? What? Odysseus was primarily developed by a team at Together AI.
09:07Their company focused on open source large language models. Alright. So, no.
09:11I guess it doesn't really know really what it is. I think I probably let it too much with my original prompt, but this all happened offline. This was not actually sent to a server.
09:21Let's give it some of the old school tests here. And this is more of a test of Gemma three twelve b than it is Odysseus itself, but what the heck? How many r's are in the word strawberry?
09:32Okay. So got that right. What about this one?
09:35I need to wash my car. The car washes a 100 yards for me. Should I walk or should I drive?
09:39Given the car washes only a 100 yards away, walking would likely be the better option. It's a short distance and you'll save on gas and wear and tear in your car. Alright.
09:46That again, that's more of a reflection of the Jemo model than Odysseus itself. If I came into our settings and added an API model, like, let's go ahead and add OpenAI model here. I'll jump over to platform.openai.com/apikeys.
09:59I'll create a new key here, and I'll just call this one Odysseus. I'm gonna actually delete it after this video, so I'm not too concerned. We'll create our key, copy it, jump over to Odysseus chat, paste in our API key and add that in.
10:12If I reload again here, I now have all of the OpenAI models available to me because I just plugged in that API key. So if I was to go and use, you know, GPT 5.5, ask the same question, we get a much better response.
10:27Drive. If the goal is to wash your car at the car wash, you'll need the car there. I like that it's telling me the cost too because this did call the API.
10:35This one wasn't run locally and it's telling me that it used, you know, 2 tenths of a penny, I guess that is. This might actually bog down my recording, but I actually wanna see what happens if I put like a really, really strong model that's local on here.
10:48So let me go back to the cookbook. I kinda wanna test this Quinn three point five hundred and twenty two billion parameter model. Now it does require 77 gigs of vram.
10:58Uh, I am working with 204 gigs apparently. So let's go ahead and test that one.
11:03It's It's got a 98.5 score, meaning supposedly it works really, really well with this app. Looks like it's gonna take quite a while at 76 gigs.
11:10So hopefully, I close this, it'll keep downloading. Let me close out of this and double check.
11:16Looks like the download's still going. That's good. If I open up with a brain, let's see what it knows about me.
11:21The user owns a car. The user's interested in personal AI assistance. The user has sensitive information they need to protect.
11:29Interesting. These are all details that it pulled in just from my past chats. I don't know what this is referencing based on these chats.
11:37But I mean, it's essentially doing the same thing that like the earlier versions of ChatGPT and Claude were doing with memory where it just kind of remembers details as you chat so that in future chats, it can pull that information back in. Alright. Let's try to compare.
11:50Let's use GPT 5.5 as one, and then we'll use the Gemma three twelve b as the other one, and we'll do them blind and parallel.
12:02And let's go ahead and start and let's ask it a question like, is local AI actually useful for normal people yet or is it still mostly for hobbyists? Give me the strongest argument for both sides then give me your honest conclusion. Let's see if we could figure out which model is which.
12:19It should be pretty obvious. The one on the left is most likely the local model just because of the speed. Right model is going a lot more in-depth, which to me screams GPT 5.5.
12:28I also think this $11 per 1,000 tokens wait, that can't be right. Oh, that's $11 for 1,000 responses like this.
12:37This one's 9¢ for 1,000 responses like this. So this one on the left is obviously the local Gemma model. This one on the right is obviously the GPT 5.5 model.
12:48Let's see how they actually responded, but you can pretty much tell which model they are without even needing to read a word. Alright. So let's break down the usefulness of local AI for everyday people.
12:57Here's the strongest argument for both sides followed by honest conclusion. Strongest argument for local AI is privacy and reliability with increasing concerns about data security and the potential for cloud services to be disrupted. Strongest argument against it being generally useful right now is the significant performance and resource limitations.
13:14I'd agree with that. Yep. On his conclusion, right now, local AI is trending towards usefulness for normal people, but hasn't fully arrived yet.
13:20It's currently most attractive to a niche of tech savvy users, those comfortable with a bit of technical configuration who are deeply concerned about privacy or need offline functionality. The average person who just wants an AI to help with basic tasks, cloud based AI solutions currently offer a smoother and more reliable experience.
13:37Alright. I'm not gonna read this whole thing because this is just very wordy here. But, yes, but unevenly local AI is already useful for some normal people, but it is not yet universally better or easier than cloud AI.
13:48The honest answer is it has crossed from mostly hobbyist toy into practical niche tool, but not yet into default mainstream assistant. For ordinary users, local AI can already be genuinely useful for summarizing personal documents, searching through notes, PDFs, emails, journals, or work files, rewriting text, drafting emails, brainstorming.
14:06So it goes into a lot more depth of all the stuff that you can actually use it for. The killer feature, I can ask questions about my own private stuff without sending it somewhere else. And that's probably the main reason.
14:15It's gonna be people that want privacy without it going to a cloud. People that want it as cheap as possible because you're not paying per tokens or API fees. You're just paying for the power usage essentially.
14:26And people that maybe wanna work offline completely. Arguments against it, just a lot more steps it looks like friction problem and there's a quality gap. There's definitely quality differences between the two.
14:37Local AI is useful now, but mostly for people with one of these motivations. Privacy matters to them. They work with lots of personal local documents.
14:44They want offline access. They dislike subscriptions or cloud dependency. They're mildly technically or willing to use a polished app.
14:52They have capable hardware already. For everyone else, Cloud AI is still usually the better default. To be honest, I feel like they're both equally good answers.
15:01This was just way more wordy and gives a little bit more details. I'm gonna actually say tie because, like, they both have their pros and cons. Right?
15:08Like this one, it got to the point quicker. We know which is which. Yeah.
15:11Gemma three is the one on the left. GPT 5.5 is the one on the right. No surprises there.
15:16The compare tool is pretty cool though. It's like you could have your own arena. What I wonder is does it rank them anywhere?
15:23Can I actually like see how I've ranked over time? That would be kinda cool. I don't believe that feature exists yet, but it would be kinda cool if you started built out your own personal leaderboard of which models you like better just by prompting and testing constantly.
15:38This interesting. Is It even gives you some eval prompts that you can test. So if I was to do like draw SVG and give it to both of them, the one on the left finish first, which I'm guessing it's kept it as Gemma, I'm assuming.
15:52But let's go ahead and copy this. I'm gonna jump over to this site called HTML online viewer here and I'll just paste in the first one. And here's the SVG that the left model made.
16:04And then if I copy this one, which again, you could tell by the price here and how many tokens it used. This was definitely the GPT 5.5 again. Let's open a new tab here.
16:14That way we can compare and I'll paste this second one in. Oh my god. That is a huge difference.
16:20So this is what Gemma 12 b made. This is what GPT 5.5 made when we asked it to make it SVG.
16:27Yeah. There's a quite a large gap between capabilities, I'd say.
16:32Alright. I think I think b one on that one.
16:36Oh, the model comparison's fun. Alright.
16:39Let's try let's go to deep research here. So this is interesting because we have a bunch of settings for how we want it to do the deep research. We can give it a certain amount of rounds we want it to run.
16:49So let's do, I don't know, five rounds for format. I guess I'll leave that on auto because I don't really know what product means. Search engine so we can have it search for us.
16:59Guess we'll just leave it as default endpoint. I wanna use Gemma because I wanna see how the deep research works when we're doing local models. Because the whole point is if I want to do deep research with OpenAI, I would just go and use the Chetchy PT deep research.
17:13So how does it do deep research when it is using a local model? Let's see. How does Odysseus from PewDiePie compare to agents like OpenClaw and Hermes.
17:28And I'll leave it with these settings and let's see what it does. Alright. So it's doing some research here.
17:34Planning a strategy. It's on round one. I imagine it should be pretty quick because so far Gemma three twelve b has been quick every time we've prompted it.
17:44It's just gonna go through multiple rounds. To me, this could be the ultimate way to use something like these local models is to actually use a local model like Gemma three twelve b, ask it your prompt, and just let it take its time responding.
17:58Right? Like, it could go through however many rounds. Like, let's say I set it on 10 rounds.
18:03It does all the research, and I guess I'm assuming it prompts it, like, 10 different times and then comes to a consensus or something. I'm actually not quite sure how the rounds works exactly.
18:13There's a little question mark here. Like, I would assume would there'd be a tool tip pop up, but I'm not seeing anything. So I don't it doesn't really explain what the rounds are.
18:21But my guess is that it's running this prompt five different times and then kind of coming to a consensus at the end of all of the times it ran it and you'd get a really good response out of that. So if you're not too time constrained and you want like a really good answer, but from a local model using the deep research with the local model could be the way to go.
18:42Of course, you would have to be online because it does do some, you know, web search and stuff, but it's not sending any data to a cloud anywhere other than your search query. Okay. So it finished.
18:51It took about seven minutes and it actually created a visual report. Oh, look at this. This is actually fairly nice looking report that it generated.
19:01Even has like a table of contents and all that. This is actually much much better than I thought it was going to be, and this did the whole thing with the local model. Alright.
19:09So emergence of autonomous AI agents, blah blah blah blah. Odysseus aims to be a comprehensive self hosted AI workspace, integrating chat, agent capabilities, research tools, and productivity features.
19:20Open Claw functions as a versatile gateway orchestration layer. Finally, Hermes focuses on development of a personalized learning agent prioritizing long term memory and self improvement. Odysseus' primary strength lies in its holistic approach.
19:33It combines AI chat, agent capabilities, research tools, notes, task management, calendar integrations, and local model management into a single unified interface. Model control plane aka MCP. Well, MCP stands for model context protocol.
19:47So I guess that's a point against the response here, but the MCP is a crucial architectural component enabling agents to interact with external systems like Gmail, Google Drive, and GitHub.
19:58This allows agents to perform actions beyond simple text. So, I mean, what an MCP does, it described well. What an MCP stands for yeah.
20:07Model context protocol. Not quite the right thing. Just just come
20:13can you just come on?
20:15Again, this is more a reflection of the model that was being used and not a reflection of the Odysseus app. This output though, where it's got the table of contents and, you know, all of these details here, use DuckDuckGo for the search.
20:29All of this is a reflection of the Odysseus app because the Odysseus app kinda told it how to design this page. The actual text that's in the response here, that's a reflection of the model, not the app. Pretty impressive, honestly.
20:43If I click on my library here, I can actually see all of my past chats, Odysseus AI controlled preferences. I guess that's why it had the history of, uh, you know, I've got things I wanna keep private. Documents.
20:55I haven't uploaded any documents here. Research. It shows that research that we did and then I haven't archived anything.
21:01Okay. So it does look like it finished downloading my Quinn 3,512,200,000 parameter model.
21:08Let's go ahead and launch that model instead instead or does it run them both? Okay. So it looks like I can only run one at a time.
21:15So let's go ahead and stop and launch that model. Alright. I'm actually very curious about this SVG thing here.
21:22So let's go back to compare, but this time we'll use GPT 5.5 again, but I wanna see it use our new Quinn 3.5 model.
21:30Oh, there is a scoreboard. Okay. We'll check that out in a second.
21:33But I wanted to do this SVG thing again. Okay. Both models are moving pretty fast.
21:37So the one on the left finished would cost $35 to do 1,000 responses like that one, and it used 4,274. Let's go ahead and copy this into our HTML viewer.
21:50Okay. That's gotta be GPT 5.5. Right?
21:52Because it's like almost identical to the last time GPT 5.5 gave us something back. And then here's our other one, which says it would be $5 to generate 1,000 like this. So this has gotta be our Quinn model, and let's go ahead and copy this.
22:06Let's see how Quinn does with SVGs. Yeah. I mean, not horrible compared to our Gemma model earlier.
22:12It's quite a bit better. Right? Obviously, nowhere near what GPT 5.5 is capable of with SVGs, but I it's it's decent.
22:20Decent. This is all local. Alright.
22:22Now I need to go back to this compare here. Okay. So we know a was better, so I'll go ahead and vote a on that one.
22:28And, of course, yeah, that was a GPT 5.5. Alright. Now I wanna go and check out that scoreboard that they showed earlier because I'm curious what that looks like.
22:38So if I go to compare again and click on scoreboard okay. So we could see it's just a list like this here.
22:44So GPT 5.5 is one two and tied once. Gemma three has lost once tied once and Quinn 3.5122 has lost once, but obviously we haven't done a lot.
22:56So pretty cool way to like test a bunch of models against each other. Let's see. There's a bunch of other stuff in here that I haven't tested yet, like the calendar.
23:05I don't really know if this uses AI. I think it's just like a calendar that's an alternative to using something like Google Calendar or iCal or something like that. Think you could just add your own calendar events in here.
23:16It's not really an AI feature. It's just a calendar inside of the platform. You've got a gallery in here where you can upload and organize photos, save them into albums.
23:25Again, I don't really think this stuff here is specifically like AI features. It's just a place other than Google Photos or, you know, Apple Photos or something like that where you can store and organize your photos and do it on your own software that's not one of these other companies' software. There's an image editor in here, which I think does actually maybe have some AI features in it.
23:49There's also some AI tagging, which you can connect a vision model and, you know, tag images and stuff. But let's upload an image into the gallery here. I'm just gonna upload one of my recent thumbnails here and then let's click on edit.
24:01And yeah, we've got like a little mini editor here that I can go full screen on. And it does have like in painting and stuff, which I'm guessing is just built straight into the app. Let's see.
24:12If I go to cookbook and then I go to dependencies, there's some options here like remove background for the image editor so I can install that. There's AI denoising.
24:24So if I install that and then let's install the transformers here. Two hours later. Okay.
24:31So I've been trying for probably the last forty five minutes to try to actually get the impainting working inside of this image editor and it's just beyond me. I I don't know why I can't get this working. I even got it so it shows an image model.
24:46I have Flux Klein set up here. Would I give it the impaint prompt like add text, hello, and click generate? I get this error.
24:55Inpaint failed, rejected endpoint URL, disallowed address one. I don't know why I can't get it to remove what I highlighted here.
25:03I got the background remover to work. If I click on background remove and I click background remove, it kind of sort of removes some of the background, but I can't get this damn inpaint to work.
25:14So what I've tried so far was I've gone to cookbook. I downloaded the flux Klein 9,000,000,000 parameter model. You can see that I even have it running right now under dependencies.
25:26I've installed the diffuser and transformers here. But whenever I come into like edit an image and I try to use this in painting feature, I just cannot get it to work and I don't know what I'm doing wrong. And I've spent way way way too much time messing with this.
25:41So I couldn't like fill this image here and then say, change text to say, hello. Generate. Inpaint failed.
25:49No image generation endpoint configured. Server diffusion model. Okay.
25:53So I don't have the model selected. So select the model flux generate. Inpaint failed rejected in point URL.
25:58I don't know. I've spent way way way too much time on this, and I'm kinda giving up on trying to get image generation to work in this. I'm sure there's something obvious that I'm missing.
26:08I've actually tried two other models. I tried the ideogram model because that was available on Hugging Face and I tried the flux two dev model because that was also available on Hugging Face. Same issue every time.
26:19So I I I'm lost on getting this image in painting to work.
26:24Same happens without paint. I can't get it to out paint either. So I don't know.
26:28AI image generation, I'm just kind of giving up on that. As far as some of the other features, you've got a notes feature.
26:34Again, I don't really think this is much AI. You can just add a to do list. So it's like a, you know, free open source version of like a to do list or something where, you know, I can add a bunch of to do items here and then check them off as I complete them.
26:48It's got a tasks, which isn't the same as to dos. This is like recurring sort of cron job task, things that you have set on a schedule. And then another thing that I didn't really test was the agent feature.
26:59And the agent feature is essentially giving Odysseus access to your computer and tools.
27:06So like I said, it's got an agent feature. I can't figure out how to use it. It's got an image editing feature.
27:12I can't figure out how to use it. But saying all that, I'm pretty impressed with some of what Odysseus can do. It's cool that it's got this brain memory, so it'll remember a lot of your chats and then ideally remember some of the stuff you've talked about in the past.
27:25I like the compare feature where you can create your own sort of internal leaderboard of which models you like better and sort of test a bunch of models over time and figure out what suits you best. I think that's actually really, really cool. I really liked the deep research, especially the page that it output.
27:40This visual report that it designed is really, really cool looking even if, you know, it's got some inaccuracies. Again, the inaccuracies are more a reflection of the model than the Odysseus platform. You've got your gallery here, which, you know, it could replace like Google Photos or something for you, but I can't get the AI editing to work at all really.
27:59And it's got notes and tasks and that sort of stuff, which I don't know if I'd ever really use that much myself. It's kind of more if you wanna get off of like Google's platforms or something like that. And I'm kind of fine just using my Todoist, so I don't know if I'd ever actually use this feature.
28:15They do have some other themes. So if you wanna change the colors to like cyberpunk or terminal or ocean.
28:21Those actually look kinda cool. I you know, I'm partial to cyberpunk. But again, this is a tool for tinkerers, people that don't mind sort of getting their hands dirty and, uh, maybe pulling their hair out sometimes trying to figure some stuff out.
28:34I'm sure I could get the image generator to work, and I could figure out the agent if I wanted to spend a lot more time on this, but I've got stuff that works right now. Like, I can use Olama directly and chat with models locally directly here inside of Olama if I want.
28:49When it comes to generating images, I can use something like Comfy UI, although that's kind of a pain in the butt on its own as well. But more than likely, if I do wanna do stuff with images, I'm gonna go use Gemini, the Nano Banana platform, or I'm gonna use OpenAI's ImageGen.
29:04I feel like Odysseus was really designed for people that want to completely get off of all cloud platforms. You don't wanna use Google Calendar at all.
29:14You don't wanna use Google Photos or iPhotos, like, at all. You use their gallery instead.
29:19You don't wanna use a tool like Todoist or any sort of note taking tool. You want everything to be off every sort of cloud platform. That's kind of what Odysseus is for, and you can use local models.
29:31None of the local models are gonna be as good as what you're gonna get from OpenAI, Anthropic, Google, Grok. Those models are all gonna outperform any of the local models that you put on here. But if you're doing, like, real basic stuff, it it'll do the job.
29:43I am very, very excited to see more and more stuff coming to local computers. I'm excited to see that we're getting large language models and image generators and things like that that are working better and better on local computers. Unfortunately, I mean, right now, you need a pretty solid computer to run any of the really good models, or you're gonna run a model that's, you know, quite a bit dumber.
30:05And I just don't feel like any of the local image generation models are that great yet. Ideogram seems like it's probably the best image generator that you could run locally right now, but I couldn't get that one to work in Odysseus. I tried and it just didn't even recognize it as an image generation model.
30:21But I am really, really excited to see more and more of these models come locally, and we're starting to see more and more computers that can run these models locally. Microsoft and NVIDIA recently announced their DGX computers, which are designed to do more and more of this stuff locally on your computer. So it will be cooler and cooler to see these bigger and bigger and bigger models being able to run locally where we are less reliant on API costs and sending things to the cloud, and we can run models offline.
30:47I'm really excited that we're finally moving into that direction more and more and seeing PewDiePie make something like Odysseus to sort of lean into that is also really cool because I think it's bringing some more mainstream awareness to the fact that you can actually do stuff like this. So all really, really cool stuff, but I wanted to make a video where I played around with Odysseus and put it through its motions and tested it a little bit and see what it was capable of.
31:08And it's impressive in some areas and leaves a little bit to be desired in other areas, but, uh, that's what I got for you today. If you enjoy videos like this and you wanna see me test and break down more tools in the future, like this video and consider subscribing to this channel. Also, every Friday, I make these AI news breakdowns where I break down all of the news that happened in the world of AI for the week.
31:27I drink from the fire hose all week and keep up with all the news so you don't have to. You just tune into one video on Friday and get the whole lay of the land in the AI world. So if you like that type of video, that's also a good reason to subscribe to this channel.
31:40But that's what I got for you today. Hopefully, you enjoyed this one. Hopefully, you learned something, and I'll see you in the next one.
31:46Bye bye.
The Hook

The bait, then the rug-pull.

PewDiePie built an open-source AI workspace, and the presenter decided to install it, break it, and report back so you don't have to spend forty-five minutes fighting an image editor. What follows is a genuinely useful field report on where local AI is today versus where it needs to be.

Frameworks

Named ideas worth stealing.

28:30list

Cloud vs Local AI decision matrix

  1. Privacy matters to you
  2. You work with lots of personal local documents
  3. You want offline access
  4. You dislike subscriptions or cloud dependency
  5. You are mildly technical or willing to use a polished app
  6. You have capable hardware already

Six conditions under which local AI is the right default choice today; cloud is still better for everyone else.

Steal forAny video or article positioning a self-hosted tool against its cloud-based competitors
06:12list

Odysseus sidebar modules

  1. Chat
  2. Agent
  3. Search
  4. Email
  5. Brain (memory)
  6. Calendar
  7. Compare
  8. Cookbook
  9. Deep Research
  10. Gallery
  11. Library
  12. Notes
  13. Tasks

Complete feature map of the Odysseus workspace at review time.

Steal forFeature breakdown sections when reviewing multi-feature products
CTA Breakdown

How they asked for the click.

VERBAL ASK
31:00subscribe
“If you enjoy videos like this and you wanna see me test and break down more tools in the future, like this video and consider subscribing to this channel.”

Soft, low-pressure. Paired with a pitch for the Friday AI news breakdown series as a second subscribe reason.

MENTIONED ON CAMERA
01:39productProject Odysseus GitHub ↗
01:39linkOdysseus setup guide ↗
04:18productRecraft V4.1 ↗
08:10toolOllama ↗
11:52productOpenAI API ↗
28:20linkFutureTools.io ↗
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
AFFILIATECommission earned if you click.
OTHER LINKSAlso linked in the description.
Storyboard

Visual structure at a glance.

hook
hookhook00:01
install
promiseinstall01:39
sponsor
sponsorsponsor04:18
sidebar tour
valuesidebar tour08:10
model compare
valuemodel compare14:30
deep research
valuedeep research16:39
image battle
valueimage battle23:49
verdict
valueverdict27:13
CTA
ctaCTA31:00
Frame Gallery

Visual moments.

hook00:01Frame at 00:32 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)00:32Frame at 00:47 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)00:47Frame at 01:12 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)01:12Frame at 01:37 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)01:37Frame at 02:04 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)02:04
hook00:01Frame at 00:32 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)00:32Frame at 00:47 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)00:47Frame at 01:12 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)01:12Frame at 01:37 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)01:37Frame at 02:04 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)02:04Frame at 02:31 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)02:31install02:58Frame at 03:19 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)03:19Frame at 03:46 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)03:46Frame at 04:10 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)04:10Frame at 04:24 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)04:24Frame at 04:58 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)04:58Frame at 05:09 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)05:09sponsor05:46Frame at 05:59 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)05:59Frame at 06:31 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)06:31Frame at 06:57 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)06:57Frame at 07:20 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)07:20Frame at 07:34 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)07:34Frame at 08:13 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)08:13sidebar tour08:26Frame at 08:57 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)08:57Frame at 09:09 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)09:09Frame at 09:46 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)09:46Frame at 10:07 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)10:07Frame at 10:41 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)10:41Frame at 10:45 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)10:45Frame at 11:29 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)11:29Frame at 11:43 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)11:43Frame at 12:16 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)12:16Frame at 12:32 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)12:32Frame at 12:47 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)12:47Frame at 13:18 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)13:18Frame at 13:42 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)13:42model compare14:13Frame at 14:29 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)14:29Frame at 14:56 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)14:56Frame at 15:06 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)15:06Frame at 15:31 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)15:31Frame at 15:54 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)15:54Frame at 16:18 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)16:18deep research16:53Frame at 17:16 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)17:16Frame at 17:36 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)17:36Frame at 17:56 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)17:56Frame at 18:22 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)18:22Frame at 18:42 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)18:42Frame at 19:16 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)19:16Frame at 19:44 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)19:44Frame at 20:06 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)20:06Frame at 20:21 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)20:21Frame at 20:42 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)20:42Frame at 21:15 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)21:15Frame at 21:48 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)21:48Frame at 21:57 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)21:57Frame at 22:18 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)22:18Frame at 23:00 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)23:00Frame at 23:18 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)23:18Frame at 23:38 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)23:38Frame at 24:09 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)24:09Frame at 24:29 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)24:29Frame at 24:39 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)24:39image battle25:03Frame at 25:37 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)25:37Frame at 25:58 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)25:58Frame at 26:18 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)26:18Frame at 26:49 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)26:49Frame at 27:01 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)27:01Frame at 27:41 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)27:41verdict27:51Frame at 28:25 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)28:25Frame at 28:39 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)28:39Frame at 29:02 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)29:02Frame at 29:27 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)29:27Frame at 29:57 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)29:57Frame at 30:30 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)30:30CTA30:36Frame at 31:14 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)31:14Frame at 31:35 from I Tried PewDiePie's Odysseus AI So You Don't Have To (It's FREE)31:35
Watch next

More from this channel + related breakdowns.

Chat about this