BREAKING: OpenAI Launches FREE Open Offline Model!
Related videos
AI News: 18 Breaking Stories You Missed This Week
Matt Wolfe
14.8k views
OpenAI Just Killed Sora
Matt Wolfe
35.5k views
AI News: They All Launched the Same Thing!
Matt Wolfe
14.8k views
AI Is Frying Your Brain
Matt Wolfe
31.7k views
Breaking: Google Upgraded Nano Banana Again!
Matt Wolfe
44.7k views
Why People Are Freaking Out About Clawdbot
Matt Wolfe
43.5k views
OpenAI’s Code Red Explained
Matt Wolfe
60.0k views
REVEALED: Sam Altman's OpenAI Is 'MONEY LOSS MACHINE'
Breaking Points
28.6k views
The 11 Best FREE AI Tools of 2025
Matt Wolfe
23.0k views
AI News: New FREE Agent Is Blowing My Mind
Matt Wolfe
184.1k views
OpenAI Releases GPT-OSS: Analyzing the New Open-Weight Leader
Assess the performance of OpenAI's new open-weight model, GPT-OSS, against closed-source leaders and learn how to run these powerful versions locally and privately on your own hardware.
Short Summary
- [00:00:39] GPT-OSS establishes a new standard for open models, widely expected to become the "open-weight king" over current Llama and Mistral models.
- [00:01:21] Local execution ensures complete data privacy because conversations are never processed or sent to external cloud servers.
- [00:04:22] Successful usage requires significant GPU resources, with the 20B model needing 16GB VRAM and the 120B model needing 80GB VRAM.
This overview details OpenAI’s landmark release of GPT-OSS, available under the Apache 2.0 license, demonstrating its competitive strength across coding, reasoning, and specialized exams while providing practical steps for local, offline installation using tools like LM Studio.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Related videos
AI News: 18 Breaking Stories You Missed This Week
Matt Wolfe
14.8k views
OpenAI Just Killed Sora
Matt Wolfe
35.5k views
AI News: They All Launched the Same Thing!
Matt Wolfe
14.8k views
AI Is Frying Your Brain
Matt Wolfe
31.7k views
Breaking: Google Upgraded Nano Banana Again!
Matt Wolfe
44.7k views
Why People Are Freaking Out About Clawdbot
Matt Wolfe
43.5k views
OpenAI’s Code Red Explained
Matt Wolfe
60.0k views
REVEALED: Sam Altman's OpenAI Is 'MONEY LOSS MACHINE'
Breaking Points
28.6k views
The 11 Best FREE AI Tools of 2025
Matt Wolfe
23.0k views
AI News: New FREE Agent Is Blowing My Mind
Matt Wolfe
184.1k views
Top Comments (10)
First video I have seen where someone actually downloaded it, explained the requirements, installed it, and let us see it in action instead of overhyped videos with nothing to show. Thank you!
Anyone remember encyclopedia britannica CDs?
I was broke and without a job and then I found the book called Inside the AI Matrix Vault by accident and tried one of the AI methods in it. Six days later I made my first $200 online. Now i get much more than just that. Throw an eye on that book.
The reason a lot of software developers don't use web-based AI is due to privacy/confidentiality reasons. If you're working on proprietary code, your employer may forbid you from transmitting it outside of the organization, which includes showing any of it to ChatGPT/Claude/Gemini/etc., because there is a risk that they could record it and be able to see the code in question (even if they claim they don't, there's no way to be sure, the data is transmitted, and you're just relying on their word that they don't keep a copy). With Local LLMs though, this is a non-issue, since you can run them air-gapped so that it is literally impossible for them to transmit anything outside of the organization.
We love your show, Matt! Thank you for all you create and share!!! 🎉
*Cries in 8GB of VRAM*
Can't pretend to be king with the Open source community. You're either good or not.
Ok but what does "open-weight" mean? I guess I will just go and ask my AI chatbot...
Thanks for this video. I downloaded the model and tried it but honestly I'm not at all impressed. Not only does it have no long term memory and no large scale document retrieval capability, but even basic conversations were running into token limitations and even a couple of errors (mean it legit just spat out an error code and nothing else).
Your PC specs is what I want to be when I grow up.
Unlock the Data Inside
Turn Videos into Knowledge
- Get FREE 10/day: transcripts, summaries, chats
- Chat with videos, export text & PDF
- $1 free API credit for RAG, chatbots & research
Free forever plan • All features unlocked
Top Comments (10)
First video I have seen where someone actually downloaded it, explained the requirements, installed it, and let us see it in action instead of overhyped videos with nothing to show. Thank you!
Anyone remember encyclopedia britannica CDs?
I was broke and without a job and then I found the book called Inside the AI Matrix Vault by accident and tried one of the AI methods in it. Six days later I made my first $200 online. Now i get much more than just that. Throw an eye on that book.
The reason a lot of software developers don't use web-based AI is due to privacy/confidentiality reasons. If you're working on proprietary code, your employer may forbid you from transmitting it outside of the organization, which includes showing any of it to ChatGPT/Claude/Gemini/etc., because there is a risk that they could record it and be able to see the code in question (even if they claim they don't, there's no way to be sure, the data is transmitted, and you're just relying on their word that they don't keep a copy). With Local LLMs though, this is a non-issue, since you can run them air-gapped so that it is literally impossible for them to transmit anything outside of the organization.
We love your show, Matt! Thank you for all you create and share!!! 🎉
*Cries in 8GB of VRAM*
Can't pretend to be king with the Open source community. You're either good or not.
Ok but what does "open-weight" mean? I guess I will just go and ask my AI chatbot...
Thanks for this video. I downloaded the model and tried it but honestly I'm not at all impressed. Not only does it have no long term memory and no large scale document retrieval capability, but even basic conversations were running into token limitations and even a couple of errors (mean it legit just spat out an error code and nothing else).
Your PC specs is what I want to be when I grow up.