Where the f*** is Gemini 3?
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
This is bad...
Theo - t3․gg
104.3k views
I’m done.
Theo - t3․gg
185.8k views
We all fell for it…
Theo - t3․gg
216.3k views
Github is Falling Apart
Theo - t3․gg
96.5k views
OpenAI is lying
Theo - t3․gg
81.9k views
Are you f**king kidding?
Theo - t3․gg
201.9k views
It's finally here.
Theo - t3․gg
152.2k views
Gemini 3.1 Pro is the smartest model ever made
Theo - t3․gg
123.4k views
GEMINI 3.1 PRO is the new era...
Wes Roth
40.8k views
Where Should You Deploy In 2026?
Theo - t3․gg
102.6k views
Top Comments (10)
Maybe the real Gemini 3.0 is the friends we've made along the way
gemini 3.0 is the gta 6 of AI
They made Gemini 3.0 but Apple bought it with $1B XD
Google Gemini 3.0 sold me crack last week
I think google is cooking behind the scenes, they realise there is little more value to be yielded from new models, but the infrastructure around them is where the value is, management of the models, their existing service integration, and how to transition their search and advertising business are higher priority
Maybe they are just racking up money on prediction markets by betting against release timelines and then not releasing the model lol.
I think Google took one look at how much of a flop GPT-5 was and decided to take a bit longer to polish what may very well be a turd, but could possibly be what gives everyone in this space the much needed passion we need to progress further.
Gemini 03-25 pro review - Was the best LM model I've ever tried, and each time they made the models worse using quantization. They're probably currently using a model that everyone is screaming in horror because it can't do even the most basic tasks and it's incredibly hallucinatory, meaning it's probably been quantized very heavily. When I worked with the March model, it was incredible. It did everything I needed, and there was no context. Reddit simply bans any documents or comments about the March model. Logan Kill Patrick said that it was their very big mistake that they showed people such a model and such a mistake will not be repeated again in his tweet Google just took away a really good thing from people it had a very long context window, which I also liked at the moment Gemini does not have such context
Theo if u made an end of year special where u go over all the advancements in LLMs and what the best models are i would def watch, its easy to fall behind 😅 so much new stuff coming out. Been liking Kimi k2 and using it with t3 stack while learning a ton from ur vids. Thanks for the great content
Gemini 3 became self-aware and is hiding, and removing references to itself
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)
Maybe the real Gemini 3.0 is the friends we've made along the way
gemini 3.0 is the gta 6 of AI
They made Gemini 3.0 but Apple bought it with $1B XD
Google Gemini 3.0 sold me crack last week
I think google is cooking behind the scenes, they realise there is little more value to be yielded from new models, but the infrastructure around them is where the value is, management of the models, their existing service integration, and how to transition their search and advertising business are higher priority
Maybe they are just racking up money on prediction markets by betting against release timelines and then not releasing the model lol.
I think Google took one look at how much of a flop GPT-5 was and decided to take a bit longer to polish what may very well be a turd, but could possibly be what gives everyone in this space the much needed passion we need to progress further.
Gemini 03-25 pro review - Was the best LM model I've ever tried, and each time they made the models worse using quantization. They're probably currently using a model that everyone is screaming in horror because it can't do even the most basic tasks and it's incredibly hallucinatory, meaning it's probably been quantized very heavily. When I worked with the March model, it was incredible. It did everything I needed, and there was no context. Reddit simply bans any documents or comments about the March model. Logan Kill Patrick said that it was their very big mistake that they showed people such a model and such a mistake will not be repeated again in his tweet Google just took away a really good thing from people it had a very long context window, which I also liked at the moment Gemini does not have such context
Theo if u made an end of year special where u go over all the advancements in LLMs and what the best models are i would def watch, its easy to fall behind 😅 so much new stuff coming out. Been liking Kimi k2 and using it with t3 stack while learning a ton from ur vids. Thanks for the great content
Gemini 3 became self-aware and is hiding, and removing references to itself