https://developers.googleblog.com/bringing-gemma-4-12b-to-your-laptop-unlocking-local-agentic-workflows-with-google-ai-edge/According to Google, the new model delivers performance close to much larger AI systems while requiring significantly less memory. The company says Gemma 4 12B can run locally on devices equipped with just 16GB of VRAM, making advanced AI more accessible to developers, researchers and businesses. The launch highlights a growing trend across the AI industry: bringing powerful AI models directly to personal computers instead of relying solely on remote data centers.Gemma is Google's family of open AI models built using technology and research from its Gemini program. The new Gemma 4 12B model contains 12 billion parameters and has been designed to handle multiple types of information, including text, images and audio. Unlike traditional AI systems that focus only on text, Gemma 4 12B can understand visual content, process audio inputs and perform advanced reasoning tasks. This makes it suitable for a wider range of applications, from software development and content creation to research and automation. Google says the model is available under the Apache 2.0 licence, allowing developers and organizations to use, modify and deploy it with relatively few restrictions.One of the most significant technical changes in Gemma 4 12B is its new unified architecture. Traditionally, multimodal AI systems use separate components known as encoders to process images, audio and text before combining the information. >Google says Gemma 4 12B removes the need for separate multimodal encoders.Instead, the model processes different types of information through a unified architecture. According to the company, this helps improve efficiency while reducing memory requirements and computational overhead. The result is a model that can deliver advanced multimodal capabilities while remaining small enough to run locally on modern hardware.
>GoogleWoke leftist AI is gay af. lame!
>>536368768Google's AI is without a doubt the shittiest of all of them. It hallucinates constantly and is so contained that it cannot even really hold a decent conversation, or present information in a way that isn't cliffnotes. I suspect that it also is unable to reason on a level that provides any sort of textual analysis. This is probably just the same shitty AI but using less resources.
8.5 katrillion dollars funneled into the company just so they can prevent the ai from saying "boobs" or "niggers" btw"ai safety" is a form of psychosis.
>frontier models for me, not for theeso pretty much what we've been saying.you're going to pay premium to run copilot (gimped) on local while openai, antrophic et al make the big (read: only) bucks serving you slopwaregg no re
>>536368768>AI IS DOING THINGZ!Nigger, I just want a fucking border wall and the 2000s back.
>>536371489not gonna happen summerfag
16 gb vram? Nothing burger . Get back to me when u can run this shit on 16gb ram cpu.
>>536369564nigga these companies are crawling with 3rd world DEI hires that are itching for the chance to sell the code for a $500 giftcard
>>536371569>16 gb vram? Nothing burger . Get back to me when u can run this shit on 16gb ram cpu.OP was lazy (and possibly retarded). I have a measley FOUR GB VRAM laptop with 32GB (36GB of memory, total) and I can run ANY LLM up to about 14 GB because I need space for the OS. I can still have a couple of browser tabs open and shit but the O/S + Drivers + etc eat up about 8-9GB on my system. I've been using KoboldCPP almost since it was released and it's pretty easy to get started with. And it runs just fine, speed-wise. Depending on how much the O/S takes up on that theoretical 16GB machine you mention, it's likely Gemma 4 12B with like a 4_K_M quant GGUF (or a little bigger or smaller) would work fine. With thinking turned off, these Gemma 4 model variants are pretty speedy.
>>536369564How expensive is it to train an AI?Why does no one train an uncensored based and red-pilled AI and release it to the world?Could a Free AI be trained distributively, like SETI used to distribute its data for analysis to home computers worldwide during their downtime?
>>536369138>Google's AI is without a doubt the shittiest of all of themYou haven't tuned your temps et al. because you're retarded so, of course, the model will 'perform' poorly.
>>536368768didn't some nobody make an AI that did exactly this at the start of this year? who cares about Google? their search functions don't even work anymore, and those are simple booleans, which were solved back on punch-card machines.
I trust Russian and Chinese AI over Kike AI any day. It's smarter and more efficient too.
>>536368768>16GB VRAMNice, thats me!
>>536368768They cant even get Chrome to work without hogging all the RAM and you believe this shit.
>>536368811>be me>trying out new Woke leftist AI>hey Gemma, how to rid of kudzu?>6 gorillion tokens spent lecturing me on pronouns and kalergislop>end of essay, 1 sentence: "idk maybe glyphosate?":|
