1B Gemma downloads ๐ Congrats @googlegemma and thank you.
We've been Gemma builders since day one: trained on it, shipped products on it, ran it in our runtime. After @Qualcomm acquired our startup, we launched GenieX at Qualcomm AI Hub to bring Gemma to Qualcomm PCs, phones, and edge devices. Our story is a Gemma story.
@alexchen4ai got to tell it on stage at @SHACK15sf@cerebral_valley last night alongside @DynamicWebPaige and @clmt. The energy in that room was unreal.
Also got to hang with @osanseviero@o_lacombe@GlennCameronjr@triswarkentin@danielhanchen@jeffboudier and so many others from the Gemma community and team. Warm, caring, and genuinely fun to build with. This is what a developer community should feel like.
Builders: go try Gemma on Qualcomm devices with GenieX now, a few lines of code, NPU-accelerated, and fully open-sourced. Link below ๐
Something I told 14 yo: People are going to stop reading books. I wish this wasn't so, but I fear it is. The silver lining in this cloud is that if you're one of the few people who still read, you'll have a huge advantage over everyone else.
The winning consumer AI will be a lifestyle product. Users love simplicity. 1 subscription, making the best decision for me, and having the access to the best model via API.
Some news I've been waiting to share: Nexa AI was acquired by @Qualcomm this March. And we just shipped our first product together โ Qualcomm AI Hub GenieX
Two years ago, we started with a hypothesis: generative AI should run on and empower every physical device. So we dove right in.
Along the way, we launched the Octopus series of small language models that climbed to the top of open-source download charts. We built the Hyperlink agent, the first local AI agent that could search and understand the full context of every file on your computer. We open-sourced NexaSDK โ inference toolkit that runs any model on any device โ which hit #1 on GitHub Trending. And we got to share stages with, and be featured by, many of the biggest names in the industry. Two years ago, we could only dream of that.
What I want to say most is thank you.
Thanks @alexchen4ai and @zacklearner for the daily inspirations that great entrepreneurs obsess over hard problems until they crack, with strong conviction and resilience. Thank you to every partner who bet on us early and I'll never forget the wins we celebrated together. Thank you every teammate and friend who carried us through the hard nights. The lessons I learned from all of you in these two years are more than in the ten before.
And to my new colleagues at @Qualcomm: thank you for the warm welcome, and for making this launch happen together.
The future is agentic, physical AI that every developer can build. We launched Qualcomm AI Hub GenieX, an open-source GenAI inference engine and SDK: a few lines of code to run frontier models on Qualcomm devices across NPU, GPU, and CPU. This is just the beginning.
If you've made it this far โ reach out! DMs always open โ. I'd love to learn what you're building, help you get the most out of GenieX, or explore ways we can partner.
Learn more in blog:
https://t.co/SwKdBJosgF
Big day for Ollama! When we started, open models and the open source AI ecosystem were in their early days with few believers.
Our belief in open source has never wavered.
With today's fundraising announcement and our 9M+ active builders, weโre ready to scale open models into AI that you can own.
All aboard open models!
๐งต
Less AI plumbing. More AI building.
GenieX is a new open-source runtime that helps developers run and integrate Gen AI models across @Qualcomm devices with just a few lines of code.
Developer preview is now live: https://t.co/2Rbgm7MevY