Engineering Lead, Gen AI inference @ Qualcomm | Nexa AI co-founder & CTO (acquired by Qualcomm). Opinions are my own.
Santa Clara, CA, US
Joined October 2021
- Tweets336
- Following2.6K
- Followers3.6K
- Likes586
Pinned Tweet
New GenieX release from @Qualcomm AI Hub will be live today, bringing @liquidai 's LFM3 to the Qualcomm NPU / GPU / CPU for fast, private, on-device AI. 🚀
Give it a try and let us know what you build!
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence.
> 6 model instances, across text, audio, vision, and omni modalities, generative and decision models.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding
> 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages.
> 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language.
> 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input.
> fine-tune on 1 GPU.
> a new tri-hybrid architecture
> architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒
Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device.
1/ 🧵
After being mocked for falling behind in AI for over a year, Google is back at #1. I think three things made the difference:
1. Distribution and real-world feedback. Search, Gemini, Workspace, and Cloud give Google access to massive real-world usage and feedback loops that most AI labs can’t match.
2. Full-stack control. From TPUs and JAX/XLA to model training and inference, Google can optimize hardware and software together, improving efficiency and iteration speed.
3. Recursive self-improvement. Better RL, reasoning, and coding capabilities mean Google’s models can increasingly help optimize their own training infrastructure and accelerate future research.
Google’s biggest advantage isn’t just having the best model today. It’s having the infrastructure and feedback loops to keep building better ones.
Zack Li retweeted
LFM3 is here! this is an insanely elaborate release, get yourself a coffee ☕️ spend a few hours going through the blog where we tell you everything about it.
The amount of useful intelligence you can pack inside 2B parameters is really insane!
the highlights for me:
> LFM3 architecture got out of our propriety automated hardware in the loop architecture search algorithm optimized for 14 different types of processors. This is exactly the vision we embarked on: building an intelligence fabric on the fragmented device world.
> LFM3-VL-2B is truly a beast performing on par with GPT 5.6 Luna and Gemini flash ultra! at 2B parameters, wtf!!
> It is sooo easy and chip to fine-tune LFM3 instances. you can now truly build specialized intelligence for anything.
> LFM3 is enjoyably fast to train, and deploy.
> LFM3 omni 2B packs all modalities in one model enabling reliable spacial intelligence on device!
> It is now the era of products. lets build together 💫🚀
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence.
> 6 model instances, across text, audio, vision, and omni modalities, generative and decision models.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding
> 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages.
> 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language.
> 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input.
> fine-tune on 1 GPU.
> a new tri-hybrid architecture
> architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒
Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device.
1/ 🧵
Zack Li retweeted
So excited to get this out:
LFM3 is finally here!
The first release in the LFM3 family contains six models in 2B-class size:
• LFM3-1.7B: to run thousands of tiny agents in parallel
• LFM3-VL-2B: now with video support
• LFM3-Audio-2B: Speech in AND speech out
• LFM3-Omni-2B: our first ever omni model
• d1-2B and d1-Omni-2B: with new types
They are all built on the same shared LFM3-1.7B-Base.
These are super powerful to fine-tune them into tiny specialists in your domain.
Weights on Hugging Face: huggingface.co/collections/L…
Introducing 𝙻𝙵𝙼𝟹: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence.
> 6 model instances, across text, audio, vision, and omni modalities, generative and decision models.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱: text, 256k context window.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱: vision-language model with video-input support for spatial understanding
> 𝙻𝙵𝙼𝟹-𝙰𝚞𝚍𝚒𝚘-𝟸𝙱: tool calls, voice chat, reasoning in 18 languages.
> 𝙻𝙵𝙼𝟹-𝙾𝚖𝚗𝚒-𝟸𝙱: an omni model combines vision, audio, and language.
> 𝚍𝟷-𝟸𝙱 and 𝚍𝟷-𝙾𝚖𝚗𝚒-𝟸𝙱: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input.
> fine-tune on 1 GPU.
> a new tri-hybrid architecture
> architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs.
> 𝙻𝙵𝙼𝟹-𝚅𝙻-𝟸𝙱 is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters.
> 𝙻𝙵𝙼𝟹-𝟷.𝟽𝙱 is #1 on the Pareto frontier of device intelligence 𝚙𝚒𝚙𝚎𝚝𝚝𝚎.𝚕𝚒𝚚𝚞𝚒𝚍.𝚊𝚒
Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device.
1/ 🧵
Try Liquid AI's LFM3 model with the latest GenieX v0.10.0 — just one command!
geniex infer LiquidAI/LFM3-VL-2B-GGUF
Download GenieX here: geniex.aihub.qualcomm.com/en…
Zack Li retweeted
Tomorrow is a huge day!
Something we've been building quietly for a while. Stay tuned, the 24h clock starts now 🚀
I recently came across a talk by @jaschasd from @AnthropicAI at @Stanford. The phrase “last days of the Anthropocene” immediately caught my eye, and this particular slide is what really stayed with me.
As AI capabilities advance exponentially, two years of dedicated research could be overtaken by a newer foundation model before the project is even finished.
It raises a profound question: It’s no longer just “Can I solve this problem?” but “Will it still matter by the time I solve it?”
My biggest takeaway: Time is precious. Live every day fully, pursue your biggest ambitions, and focus on the hardest problems that truly matter—especially those whose value grows as AI becomes more capable.
@jaschasd 's talk: threadreaderapp.com/thread/1…
Fantastic, cannot wait to learn and try the fully open sourced version!
Join us today at AGI House for the Google Gemma Hardware Hackathon, co-hosted by Google DeepMind and AGI House! 🚀
Explore on-device AI, edge AI, and physical AI, and build exciting applications powered by Gemma.
📅 Oct 24 | 10 AM – 10 PM PT
📍 AGI House, Hillsborough, CA
Register: luma.com/6ccl0jw5
Congrats, @AndrewDai! Leaving a senior role at @GoogleDeepMind to build something from scratch takes real courage and ambition. Exciting to see @ElorianAI pushing the boundaries of visual reasoning with Glimpse. Wishing you and the team all the best!
Tesla just took “car launch” to a whole new level. 🚀 The Roadster isn’t just pushing the limits of speed—it’s challenging gravity itself. ⚡
Zack Li retweeted
Excited to be part of this! 🚀
Join us on Oct 24 for the Google Gemma Hardware Hackathon at AGI House, bringing together Google DeepMind, NVIDIA, and Qualcomm to push the boundaries of on-device AI.
$45K in prizes — can’t wait to see what everyone builds!
$45K cash prize pool + hardware! We are hosting the most cash prized Google Gemma build day at AGI House with @googlegemma @GoogleDeepMind @nvidia @Qualcomm on 10/24!
RSVP: agihouse.org/events/the-goog…
Cc @alexchen4ai @zacklearner @XeniaBulatov @UFBots @JessicaChenX
Excited to be co-hosting the Google Gemma Hardware Hackathon with @GoogleDeepMind and @agihouse_org on Oct 24!
AI is moving beyond the cloud and into the physical world. The next wave of AI innovation will be powered by on-device intelligence, robotics, and Physical AI. Looking forward to seeing builders turn cutting-edge AI models into real-world applications!
🔗 agihouse.org/events/the-goog…
cc @agihouse_org @RockyYu
$45K cash prize pool + hardware! We are hosting the most cash prized Google Gemma build day at AGI House with @googlegemma @GoogleDeepMind @nvidia @Qualcomm on 10/24!
RSVP: agihouse.org/events/the-goog…
Cc @alexchen4ai @zacklearner @XeniaBulatov @UFBots @JessicaChenX
Great insights from @xia_char
The real battle is shifting from who has the best model to who can deliver intelligence at the best cost, latency, and scale.
Nvidia is reportedly in talks to acquire Reflection AI. On Vercel’s AI Gateway, open-weight share of tokens went from 7% in December to 56% in August, while taking just 14% of spend. By one day in mid-September, it hit 78%.
Anthropic shows the split. From June to July, its spend share rose from 61% to 64% while its token share slipped from 32% to about 30%. Pricing power is holding for now.
The usage data tells us an interesting pattern: open wins the token volume, closed keep the dollars.
The driver is agents looping in the background, planning, retrieving, and revising. Developers route that repetitive middle to open models at lower cost and save closed APIs for the calls where a mistake is costly.
So how would the token volume and dollars flow in the next few years?
