President, @FireworksAI_HQ . Board Member, World Relief, Yelp. 2 Corinthians 4:7
San Francisco
Joined March 2010
- Tweets111
- Following129
- Followers5.4K
- Likes54
George Hu retweeted
Today we're launching the Specialized Intelligence Index (SII): one destination for real-work benchmarks across industries, built by the teams that use them every day.
Hear from Fireworks co-founder @the_bunny_chen on the importance of specialized benchmarks:
George Hu retweeted
We're a launch partner for Muse Glimmer and very excited to see @finkd and @alexandr_wang note that a version of Muse Spark will also be available as an open weights model.
George Hu retweeted
one of the top cybersecurity models, post-trained from open weights by @depthfirstlabs on @FireworksAI_HQ
long-horizon RL is as much an infra problem as a research one: 100+ turn rollouts, async/pipeline RL for high utilization, train-sampling numerical alignment
Fireworks training platform takes care of infra, so research can move fast
we're proud to support the ecosystem of cyber defenders like depthfirst as AI adoption accelerates
Today we're announcing dfs-large1, our newest cybersecurity model that achieves best-in-class performance on vulnerability detection tasks.
Besides frontier AI labs, only a handful of companies have built specialized models that reach the state of the art in their domain. We're proud to be the first to do it for cybersecurity.
dfs-large1 is built on GLM-5.2 and post-trained with reinforcement learning inside depthfirst's security infrastructure. We evaluated it on depthfirst-bench, our benchmark of long-horizon vulnerability discovery across complex repositories, where it achieves best-in-class performance.
Training improvements have not plateaued yet and we expect additional performance gains as we continue training.
A huge thank you to @FireworksAI_HQ for being an outstanding training partner. Their infrastructure enabled us run large-scale reinforcement learning efficiently and iterate much faster.
dfs-large1 is now in preview within the @depthfirstlabs platform
George Hu retweeted
What a week with Kimi K3 and Opus 5 launch!
We compared these two great models across SWE (480), Algorithmic(100), Terminal (83), the task level quality is very close with Opus 5 being on par or better, and K3's per-task cost is 2x to 4.6x cheaper, using serverless pricing.
Key measurement is per-task cost, not per-token cost. In general, open models tend to be more verbose than close models. Hence the quick study.
We aim to continue to increase per-task serving efficiency (via Fireworks Inference) and quality (via Fireworks Training).
Share what you find out for the tasks you care about. 👇
More details of our study -- lnkd.in/eYtHQR2D
George Hu retweeted
Kimi K3 is live on Fireworks. Day 0, inference and training. US-hosted, and zero data retention.
This is the first frontier open model in the 3 trillion parameter class. It sports 1M context, native vision, and reasoning that rivals the top closed models.
Boom.
George Hu retweeted
Attackers have frontier AI. Defenders need a frontier AI ecosystem—the best open and closed models, force-multiplied by a global community.
During the Hugging Face incident, closed AI blocked essential forensics. An open-weight frontier model helped contain the intrusion.
That’s why we created the Open Secure AI Alliance.
AI security advances when the industry builds in the open, together.
We're introducing the Open Secure AI Alliance with industry leaders to develop new techniques and tools to safeguard software and agents.
By sharing models, tooling and research in the open, we can broaden the community of defenders.
Learn more about the founding members’ contributions: nvda.ws/4pD8Fc5
ALT Logos of various companies including Adobe, Cisco, and Microsoft.
George Hu retweeted
For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.
AI will transform every industry, power every company, and be built by every country.
Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.
The world needs both frontier closed models and frontier open models.
images.nvidia.com/pdf/Open-W…
George Hu retweeted
there are certain costs people have stopped questioning
so when one of them drops 80%, the first reaction is disbelief
open-weight models will give everyone that reaction at scale this year
excited to be partnering with @FireworksAI_HQ on this
George Hu retweeted
Open + closed models = better together.
Our previous research with @harvey showed the benefits of combining a frontier closed model as an advisor agent with fine-tuned, open-source worker agents. Three benchmarks later, the result holds. In an even simpler set up, the open + closed duo delivers superior outcomes and 40-67% lower cost than running entirely on Opus 4.8.
Read the breakdown → fireworks.ai/blog/frontier-o…
George Hu retweeted
Customization continues to improve on another SOTA model.
Own your intelligence, not rent!
Kimi 2.7 is now fully trainable on Fireworks.
Feed your data into Kimi and build a moat that beats frontier models at lower costs. SFT. DPO. RL.
Managed clicks or raw API with huge context, giant LoRA ranks & all the pro knobs.
Get started: app.fireworks.ai/dashboard/f…
George Hu retweeted
Fireworks Training Platform keeps expanding.
Leading US open weight model Nemotron 3 Ultra is now ready for post-training: SFT and DPO via LoRA or full-parameter, on the same infrastructure that serves it.
The model you train is the model you ship: fireworks.ai/train
George Hu retweeted
Introducing @PoeticHQ: a new AI system that executes complex multi-hour tasks with 99%+ accuracy and 10x fewer tokens than agents.
We raised $50M at $500M from Kleiner Perkins, Founders Fund, First Harmonic, and Genius Ventures to build AI that does complex work inside Fortune 500 companies without hallucination.
While code is too brittle, agents are too unpredictable. The work that runs the global economy - anti-money laundering, fraud investigations, underwriting - needs extreme accuracy.
So we built a new kind of software that pairs the flexibility of AI with the predictability of code.
When the world stays the same, Poetic runs fixed code: fast, cheap, identical every time. When the world changes, Poetic uses AI to regenerate its approach and find its way back to the objective.
In one year, we went from zero to an eight-figure run rate as a team of four.
Since then, we’ve scaled the team and executed the highest-stakes processes at AIG, SoFi, and Chime. At SoFi, a large US bank, Poetic reached 99%+ quality on fraud investigations in five weeks.
George Hu retweeted
Frontier models are powerful advisors.
On @harvey's Legal Agent Benchmark, a GLM 5.1 worker using Claude Opus 4.7 as a sparse advisor reached 18/100 all-pass versus 14/100 for Opus alone, at 39% of the cost.
More on the harness design, advisor pattern, and training results: fireworks.ai/blog/open-sourc…
George Hu retweeted
We just hit a major milestone — @FireworksAI_HQ passed $800M annualized run rate and reached 4x revenue growth, apart from Cursor, in Q1.
We invite curious and courageous minds to join us and define new frontiers of specialized intelligence!
Excited to share some big news: I've joined @FireworksAI_HQ President.
I've spent my career at the intersection of great teams, disruptive technology, and massive market opportunity, from architecting @salesforce 50x revenue growth to $5B+, to driving 10x revenue growth at @twilio in five years.
I believe the opportunity at Fireworks is even bigger.
Fireworks is the world’s frontier inference platform, created by the team that built PyTorch, the backbone of modern AI. As the demand for tokens explodes, inference is poised to be one of the biggest markets in history. Now that we’ve hit an inflection point in the quality of open models, I believe no one is better positioned than Fireworks.
Fireworks brings together state-of-the-art training and inference on one platform, enabling companies to transform open models into private models that surpass frontier model performance. Owning your model is the key to being able to control your AI quality, performance, cost, and future. That’s why Fireworks has become the inference platform of choice for AI leaders like Cursor, Vercel, DoorDash, Genspark, Uber, Shopify, and thousands more.
Fireworks is in the early stages of this hyper-growth journey, now processing 30 trillion tokens a day, up more than 100% from just a few months ago.
I want to thank our visionary CEO @lqiao for this opportunity. I’ve had the privilege of partnering with some amazing CEOs like @Benioff and @jeffiel, and I have the same confidence that Lin and I will be amazing partners in building a generational company.
If you're looking for the ride of a lifetime, we're hiring across the board.
Let's bring the 🔥
This #GivingTuesday, I invite you to consider giving to World Relief. In the face of so much suffering around the globe, it’s easy to lose hope. But churches and people like you are making a difference all across the globe in partnership with World Relief. worldrelief.org/blog-5-inspi…
Heart broken to hear the news about the Earthquake that has affected Turkey and Syria. As a board member of World Relief, I am honored to join their efforts to quickly respond to the crisis. Please consider standing with us. worldrelief.org/respond/#cur…
