Sancus Ventures

Excited to finally announce Dyna Robotics $23.5M seed round co-led by CRV and First Round Capital. Big thanks to Sharon Goldman for breaking the news! https://lnkd.in/gwx5vCtf Yes, robotics is becoming a crowded field. However, current technical challenges significantly hinder the widespread commercial viability of robots. Most robotic foundation models today operate at just 10-30% of human-level speeds, impacting productivity on high-speed production lines. They also lack generalization capabilities, requiring extensive data collection even for minor environmental or task variations. Additionally, robot hardware remains expensive and insufficiently durable, limiting deployments to scenarios where ROI is easily justified... And these are just a few of the issues. After conversations with hundreds of potential customers, we discovered that people don’t care about fancy designs—they just want robots to perform reliably, maintain high throughput, and deliver clear ROI. To succeed, we knew we had to simplify by deciding what not to do (at least initially): - No legs: Over 60% of manual tasks involve repetitive actions with minimal mobility. - No human-like grippers: Complex finger designs significantly raise costs without proportional benefits. - No unrealistic expectations: Customers prefer robots excelling at one task (instead of many) repeatedly, reliably, and efficiently. We believe mastery of one beats mediocrity in many. We’re laser-focused on achieving general-purpose capabilities by perfecting one task at a time. Today, we’re bringing cost-effective, easy-to-deploy robotics solutions to businesses of all sizes. How do you know you have product-market fit? When your robot looks like shit but delivers so much value that customers happily pay for it.

Starting today, you can build your own custom AI models without collecting labeled data – with the first end-to-end reinforcement fine-tuning platform. DeepSeek-R1 showed the world how RL can solve challenging reasoning tasks, and now we’ve baked these capabilities into an intuitive platform so anyone can leverage self-improving models for their use cases. RFT guides a model with reward functions and unlabeled data in a new interactive experience that’s a game-changer for tasks like code generation, complex RAG and more. Our early results with the platform have blown us away. We used it ourselves to build a 32B-param model to write CuDA code 3x more accurately and faster than OpenAI’s o1 or DeepSeek-R1, that we’re open sourcing today. And we’ve already seeing great traction with early customers. I’m very excited to launch the first version of this experience and give you an early look at what we’re building. Check out how it works in the thread below and let’s shape the future of custom AI together.

We're finalizing our next AI webinar featuring some of our favorite thought leaders in the field. https://lnkd.in/eNiBtkBg Expect engaging discussions on topics like, is my CS degree now worthless? Should I have put it on the blockchain and sold it off in 2021? I’m still searching for a final speaker—ideally, a No Code expert. If you have any suggestions or connections, I’d love your recommendations! Cc: Lovable, Bolt, Vercel, Cursor, Replit, Codeium, Augment Code Thank you to our speakers, Lake Dai, Richard Socher, Ted Nyman! cc: Akkadian Ventures, AIX Ventures, Sancus Ventures, Cased

Great investors don't chase trends—they find signal amid the noise. So what's the framework behind placing the ‘right bets’ in venture investing? Join us next Thursday, March 13 as Lake Dai, Managing Partner at Sancus Ventures, reveals her strategic framework for early-stage investing. Learn how she identifies promising markets, selects startups with real advantages, and partners with exceptional founders—a proven approach that consistently outperforms trend-chasing. With a dynamic background—from leading the first product team at Alibaba to serving as an adjunct AI professor at Carnegie Mellon and investing in venture capital with institutional LPs—Lake offers a unique perspective you won't want to miss! https://lnkd.in/gwE8CtRP

AI models are getting smarter, but also slower. That’s the tradeoff with advanced reasoning models like DeepSeek-R1—they deliver step-by-step logic but at the cost of inference speed. ?? If you're working with large models and want to squeeze every last bit of performance out of them ?? on March 5 at 10 am PT we hold an exclusive live session where we’ll show you how Predibase Turbo ?? unlocks 2x faster inference speeds without compromising accuracy. We believe that we should not have to choose between structured reasoning and fast performance. What to expect: ?? Why DeepSeek-R1 and other reasoning models are slow out of the box ?? How speculative decoding turbocharges inference speeds ?? How you can apply Predibase Turbo to fine-tune models faster than ever Come prepared with questions and use cases for the Q&A at the end! #AI #MachineLearning #LLM #DeepSeek #FineTuning #TurboLoRA #InferenceAcceleration #SpeculativeDecoding

8.4k stars on our deep research github repo in 6 days ?? This is now the most popular open source deep research project. Every AI tool says they do "research", most are pretty disappointing. Aomni is one of the few AI native companies that's actually pushing the boundaries of frontier models.

The future belongs to AI-powered sellers, and we just raised $4M in seed funding led by Decibel Partners to build it. Aomni gives every revenue org a team of custom AI agents that works 24/7 to 10x the output of top sellers. ??Proactively find, enrich and qualify enterprise leads ? Deep research people and companies ??Create hyper-targeted campaigns in minutes ??Research accounts in depth to map orgs and find decision makers Let Aomni do all the tedious repetitive tasks, so sellers can focus on building relationships with customers. We're so grateful to be partnering with incredible investors like Jéssica Le?o, Jon Sakoda, Stacey Wueste, Lake Dai, Lei Du, Jia Li, Paul Rosen, Chris Messina, Nathan Pitzer, and so many other amazing angels! LINK IN COMMENTS BELOW:

For years, I’ve been working with engineering leaders and seeing the challenges they face daily. It’s a tough gig. The ask: * Achieve ambitious goals while navigating the everyday toil of the job * Create resilient and highly effective teams * Encourage and support developers’ professional growth * Keep your sanity while doing more with less Now, there’s real anxiety about how AI is transforming our future. While new AI advances are exciting, the pairing with company announcements about replacing teams with AI and laying off “low performers” creates a lot of stress. The power balance has clearly swung the other direction. That’s what sparked the vision for Exceeds: How do we champion the individual? It’s time to use AI to help people reach their potential, not replace them. Today, I’m excited to share an important milestone in that journey: we recently closed our $4.6M seed round. This round was led by Venrock, with support from SemperVirens Venture Capital , Sancus Ventures, inVest Ventures, and a group of inspiring pre-seed angel investors with amazing leadership experience from LinkedIn, Coinbase, Meta, ServiceNow, Netflix and beyond. In our first chapter, we’ve built a platform that doesn’t blindly focus on lines of code or solely on subjective feedback, but instead leverages your everyday work to build meaningful insight and context to help them succeed. For our early customers, the impact has been meaningful. We’ve turned legacy performance processes that took weeks, and drove higher quality results within minutes. We’re still very early on, and there’s much more of this mountain to climb. We’re grateful to our early supporters who share our conviction, our customers who inspire us to go further, and the community we’re here to serve. ?? Come join us on our journey. Senior Full Stack Engineer: https://lnkd.in/gKt2A9ZU Founding AI and Data Engineer: https://lnkd.in/g6S5FJYs