Orange Collective

"Founder Mode Party" is an oxymoron, except for this founder who pulled out their laptop and shipped code in the middle of the party 🫡 Thanks to the portfolio founders, LPs, and YC alumni who joined us during the 2024 YC Alumni Reunion—and the OC members who helped make this possible! https://lnkd.in/gNEhnmcv

You can finetune Llama-3.2 1B & 3B for free on Colab now! 🦥Unsloth AI makes finetuning 2x faster and uses 60% less VRAM with no accuracy degradation. Llama 3.2 (1B) QLoRA finetuning fits on a 4GB GPU, and (3B) fits on 7GB. Vision model support for Meta's (11B) and (90B) coming soon!! Colab conversational notebook: https://lnkd.in/gwHrmXjU Kaggle conversational notebook: https://lnkd.in/gHVQTrxh Also uploaded 4bit pre-quantized weights for 1B, 3B, 11B and 90B to huggingface.co/unsloth. There's GGUFs as well! Unsloth github.com/unslothai also has notebooks for: Continued pretraining, text completion, conversational chat, reward modelling with ORPO, DPO and more! Vision model support coming soon!

This is YC Alumni Week and hundreds of YC founders practiced Founder Mode at tonight’s reunion, showing off our relentless resourcefulness (and fearlessly asking for more bottomless drinks !? ) It was so fun catching up with old friends and meeting new founders. Huge shoutouts to Dave Yen Orange Collective Amy Yan Anna Sun Nowadays (YC S23) for hosting such an incredible event for the community!

Catching up with friends and former YC hoops teammates at the Orange Collective event as part of Y Combinator reunion weekend. An exciting time to be a part of Orange Collective. Great times to come. 🚀

Why is there a disconnect between CRO and CFO offices? How can we provide assurance and guarantee trust? This milestone is a testament to our team's dedication and expertise. We've worked closely with every customer, delving deep into their systems' complexities and each step's friction point. At JustPaid We learned that while time savings are significant—we save about 40+ hours per month for both billing workflows (usage, platform, tiered, discounts, trials, etc.) and collections—most importantly, we provide data reconciliation. How can you create trust? Where do you feel friction? Thank you Y Combinator , Rebel Venture Fund, Orange Collective, DECACORN, Pioneer Fund for your support

The internet posed an existential threat to traditional media and journalism, much like how AI does today, reshaping the industry and how we consume news and media. That’s why we’re excited to back Capitol AI (YC S24), as they help organizations like POLITICO drive the next evolution of media in the age of AI.

Powder users have generated proposals for over $20 billion in assets—and the numbers keep rising. The sheer volume of assets in motion is staggering. 😎 The first billion took 3 months. The next $19B took 6 months!! 🤯 If you're not optimizing the most critical part of your sales process, you're leaving profits on the table. We are happy to help you grow faster! Book a demo here at https://lnkd.in/gDEfkkDh

I'm thrilled to announce that we’ve raised $3.5M Seed led by Gradient Ventures, with the participation of Y Combinator, Ardent Venture Partners, Orange Collective, key executives from Snowflake, Group 1001, Superlinked, Frigade, and Kevin Moore. Today, more than 400 exabytes of data are generated daily. Over 15 million data professionals are leveraging this data to guide their organizations toward their true north. Coming from Snowflake, LinkedIn, and Meta, my co-founder and I have tight bonds with data teams, and seeing them calculate the right product features to build just based on data is nothing but astonishing. However, for the last two decades, unstructured data has largely been left out of this equation. It's incredibly challenging to translate into something machines can use for quantitative analysis. We, humans, process 74GB of data through our eyes and ears every day. We are naturally great unstructured data processors, but unfortunately, we don't scale. But just recently, we've seen how multimodal LLM can understand the pain points from a 45-minute-long user journey video and sense the sadness from a phone call just by the tone. It is a perfect, unstructured data processor that can scale. It is a truly Cambrian moment for the data community because now it's possible to do quantitative analysis on unstructured data! This transformation may take many years because today's data tooling is so deadlocked into structured data. We will not wait for the incumbents. We will build this future, and that's why we're building Roe AI, now. This mission to Mars is going to be thrilling and bumpy, but I’m grateful not to be on this journey alone. Thanks to my co-founder Jason Wang, my family, mentors, friends and fellows at YC, Snowflake, LinkedIn, and UC Berkeley, and, of course, Roe AI’s early customers. You are our solar power 🖖 Live long and prosper! Fundraising blog: https://lnkd.in/eKArsR8s Find time to chat with me: https://lnkd.in/eXyURM6z We're hiring a Founding Senior SWE (L5+) full-time / contractor, a fall or winter intern. Apply here https://lnkd.in/e7kFTEXU Adrien Treuille Pieter Verhoeven Jay Prassl Gu Xie Karthik Murugesan Natchie Subramaniam Thiagarajah Nikolas Moore Tyler Bosmeny Pete Koomen Vivek Raghunathan Sridhar Ramaswamy shrikant shanbhag Yusuf Ozuysal Vladimir Katardjiev Abhishek Nagaraj Adam Blake Rohan Rajiv Emre Kazdagli Alp Artar Christian Mathiesen Namrata Ganatra Jim Fan Darian Shirazi Kyle Duffy Zachary Bratun-Glennon Phil Bronner Dave Yen Benjamin Rogojan Ryan Wexler

MICSI is thrilled to announce that MICSI-RMT has received FDA 510(k) clearance. It requires no training data and leverages the existing information within the MRI exam to remove noise and enhance image quality. MICSI-RMT includes modules for both functional and diffusion MRI, improving activation map accuracy in functional MRI and supporting both weighted linear least squares and Bayesian fitting techniques to facilitate high-quality parametric maps for diffusion weighted imaging. We would like to thank our team and collaborators for helping us cross the finish line. This marks a significant milestone in our journey to enhance the elevate the standard of medical imaging and nudge it into the direction of objective high-quality imaging biomarkers. MICSI-RMT is here to super-charge your radiology practice with cutting-edge MRI noise reduction technology. Reach out to us directly here on LinkedIn or through our contact page to learn more about our product and how it can benefit your practice. #MedicalImaging #HealthTech #Radiology #InnovationInHealthcare #MedicalDevices #DigitalHealth #MRI #RSNA2024 Gregory Lemberskiy, PhD Benjamin Aron

Super excited to announce that we've raised $2.75 million in seed funding to accelerate our work on multi-modal web agents. This brings our total raised to over $4 million. Our LLM systems are live in production today with numerous companies building new web data constrained products. If you have any need for domain specific, highly structured web data, don't hesitate to reach out to us at Reworkd (YC S23) We're also super excited to get to work alongside amazing investors like Paul Graham himself, Nat Friedman, Daniel Gross, SV Angel, General Catalyst, Panache Ventures and many more. Learn more here: https://lnkd.in/gfAYEQfK and shout out Maxwell Zeff for the write up