WndrCo

Founding Partners Jeffrey Katzenberg and Sujay Jaswa joined Caroline Hyde on Bloomberg this week to share more about WndrCo's recent raise of over $460 million for our new Seed and Venture funds. https://lnkd.in/ge2PbkSs

We’re excited to announce that WndrCo has closed over $460 million in new capital for our Seed and Venture funds. Read how this supports our mission in the future of work, consumer tech, cybersecurity, and developer infrastructure in this CNBC coverage: https://bit.ly/3VolPfz

So great to host another AI dinner with my partner Anthony and our friends at Outset Capital - Ali Rohde and Kanjun Qiu. Code-gen (end of software engineering?), LLMs beyond transformers, data curation/pre-training, and applications in computer vision were just some of the fun dinner conversations. And then Kevin Kwok pulled out a stuffed parrot with a 7B model that follows voice instructions...only in San Francisco!! Thank you to everyone who attended - we had a blast! Josh McGrath Aarti Bagul Ulrik Stig Hansen Scott Johnston Raphael Gontijo Lopes Devon Rifkin Tori Seidenstein Michael Makhlevich Itay Shemer Ian Thompson Isaac O. Daniel C. Liem Tina Jiang Evan Ellis Michael Andregg

Jeffrey Katzenberg Sujay Jaswa and Tomás P. giving the Alembic Technologies, Inc. team a masterclass in business and the current state of tech! #startuplife #masterclass #marketingroi

WndrCo is excited to announce our $14M Series A investment in Alembic Technologies, Inc. I'm also thrilled to be joining the company's Board of Directors. “Deploying AI allows companies to increase productivity in ways not possible before. Alembic is using AI developed for scientific research applications to predict ROI from marketing. NVIDIA marketing is using Alembic with great success.” - Jensen Huang, founder and CEO of NVIDIA At WndrCo, we continually ask Fortune 500 CMOs about their top challenges. The most common response is around proving the effectiveness of marketing spend. With Alembic, enterprise brands can now leverage the power of AI and cutting-edge statistics to understand marketing effectiveness and predict future ROI across all of their campaigns (social, TV, radio, out-of-home, sponsorships, etc). This is made possible by groundbreaking techniques originally developed during COVID-19 to trace causes and patterns in disparate public health datasets. “Alembic has a breakthrough solution for a problem I’ve seen throughout my career: Measuring results from marketing initiatives. All enterprise marketing organizations should consider Alembic to solve this problem in marketing." - Jeffrey Katzenberg Looking forward to partnering with Tomás Puig, John A., Lloyd Taylor, Seth Little, Jayden Ziegler, PhD, and the Alembic team on what Jeffrey describes as the "Holy Grail of Marketing": enabling marketers to understand the value of their spend. Joe Montana, ChenLi Wang, Laura Maness, Mark Ghermezian, Aristomenis Capogeannis

🚨 Hiring alert 🚨 We are growing our investing team at WndrCo! - Come work with Sujay Jaswa and me as we partner with the next generation of founders from seed to growth - Learn from our prior experience building Dropbox and DreamWorks and see how we help companies with storytelling (Jeffrey Katzenberg) and supercharging GTM (Justin Wexler) - Looking for curiosity, hustle, analytical thinking, strong personal networks, and a passion for both technology and investing We are back to working in the office every day (in downtown Redwood City), and you will get to work directly with Sujay and me, along with our team in LA/NY - Anthony Saleh, Jeff Nykun, Erika Panico, Armaan Grewal, and Sharon Pan. If you know anyone who we have to hire, tell them to apply below! https://lnkd.in/dz2GTRej

I'm very honored to be among Business Insider's 2023 Rising Stars of Venture Capital alongside some of my favorite peers in the industry. Grateful to the team at WndrCo and our portfolio companies! https://lnkd.in/efVhkyCD

Last night, May Habib (Writer) and I hosted some of the most forward-thinking enterprise leaders for an evening focused on deploying generative AI in the workforce. It was awesome hearing all the perspectives from across different industries such as marketing (WPP, IPG), travel (Clear), professional services (KPMG, Russell Reynolds, Accenture, BlackRock, UBS, Northwestern Mutual), tech (Dropbox, X), media (Forbes, Gannett), and medical (J&J, Cigna).

More fascinating conversations from AI founder dinners. - David Kanter is the founder of MLCommons, the creator and steward of MLPerf benchmarks for training models on different hardware configs. It's a reminder that the advances in AI have deep roots in innovation at the silicon level - And speaking of hardware, Abhinav Venigalla just published a new blog post detailing how MosaicML/Databricks have been scaling up LLM model training on AMD GPUs - Kevin Fischer is building Opensouls/SocialAGI - infusing LLMs with human emotions - "soul". The way he is building this struck me as blurring the boundary between how you a human being develops and how you would train a ML model. - In a similar vein, Nyla Worker at Convai is building realistic NPCs for virtual worlds infused by AI - can't wait to see how immersive games are going to evolve over the next few years Always a pleasure to co-host with Ali Rohde and Josh Albrecht at Outset Capital and thanks to everyone who attended! Nancy Hung Yifan Wu Anirudh Jain Simon Suo Susan Zhang Jennifer Zhou Sam Khosroshahi James Liu Robert Brooks IV David Chouinard Tristan Hume Kevin Meng James Bradbury Tiantian Zha

It feels both exhilarating and dizzying to keep up with all the progress in AI! Some learnings from a couple recent founder dinners: - As excitement from ChatGPT has tempered, there is tremendous enthusiasm from businesses in building or buying "ChatGPT for my data". The new buzzword is Retrieval Augmented Generation (RAG), which includes data formatters and connectors provided by LangChain (Harrison Chase) and LlamaIndex (Jerry Liu), as well as data/retrieval platforms like Chroma (Anton Troynikov). The RAG paradigm enables LLM-powered apps like chatbots and agents to reference and use an enterprise's proprietary data and reduces hallucinations. - Everyone is struggling with the "Evals" problem right now - as you train and fine-tune LLMs, how do you assess their quality in a consistent and scalable fashion? How do you generate the questions to ask, what metrics do you use to grade responses? You can try using another LLM to do Evals (i.e. one model grades another), but that doesn't seem to work well. Doing it manually (which many AI researchers do!) is very time consuming. This reminds me of how grading AP English Literature or SAT Writing is far harder to automate than multiple choice tests. - The most mind-blowing thing I learned is that for some complex reasoning problems, you can take advantage of LLM's stochastic nature by spawning a number of agents, giving the problem to each one with tweaked prompts, and instruct them to work collaboratively on solving it. This actually works better than single agents! It just amazes me that human constructs (teamwork + collaboration) translate so directly to how you get the most out of LLMs. - Great seeing WndrCo portfolio founders Waseem Alshikh (Writer) and Feross Aboukhadijeh (Socket) applying AI to solve very different but impactful problems, from enterprise content to software supply chain! Thanks to Ali, Kanjun, and Josh for co-hosting, and to everyone who joined us. Anvisha Pai Joseph Spisak Sumeet Vaidya Anant Bhardwaj John Wang Pujaa Rajan Robert Wachen Gavin Uberti Noor Siddiqui Nikhila Ravi Simon Last Winnie X. Yaroslav Bulatov Taylor Rogalski