Ferocity Capital

Thank you very much Panos Panay and the entire team at Amazon & AWS who's been supporting NinjaTech AI since our launch on May 2024: Matt Garman Rohit Prasad Swami Sivasubramanian Frédéric Deramat Nafea Bshara and everyone else especially Sam Lisowski Zoheb Munshi Raghvender Arni Firat Elbey Deepak Mohan Danny Meurer Harrison Wheaton Paul Bernard and 50+ more friends. Also, NinjaTech AI would not be happening without my brothers Arash Sadrieh and Sam Naghshineh and our sister Divya Chirasani along with our wonderful Ninja team including our key investors and supporters especially Kris V. Thomas (Seungpill) Jueng Faraj Aalaei Laszlo Bock Paul Muret Todd Stavish Ali Tamaseb Spencer Holbert and others. Lastly, we just launched our Deep Researcher 2.0, check it out ?? ?? ?? : https://lnkd.in/eKRFGJ9A

It was a pleasure spending time with 30 school district leaders across the United States last week. I served on a panel along with Cindi Chang, who's doing incredible work preparing kids in Nevada for the workforce. Thanks to the National Center on Education and the Economy for hosting the event. When asked about my prognosis for the future, I shared these things and school leaders generally seemed to agree. Today, AI has done more harm than good. It has made it easy for students to cheat on homework, breaking the current homework paradigm. In the next 2-3 years: AI will be used by teachers in a widespread way for curriculum creation and professional development. Some states have rules against using AI for curriculum, but common sense should prevail in a majority of states. In the next 5-8 years: AI video tutors will be commonplace, where students talk with a lifelike avatar for tutoring - and results will be pretty accurate. Costs for tutors can go down 5x-10x using this... and tutors will be available for dramatically more students than today. 10 years from now: In-person human teachers will still reign supreme. Teaching is all about human connection and mentoring.

It's that Empire State of Mind ???? Some of our team gathered at our NYC HQ as we continue to accelerate our growth and expansion ?? We talked all-things Cloud Detection and Response, strategized our go-to-market, and maybe even had a few slices of pizza along the way... But the best part? We had the opportunity to be in Times Square while our Series B announcement appeared front and center ?? There's no stopping us—we've got those New York lights on us and we're ready to dominate! ?? #Cybersecurity #CloudDetectionandResponse #CloudSecurity

?? ????????????????????-?? ??.??: A Breakthrough in Cost-Effective, High-Performance Multi-Gear Compound AI Reasoning Model.? Available today on: https://www.myninja.ai/ ?? ??????????????: We can now ?????????? ?????????????? ?????????????????? ???????? ?????????????????? ?????? ?????????????? ???????? ??????????????????????: SuperAgent-R 2.0 Matches OpenAI O1 in AIME 2024 but achieves this with a fraction of the cost (~$0.49 per 1M token vs. OpenAI O1 $60 per 1M token). It runs securely & safely on Amazon Web Services (AWS) end-to-end and it scales beautifully. ?? ???????? ?????? ???????? ???????? ????????: https://lnkd.in/ep2t6TTq ???? ?????????? ???????????????? & ???????????????? ?????????????? ??????????????????, ???? ?????? ?? ?????????????????? & ?????????????????? ???????? ???? ???? ???????? ????????, ???? ???? ?????? ???? ???????????????? ?????? ?????? ?????? ???? ???? ?? : Ninja AI aims to deliver powerful yet accessible AI that enhances productivity affordably. After two years developing an agentic productivity system with features like Deep Research and meeting scheduling capabilities, we identified a critical need to improve our AI's decision-making intelligence. We discovered that "step-by-step thinking" significantly improves accuracy in complex tasks. While recent reasoning models apply this approach successfully, they present key limitations for our needs: 1. Cost prohibitions: Using models like OpenAI's O1 could cost $0.75-$2.25 per task, making it economically unsustainable. 2. Performance constraints: More affordable options like DeepSeek R1 require expensive hardware, suffer from high latency, and struggle with general capability tasks. 3. Lack of customization: Existing models don't allow the fine-tuning necessary for our specialized productivity applications. These limitations led us to develop SuperAgent-R 2.0, our own reasoning system designed to be fast, affordable, and customizable for sustainable agentic workflows. ?? ?????? ?????????????????? ??????????????:? ???????? ????????: 86.67% ????????: 68.69% ????????????????????: 1633 ??????????????????????????: 54.78 ?????????? ????????: 86 We'll continue to improve this (we already have version 2.5 in the works), but this now unlocks us to launch our upcoming Deep Researcher 2.0 that agentically can solve super complex queries (trip planning, financial analysis, scientific queries, etc). We'll announce that in less than 2 weeks ?? ?? We'd love your feedback as it helps us improve this new compound model. We analyze every single user rating (incorrect to Superb): https://www.myninja.ai/ NinjaTech AI Amazon Web Services (AWS) AWS Startups AWS AI

Happy to have contributed to this year's #MilitaryBalance on Japan's defense capability. It was also great to wide coverage of the launch event by NHK last evening and to see former IISS colleagues appearing on Japanese. Impressive research and analyses at an important time. Congratulations on the launch!! ????

This week, OSI Washington - Office of Science and Innovation is participating in Bifrost Summit in Silicon Valley. The annual Summit brings together innovators, entrepreneurs, and investors from the Nordics and Silicon Valley to accelerate innovation. This year, focus is on critical and emerging tech, such as dual-use technologies, AI , and quantum. Sweden's Science and Innovation Counselor Maria Brogren moderated a panel on the Silicon Valley ecosystem for dual-use startups, where panelists Julie Zinamon, Kris V., Scott Tait, and Hannah Weber shared their insights on opportunities - and what to avoid. Thank you to our great panelists and thank you Nordic Innovation House - Silicon Valley for organizing! #scienceinnovationsweden #innovation #nordiccollaboration #ecosystem #dualuse

Bifrost Summit 2025 Week Starts Now! ?? We’re thrilled to kick off Bifrost Summit 2025 – a week dedicated to New Critical and Emerging Technologies! One of the highlights? Our Main Conference on Tuesday, March 4th, at Nasdaq Entrepreneurial Center, where we’re bringing together an incredible lineup of speakers who are shaping the future of tech. Just look at these stars! ? This is a speaker lineup you don’t want to miss. ? Susan Lindner, CEO Innovation Storytellers ? Moa Persdotter, Program Director of ‘Acceleration of Deep Tech Startups, Vinnova ? Ari Jonsson, Supervisory Board Member, NATO Innovation Fund (NIF), COO, Videntifier ? Heather Jo Richman, Gordian Knot/BMNT ? Jenny Elfsberg ?? Director & Head of Innovation Management Division, Vinnova ? Derek Kenmotsu, Senior Advisor, Global Partnerships, Defense Innovation Unit (DIU) ? Julie Zinamon, MBA, Co-founder & COO, Louva ? Hannah Weber, Co-Founder, VP Strategy & BD, HOPO Therapeutics ? Kris V. Managing Partner, Ferocity Capital ? Scott Tait Executive Director, National Security Innovation Catalyst ? Maria Brogren, Counselor for Science, Innovation, and Higher Education, Embassy of Sweden in Washington, D.C. ? Benjamin Cedric Larsen, PhD, AI/ML Project Lead, World Economic Forum ? Valerie C., Co-founder and CEO, Persimmons, Inc. ? P?r H?kansson, VP, Corporate Strategy, Nordic Semiconductor ? Babak Soltanian, CEO, Tayen.ai ? Darryl Weatherspoon, CTO, Aquabyte ? Sohini Roychowdhury, PhD, Engineering Leader, Ads and Personalization, Meta ? Magnus Bergman, Founding and General Partner, Luminar Ventures ? Lu Zhang, Founder and Managing Partner, Fusion Fund ? Natalia Burina, Venture Partner, Storm Ventures ? Shankar Chandran, Partner, Walden Catalyst Ventures ? Sam Stanwyck, Group Product Manager, Quantum Computing, NVIDIA ? Kasra Nowrouzi, Head of Hardware, Advanced Quantum Testbed, Lawrence Berkeley National Laboratory ? Lars Jensen, Founding Partner, Scale Capital ? Jonathan King, Co-Founder and Chief Scientist,?Atom Computing ? Daiana Fobian Nielsen Davis, Senior Innovation Advisor, Innovation Centre Denmark Get ready for a day of groundbreaking insights and inspiring discussions. We can't wait to see you there! ??? Check out www.bifrostsummit.com

?? Breaking: Lasso Uncovers Sensitive Private GitHub Repositories from Fortune 500 Companies Found Exposed in Microsoft Copilot via Bing Cache. In a groundbreaking discovery, Lasso’s security researcher Bar Lanyado has uncovered a major vulnerability in Microsoft Copilot. The vulnerability has exposed entire confidential archives that contain intellectual property, sensitive corporate data, access keys, and tokens from major enterprises such as IBM Google Tencent, and PayPal as well as Microsoft itself. Key Findings: ?? Over 20,000 extracted GitHub repositories ??More than 100 internal Python and Node.js packages are vulnerable to dependency confusion ?? Over 300 exposed private tokens, keys, and secrets linked to GitHub Hugging Face, Google Cloud, OpenAI, and other platforms. ?? This exposure has affected over 16,000 organizations! ?? Ophir Dror CPO & Co-Founder at Lasso warns: “Modern organizations must now operate under the assumption that any data leaving their network, even if public only momentarily, can be ingested by LLM engines and search engines, making it permanently accessible" Read the full ?TechCrunch report by Carly Page here: https://lnkd.in/d64hWUas #GenAIsecurity #Cybersecurity #GitHub #Copilot #DataLeak #LassoSecurity #Microsoft #MicrosoftCoPilot #LLMsecurity #GenAI

We're #hiring a new Executive in Residence in San Francisco Bay Area. Apply today or share this post with your network.

DeepSeek R1 is now available, locally hosted in the USA for security and reliability. Try this powerful new reasoning model today with a Ninja Ultra subscription --> https://myninja.ai/