This time, we don’t ask. We state. DePIN is indeed an opportunity for Web3 to throw down the gauntlet to such players as Google or Amazon ⚔️ To back up this thesis, together with the Onchain Foundation Research Team, we: - Created an overview of business models used by DePIN projects. - Evaluated their efficiency and indicated the best practices. - Designed the recommended business models entrepreneurs can use to build proper Web3 ventures. Then, can Akash Network fight with AWS and Azure? Will NATIX Network compete with Google Maps? And may Livepeer become the first choice for everyone who wants to leverage the generative AI video? It’s all possible, but… ➡️ … you need to check the entire piece to understand how DePIN projects can cross the mythical chasm and attract the masses: https://lnkd.in/epnHX6Vc Many thanks to the entire Research team: Dr. Ananya Shrivastava, Arin Soleymani, Ambreen Khral, Boris Agatić, and the Content and Design masterminds: Ruth M. Trucks, Lucas De Melo, Chris Braithwaite, and Sven Kamieth, for contributing to the piece. Also, a huge shoutout to our partners, Alireza Ghods, Ph.D. from NATIX Network, Greg Osuri from Akash, Alex Rawitz from DIMO, and the entire DFINITY Team, who provided a ton of valuable insights. It’s been a pleasure to collaborate with all of you! 🤝
Michał Moneta, PhD’s Post
More Relevant Posts
-
So a week after the US tech giants announce the Stargate consortium and their $500 billion investment plans, the world has just noticed Deepseek, an AI platform from China. Another AI platform….so what. For me, the thing that makes this interesting is the fact it’s open-source. Its been on my list to have a look at for the last week. Deepseek has been developed at a much lower cost as the code is readily available and can be shared and improved. Also it’s performance per price point is much better than OpenAI and that’s worried investors. A lot. Nividia (who make the machines these platforms run on) had $500 billion wiped off its valuation today and the ripples have been felt by most of the tech companies today. I’ve been using open-source software since the 90’s and have never seen as big an impact on the market leaders so suddenly. Still think AI is something just for the nerds to play about with? I’m curious…are you using AI as part of your daily business toolset? 👇 🦍 👇 I’m Craig from GorillaHub, and we build and manage Hardworking Business Websites to let business owners do what they do best.
To view or add a comment, sign in
-
-
👀 TOP-10 AI Startups to Watch: Bloomberg edition and my opinion 1. Suno - AI Music Generation 🎻 Suno is incredibly helpful for young musicians in their creative process, generating soundtracks for short films, or just for fun. 2. Perplexity - AI Search Engine 💻 While Perplexity will need significant time to gain trust and become popular among the general public, it is growing rapidly and remains interesting to watch. 3. xAI - Elon Musk’s Grok Chatbot 📱 You either appreciate Elon Musk and his unique sense of humor or you don’t. 4. Scale AI - Data Labeling Service for AI Products 🧹 Organized and cleaned up vast amounts of data... I’m loving it! 5. Cohere - Enterprise Models Including Command R+ 🦾 Salesforce Inc. and Accenture have become their clients, and Nvidia Corp. and Oracle Corp. have invested in the company. Cohere is a star in developing commercial decisions for business-focused tasks! 6. CoreWeave - Compute Infrastructure Services ☁️ This is the story of some guys who were focused on Bitcoin mining. Mining bitcoins has become unfashionable, so now CoreWeave provides cloud computing horsepower, offering chips for hire to companies like Microsoft, which are desperate for capacity. 7. ElevenLabs - AI Voice-Generating Software 🎶 While ElevenLabs has room to grow, it faces significant ethical challenges, as many may attempt to misuse this technology. 🛠Finally! This shortlist also includes OpenAI, Anthropic, and Mistral, but all of these are obvious winners and need no introduction from me. Which one is your favorite? Tell me in the comments 👇 #topaistartups #aistartup #ainews #CoreWeave #Cohere #scaleai #airating #startupreview #aientrepreneurs
To view or add a comment, sign in
-
-
LLaMa 3 is a key part of Meta's AI strategy and represents tremendous progress in just a few months time. Who do you think will win the AI foundation model arms race? To me, OpenAI is trying to be 'foundation-for-all' akin to the IBM of old. Anthropic is making waves but with an emphasis toward control and safety. Google is pushing Gemini as a means of competing with Azure and AWS, and Meta is pushing open source AI as a way to be like the Android of the market. Feels like we are going to have a pyramid of foundation models with multiple large multi-purpose models from multiple large players at the bottom and more specialized models from various providers moving up as the pyramid narrows.
NVIDIA Senior Research Manager & Lead of Embodied AI (GEAR Lab). Stanford Ph.D. Building Humanoid Robots and Physical AI. OpenAI's first intern. Sharing insights on the bleeding edge of AI.
Llama 3 is here!! The upcoming 400B+ tier will mark the watershed moment that the community gains open-weight access to a GPT-4-class model. It will change the calculus for many research efforts and grassroot startups. I pulled the numbers on Claude 3 Opus, GPT-4-2024-04-09, and Gemini. Llama-3-400B is still training and will hopefully get even better in the next few months. There is so much research potential that can be unlocked with such a powerful backbone. Expecting a surge in builder energy across the ecosystem! Meta is so back: https://lnkd.in/gY_Vn8Gz
To view or add a comment, sign in
-
-
Feels great to come out of stealth mode! https://lnkd.in/ghvXjyDC. The AI revolution was kick-started by language models that can reason in human-like ways. But for this revolution to see its full impact, computers need to see in human-like ways too. Vision language models are showing they can solve Vision AI much like LLMs did to text. Their foundational nature means they can see, read, count, and interpret vision in a generalized way. At Moondream, our mission is to bring the power of VLMs to developers everywhere. Our customers are chomping at the bit to build new vision features into their products. From transportation, retail, manufacturing, security and more, they see Vision AI as a critical factor to their success. Our customers want powerful VLMs that solve their needs, and they want the ability to run that everywhere: cloud, on-prem, desktop, mobile and edge. Moondream is excited to meet their needs by building state-of-the-art Vision AI capabilities, combined with an easy distillation process that reduces the models to tiny sizes that can run quickly everywhere.
To view or add a comment, sign in
-
🌟 Introducing Perplexica: Your Open-Source AI-Powered Search Tool! Perplexica is an innovative open-source project that replicates the functionality of the $2.5B startup Perplexity, enabling developers to easily launch AI-powered search tools. What Perplexica Offers: 🔍 Intelligent Search: Provides accurate and clear answers by thoroughly scanning the internet and understanding your queries, complete with cited sources. ⚡ Efficient Performance: Supports local models like Llama3 and Mixtral, ensuring faster and more affordable inference. 🎯 Tailored Search Modes: Six specialized modes to suit different needs: 1. All Mode: Searches the entire web for comprehensive results. 2. Writing Assistant Mode: Helps with writing tasks without needing web searches. 3. Academic Search Mode: Perfect for finding scholarly articles and research papers. 4. YouTube Search Mode: Locates YouTube videos based on your queries. 5. Wolfram Alpha Search Mode: Utilizes Wolfram Alpha for precise calculations and data analysis. 6. Reddit Search Mode: Scours Reddit for discussions and opinions. Perplexica stands out by using SearxNG, a metasearch engine, to deliver the most up-to-date information. Explore the GitHub repository: https://lnkd.in/gdHPJFqV #OpenSource #AI #Perplexica #SearchTools #AIInnovation #Developers #Tech #llm #machinelearning #llama3
To view or add a comment, sign in
-
-
We are ready to run it back again and start taking applications for Cohort #2 of CryptoOracle's AI Web3 Accelerator. If you are a founder building a business at the intersection of AI and Web3 and looking to take your startup to the next level apply or reach out to me if you have any questions. #ai #web3 #startups #founder #accelerator #cryptooracle
Excited to announce the opening of applications for Cohort #2 of CryptoOracle's AI Web3 Accelerator. But before we move on to Cohort #2, I want to reflect on lessons learned from Cohort #1 (see post below). And I'd like to thank all those who helped organize and guest lecture at the Accelerator including Eric Bravick Peter Chmiel Charles Kanavel Shaun Geer Kyle Heise Steve Masur Michael Olorunninwo Allan Dumaine Dmitri Pchelintsev Player1Taco .eth Pete Townsend Lucas Tcheyan Mohamed Baioumy Matt Wright Andrew Ackerman Adam Miller Albert Liang Shayon Sengupta Pedro Herrera Michael Adler Avery Ching David Schwed Adam Benayoun Ivan Fartunov Keli Callaghan Azhar Hussain Andy Singleton John Sarson Scott Berenzweig Alexander Golding Sabrina B. Mike Anderson Jarrod Barnes Anna Bertha Aly Madhavji Drew Austin Mike Chan Dele Atanda Eric Forst Tyler Carter https://bit.ly/5LChrt1
5 Lessons Learned From Cohort #1 Of The Collective’s AI Web3 Accelerator
medium.com
To view or add a comment, sign in
-
Last Thursday I attended #MITforge2024! Thanks John Werner for bringing the entire Boston AI ecosystem together - from legends & visionaries, first time founders & serial entrepreneurs, academic leaders & students, pre-seed & private equity investors, big tech & startups looking to launch. It was the high energy tech gathering that I attended in a while. Especially impressed with Liquid AI. We at Pristine AI became familiar with #LNN last year. Exciting to see their progress over the last many months. With their explainability, learning abilities, speed and low cost training, they could pose a formidable challenge to the better known foundation models. It was great meeting with Igor J., Stephen Wolfram Ramesh Raskar Paul Baier and several fellow founders. Interesting listening to Yann LeCun on the day https://lnkd.in/eNb-iF9X was released. The battle between open source and proprietary models may well go the way the battles among various Unix flavors and Linux went. Among the various takes on the Prediction for 2025, I liked Lex Fridman's 5 trends, especially 'LLM Personalization'. Pristine AI has focused on it for a while. Most importantly, it was great to experience the vibrancy of Boston tech eco system and the sheer number of start ups coming down the pike. #founders #generativeAI #robotics
Meta Llama 3
llama.meta.com
To view or add a comment, sign in
-
Llama 3 is here!! The upcoming 400B+ tier will mark the watershed moment that the community gains open-weight access to a GPT-4-class model. It will change the calculus for many research efforts and grassroot startups. I pulled the numbers on Claude 3 Opus, GPT-4-2024-04-09, and Gemini. Llama-3-400B is still training and will hopefully get even better in the next few months. There is so much research potential that can be unlocked with such a powerful backbone. Expecting a surge in builder energy across the ecosystem! Meta is so back: https://lnkd.in/gY_Vn8Gz
To view or add a comment, sign in
-
-
#SamAltman, CEO of #OpenAI, recently shared his insights on the future of AI and startups. He believes that #GPT5 should be the underlying infrastructure for startups, rather than trying to improve ChatGPT with minor patches. Altman warns that attempting to enhance ChatGPT with small products could result in sudden startup death with the release of #GPT5. This perspective is intriguing, especially considering #Zuckerberg's latest move with Llama 3 which btw is free. #Llama 3 is proving to be nearly as powerful as #GPT-4-turbo! So startups need not be scared of OpenAI's power.
NVIDIA Senior Research Manager & Lead of Embodied AI (GEAR Lab). Stanford Ph.D. Building Humanoid Robots and Physical AI. OpenAI's first intern. Sharing insights on the bleeding edge of AI.
Llama 3 is here!! The upcoming 400B+ tier will mark the watershed moment that the community gains open-weight access to a GPT-4-class model. It will change the calculus for many research efforts and grassroot startups. I pulled the numbers on Claude 3 Opus, GPT-4-2024-04-09, and Gemini. Llama-3-400B is still training and will hopefully get even better in the next few months. There is so much research potential that can be unlocked with such a powerful backbone. Expecting a surge in builder energy across the ecosystem! Meta is so back: https://lnkd.in/gY_Vn8Gz
To view or add a comment, sign in
-