We’re launching a mini simulation for Australia in 2022 so if you’re interested, join our waitlist at www.moonshotai.org!! If you want to join us to make the impossible possible, there’s also a jobs link at the bottom of the site
NVIDIA’s Earth-2 supercomputer wants to predict climate change accurately by 2060. Armed with our 50-95% faster algorithms, we’re predicting not only climate change, but everything, quickly, efficiently and it’ll be open to the public!
Hey everyone again!!
Sadly my bro and I did not make it to YC this time!
YC said we’ve already accomplished so much and have a bright future ahead, and they loved our passion and energy and understood we were smart - except they said we need to first take a detour to sell our data / algos then reapply after 6-12 months to feed our final mission!
I want to sincerely thank everyone here for supporting us - it means tonnes! I’m also grateful to YC since they woke me up - you can always take a detour to get to the final goal - and it might be a shortcut!
We’re going to first start repackaging our stuff first, sell stuff quickly as a detour!! Our main mission to create a world simulation hasn’t changed though!
Thanks everyone again here for your continued support!!!
OOOOOOOOOOOOOOOO thanks a lot @boris!!! I’m definitely gonna try it out!!! I can see Davinci is a whopping 12K in dimensions!!! Now that’s what I’m talking about!!! It clearly has much more context than Word2Vec’s 300 dimensions!
The discussion revolves around Daniel Hanchen and his team’s exciting project, Moonshot, which aims to predict the future through simulating Earth. The goal is to answer various complex questions affecting individuals and society at large by simulating various factors such as the impact of rising sea levels on property prices or how local air pollution affects lifespan. The team initially used ConceptNet / Word2Vec for matching questions to a database, but plan to utilize and leverage GPT3 advancements for generating real-time predictions. They also developed Moonshot Hyperlearn that makes machine learning algorithms faster using less memory (Post 1).
Various applications of Moonshot were discussed including insurance, real estate, DNA Analytics, AI hardware, and governmental policy support. Their simulation technology has garnered interest from Australia’s Department of Defense, as well as from companies and organizations like Thales, Antler, and Newchip (Post 1).
User daveshapautomator expressed concerns about the misuse of trademarks like JARVIS but also offered to help by contributing his work on cognitive architectures (Post 2). Daniel Hanchen acknowledged this point and expressed an open mind to collaborate (Post 3).
Several users showed support for Moonshot and provided feedback and ideas, including a comparison to the concept of “Rehoboam” from HBO’s Westworld series (Post 5 & Post 6). Daniel also shared a demo video of Moonshot for his YC W22 application for users to get a better idea of the project (Post 15).
Towards the end of the discussion, Daniel shared a snapshot of the new website for Moonshot under development (Post 43) and expressed gratitude to the community for their support following their invite for a YC interview (Post 59).
Overall, the discussion reflects a vibrant and supportive community exploring together the possibilities and challenges presented by advanced AI technology.
Summarized with AI on Nov 24 2023
AI used: gpt-4-32k
Earlywork #48: The Simulation Arms Race