Mark Neufurth’s Post

Governing through the cloud: the role of compute providers in AI regulation 26 April 2024, 2:00pm - 3:15pm Oxford Martin School & Online As jurisdictions worldwide start regulating the most powerful AI systems, there is a growing need for effective enforcement mechanisms that can both verify compliance and respond to violations. What role can compute providers such as Google Cloud play in supporting AI governance frameworks? Is it possible to establish a harmonised international compute oversight scheme among governments with competing political interests? What lessons can be learnt from existing international cooperation frameworks in other industries to inform global AI regulation? A presentation of key authors of the Oxford Martin School recent paper: Governing Through the Cloud: The Intermediary Role of Compute Providers in AI Regulation, delving deeper into these questions in a panel discussion with Noa Zilberman and other leading scholars from computer science, international relations, and beyond. Registration is not required Livestream: https://lnkd.in/egU4AwKQ

In line with its #AI strategy to make core computing resources more accessible to customers of all sizes - and empower their long-term growth in the AI era - Alibaba Cloud has announced a new pricing strategy for international customers. You can read more about how this benefits organisations by reading a news story published by The Register: https://lnkd.in/eK3243s9 #ai #theregister #alibabacloud

Exclusive: Chinese entities turn to Amazon cloud and its rivals to access high-end US chips, AI - https://lnkd.in/egu_p3t9 A very interesting read. State linked Chinese entities are using different strategies to access generative AI models & secure advanced computing power in the past year.

Sharing this article from SiliconANGLE & theCUBE This is an example of the type of stories/articles that we hand-curate at Teckedin.com. _________________________ "Bessemer predicts that the battle for AI foundational model dominance will intensify in the coming years, since it represents a critical “land grab” that will determine which of the big tech companies reigns supreme within the wider cloud computing market. It believes that, ultimately, closed-source LLMs such as OpenAI’s GPT models will drive the bulk of LLM compute cycles, and that their developers will ultimately be the biggest beneficiaries of this trend."

The race to dominate the AI and cloud landscape is heating up 🚀 Major players like Microsoft, Google, and Meta are significantly increasing their investments in AI infrastructure, reflecting their confidence in these technologies' transformative potential. Read more here 👉🏽 https://lnkd.in/eEA6zWfv (source) Let us know your thoughts in the comments ⬇ #Cloud #AI

🌟Excited to share part 2 of Google Cloud's series on RAGs powered by Google Search technology! Learn how these cutting-edge technologies can enhance the effectiveness of language models and improve information retrieval capabilities. Dive deeper into the critical technologies essential for building a successful RAG system: https://ift.tt/nJLjqa2 🔍 #RAGs #GoogleSearch #AI #MachineLearning #GCP #GoogleCloud #66degrees #cloud #cloudcomputing

As long as the cake of AI market is growing and the cake distribution mechanism is being optimized (Q1), every actor will get a fair piece of cake thanks to its respective effort (Q2, Q3). Q1: "The moves diverge from how AI is being commercialised in the west and capitalises on concerns at Chinese companies about protecting their data." Q2: "During China’s first rollout of cloud computing, Huawei established itself as the dominant provider for private clouds, especially for deep-pocketed government and state-owned groups. Alibaba and Tencent were left vying to provide computing power mostly to other internet groups." Q3: "Procurement records show China’s large state-owned groups have a preference for private cloud AI."

Exciting to see this tangible L'Oréal case study enabled by LangChain and Google Cloud launching “Gen AI as a Service in 3 months." The roadmap including AI agents, graph databases and LangServe most definitely captured my attention too. “Our development roadmap includes building out AI agents, graph databases, and long-term memory as well." "We are also looking to deploy LangServe on Cloud Run as soon as possible." "With LangServe, we’ll be able to add a LangChain application to FastAPI with just one line of code." "This will be indispensable, as we have more and more users and more and more business cases." "We want to be able to deliver new features as quickly as possible without compromising security and accuracy.” s / o Thomas Menard + Harrison Chase

Big news in Gen AI world this week, with Meta releasing version 3.1 of Llama in 3 sizes, which included the all new big 405 Billion parameter model. The reason this is such a big deal is that the 405B model goes head-to-head with proprietary cloud only frontier models like OpenAI’s GPT 4o. Where as Llama has an open source licensing model and has very flexible deployment options. It can be deployed behind a company’s firewall, in their private cloud or hosted with many others, including the hyperscale cloud vendors, Microsoft, AWS and Google. It will require some significant hardware resources, so this will need to be taken into consideration when selecting which of the above deployment options is best for your needs. Also, it is important to consider whether a large general-purpose model is required for your use case, could a number of smaller specialised models be a better and more cost-effective solution. How do you think companies will want to deploy LLMs? For more detail on using smaller models, check out my post about Hierarchical Model Architecture here - https://lnkd.in/gNfNDjn3 #GenAI #Llama #LLM