Tom Erickson’s Post

Apple made waves at Amazon Web Services (AWS) re:Invent 2024, with Benoit Dupin, Apple’s senior director of machine learning and AI, taking the stage alongside AWS chief Matt Garman. Dupin shed light on Apple’s decade-long partnership with AWS and how it fuels innovations like Siri, iCloud Music, and Apple TV. By embracing AWS’s Graviton and Inferentia chips, Apple reported a 40% efficiency boost and doubled performance for specific AI tasks. Dupin also hinted at exploring AWS's Trainium2 chips, which promise a 50% leap in pre-training efficiency. This collaboration underscores AWS’s vital role in Apple’s AI and ML lifecycle, powering everything from training to deployment. AI and ML at scale, Apple-style. Read more now 👇

Amazon Web Services (AWS) re:Invent event brought about numerous exciting announcements, but one was particularly exciting which revealed the extensive partnership between AWS and Apple. This collaboration sheds light on how AWS plays a crucial role in powering everyday features like iCloud, Siri, Apple Maps, and Apple Music that we all rely on. Benoit Dupin, Senior Direct of AI and Machine Learning also shared a 40% efficiency boost achieved through the utilization of AWS Inferentia and Graviton chips, and the staggering scale of support for iCloud services was unveiled, with over 2 trillion objects stored in S3 storage. #aws #apple #ai #ml #genai #partnership For more details on this partnership, you can read the article here: [Link to CNBC article] #AWS #Apple #CloudComputing #TechInnovation #iCloud

As we are still in the midst of the action-packed #AWSreInvent, more and more news keeps emerging every day. Among the many announcements and sessions, one surprise reveal might have been missed by many… Apple made surprise cameo at Amazon Web Services (AWS) re:Invent Benoit Dupin, Apple's senior director of machine learning and AI, took the stage during Matt Garman’s keynote on Tuesday! In this unexpected appearance, Apple provided a rare look into how they rely heavily on AWS to power their cutting-edge services and features. Just take a look at the picture below to see SOME of the AWS services and features they are using! During his speech, Dupin shared fascinating insights into Apple's deep usage of AWS over the past decade to support their rapidly innovating and scaling needs. A quick glimpse at the slide showcased the staggering number of AWS services Apple leverages, from training to inference to storage and more. With Apple's laser focus on delivering rich, privacy-protecting user experiences globally, Dupin emphasized how AWS has been instrumental in keeping up with their highly dynamic scale and performance requirements. Some key highlights that caught my attention: 🤖 To build Apple Intelligence features like advanced writing tools and GenMojis, they scaled training of large language models and diffusion models on AWS. 🚀 Apple saw over 40% efficiency gains migrating instances from x86 to AWS Graviton, and 2x better performance for some search tasks using AWS Inferentia chips. 📈 For pre-training models powering Apple Intelligence, they are evaluating new AWS Trainium2 instances and expect up to 50% better efficiency. It was truly eye-opening to witness a technology titan like Apple go into such detail about their reliance on AWS to fuel innovation at a blistering pace. With AWS continuing to push boundaries on scalability, performance, generative AI and cloud services, this partnership seems poised to deliver more amazing experiences.

#Apple’s Senior Director of #Machine Learning and AI, Benoit Dupin, made a surprise appearance as the first guest during #AWS CEO Matt Garman’s keynote at re:Invent 2024! Dupin delivered a powerful endorsement for our cloud capabilities as he highlighted Apple’s extensive use of AWS for key offerings like Siri, #Apple Music, and Apple TV, hitting on AWS's vital role in powering these innovations.💡 Dupin also shared that Apple is evaluating #Trainium 2, our latest AI training chip, expecting up to a 50% improvement in efficiency for model pre-training. The collaboration between Apple and AWS underscores how leading tech companies can work together to leverage cloud capabilities, driving advancements in AI and machine learning. 🤖 Stay tuned for more insights from the biggest event in cloud!

Quickflick - AI News I 🌟 Apple and AWS: A Game-Changer in AI Collaboration! At the recent AWS Reinvent conference, Apple unveiled its strategic use of custom AI chips from Amazon Web Services

Amazon announces new AI guardrails and tools at Amazon Web Services (AWS) Summit, boosting developer productivity and ensuring robust security. Highlights include AWS App Studio for rapid app creation, Amazon Q updates for enhanced developer support, and new capabilities in Amazon Bedrock for generative AI applications. These advancements make AI more accessible to organizations of all sizes, emphasizing responsible AI usage and reducing harmful outputs. #AllAboutAI #AWS #AI #Appcreation #TechInnovation #Amazon #CloudComputing #ArtificialIntelligence

Apple Intelligence Foundation Language Models This paper provides technical details for Apple’s On-Device and Server Foundation Models 👉 https://lnkd.in/drSUCisP , introduced on June 10, 2024, in this post 👉 https://lnkd.in/dnVxeexc Apple presents foundation language models developed to power Apple Intelligence features, including a ∼3 billion parameter model designed to run efficiently on devices and a large server-based language model designed for Private Cloud Compute. These models are designed to perform a wide range of tasks efficiently, accurately, and responsibly. This report describes the model architecture, the data used to train the model, the training process, how the models are optimized for inference, and the evaluation results. Apple highlights focus on Responsible AI and how the principles are applied throughout the model development. #machinelearning

This came as a pleasant surprise to many at AWS #reinvent2024. Apple's Senior Director of Machine Learning and AI, Benoit Dupin showed up in Matt Garman's keynote. Here is what he shared: ➡️ Apple has relied on AWS for over a decade to scale its machine learning and AI capabilities. ➡️ Features like Siri, iCloud Music, and Apple TV depend heavily on AWS infrastructure. ➡️ This is my favourite, Apple has achieved 𝟒𝟎% 𝐞𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐜𝐲 𝐠𝐚𝐢𝐧𝐬 𝐛𝐲 𝐦𝐢𝐠𝐫𝐚𝐭𝐢𝐧𝐠 𝐟𝐫𝐨𝐦 𝐱𝟖𝟔 𝐭𝐨 𝐀𝐖𝐒'𝐬 𝐆𝐫𝐚𝐯𝐢𝐭𝐨𝐧 𝐩𝐫𝐨𝐜𝐞𝐬𝐬𝐨𝐫𝐬. ➡️ Leveraging Inferentia 2 chips for search-related tasks doubled efficiency for specific features. ➡️ Early exploration of AWS Trainium2 chips shows a 𝟓𝟎% 𝐢𝐦𝐩𝐫𝐨𝐯𝐞𝐦𝐞𝐧𝐭 𝐢𝐧 𝐩𝐫𝐞-𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠 𝐞𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐜𝐲. ➡️ AWS's expertise and services are critical for Apple's AI and ML scalability. The partnership with Apple truly exemplifies AWS's pivotal role in supporting leading tech companies' AI efforts. I am impressed. What do you think? #Apple #AWScustomerstory #Graviton #AI #machinelearning

Every year, Amazon Web Services (AWS) re:Invent showcases the incredible pace of innovation, which is a testament to our customers’ drive and ambition. And continues to live up to the promise of drinking from the fire hose. By simultaneously advancing compute infrastructure, hardware innovation, cutting-edge models, and strategic alliances - all while offering competitive pricing - Amazon is creating a comprehensive AI ecosystem. Late to the game? Not really :) My current highlights include: - Amazon Nova, a new generation of foundation models that delivers frontier intelligence and industry leading price performance, available today in Amazon Bedrock. These fast & incredibly capable models have impressive results on a wide range of industry standard benchmarks, and give our customers even more choice for their specific needs; - Bedrock's Model Distillation, which will enable models to run 500% faster and up to 75% cheaper; - the general availability of Trainium2 chips on Anthropic: four times faster than their predecessors and 50% cost reduction compared to NVIDIA; plus major partnerships secured, including with Apple; and the next-generation Trainium3 chips already in the works; - The AI Supercomputer, project Rainier: one of the world's largest distributed AI compute clusters with hundreds of thousands of Trainium chips working in harmony. The target is to get a performance five times greater than current leading systems with multi-location design for unmatched scalability. There much more coming, keep drinking at the fire hose :)

As you can imagine GenAI was a big focus topic at our AWS Partners summit in Tokyo this week. It is amazing to hear about the solutions our partners in Japan are building with Amazon Web Services (AWS) #generativeai but we are only just getting started.