Discover more from The AI Edge
Apple & Meta's Unlikely AI Alliance
Plus: ByteDance and Broadcom teams up for AI chip development & Researchers developed a new method to detect hallucinations.
Hello Engineering Leaders and AI Enthusiasts!
Welcome to the 303rd edition of The AI Edge newsletter. This edition features Apple and Meta’s potential gen AI partnership.
And a huge shoutout to our amazing readers. We appreciate you😊
In today’s edition:
🤝 Apple and Meta are discussing gen AI partnership
🤖 ByteDance and Broadcom collaborate on AI chip development
🕵️♂️ Researchers developed a new method to detect hallucinations
📚 Knowledge Nugget: 11 steps to keep Meta from stealing your data to train AI by
Let’s go!
Apple and Meta are discussing gen AI partnership
Apple is reportedly in talks with its longtime rival Meta to integrate the latter's Llama 3 AI model into Apple Intelligence. This move comes as Apple prepares to roll out its AI features across iPhones, iPads, and Macs later this year.
The potential partnership follows Apple's existing deal with OpenAI, suggesting a collaboration strategy rather than solo development in the AI race. In Apple's arrangement with OpenAI, there's no direct payment. Instead, OpenAI can offer premium subscriptions through Apple Intelligence, with Apple taking a percentage. It's unclear if Meta would agree to a similar business model, given that Llama 3 is open-source and free to access.
Why does it matter?
Despite their historical rivalry, especially after Apple's 2021 privacy changes impacted Meta's ad revenue, this potential partnership suggests a willingness to cooperate in the AI race. However, not everyone is thrilled about it. Users and investors are voicing concerns about privacy, with Meta's track record being a major sticking point. Some see this as a desperate move by Apple.
ByteDance and Broadcom collaborate on AI chip development
ByteDance is collaborating with U.S. chip designer Broadcom to develop an advanced AI processor. This partnership aims to secure a stable supply of high-end chips amid ongoing U.S.-China tensions. The project centers on creating a 5-nanometre, customized Application-Specific Integrated Chip (ASIC) that complies with U.S. export restrictions.
This chip's manufacturing is set to be outsourced to Taiwan Semiconductor Manufacturing Company (TSMC), though production is not expected to begin this year. While the design work is currently underway, the critical "tape out" phase has yet to commence.
Why does it matter?
This partnership shows how Chinese tech companies are getting creative to secure advanced AI chips. As ByteDance has launched a major push in generative AI, this new initiative brings dual benefits: reducing procurement costs and ensuring a stable supply of higher-end chips.
Researchers developed a new method to detect hallucinations
ChatGPT and Gemini can produce impressive results but often "hallucinate" false or unsubstantiated information. This research focuses on a subset of hallucinations called "confabulations," where LLMs generate answers that are both wrong and arbitrary. Researchers have developed new methods to detect confabulations using entropy-based uncertainty estimators. They introduce the concept of "semantic entropy" to measure the uncertainty of LLM generations at the meaning level.
High semantic entropy corresponds to high uncertainty and indicates a higher likelihood of confabulation. The method computes uncertainty at the level of meaning rather than specific word sequences, addressing the fact that one idea can be expressed in many ways. The method provides scalable oversight by detecting confabulations that people might otherwise find plausible.
Why does it matter?
Detecting confabulations can improve question-answering accuracy for state-of-the-art LLMs. The method helps users understand when to be cautious with LLM outputs and opens up new possibilities for using LLMs that were previously limited by unreliability.
Enjoying the daily updates?
Refer your pals to subscribe to our daily newsletter and get exclusive access to 400+ game-changing AI tools.
When you use the referral link above or the “Share” button on any post, you'll get the credit for any new subscribers. All you need to do is send the link via text or email or share it on social media with friends.
Knowledge Nugget: 11 steps to keep Meta from stealing your data to train AI
In this detailed article,
explains Meta's sneaky AI data play. Starting June 26, 2024, the tech giant plans to use your photos and posts to train their AI unless you explicitly opt out. But here's the kicker: opting out is a digital obstacle course.Copaken breaks down the 11-step process to keep your data off Meta's AI menu, from navigating Facebook's maze of settings to crafting the perfect "hands off my data" message. The process involves multiple menu dives, form-filling, and even an OTP verification - all seemingly designed to discourage users from completing the opt-out.
Why does it matter?
Your online footprint becomes a hot commodity as tech giants get hungrier for data to feed their AI models. This opt-out process shows how companies can make it challenging to protect your digital rights, highlighting the need for stronger data protection laws and user-friendly privacy controls.
What Else Is Happening❗
🆕Stability AI has appointed a new CEO
Prem Akkaraju, former CEO of Weta Digital, will step up to this role. This change comes after founder Emad Mostaque resigned in March amid financial concerns. Akkaraju is part of a new investor group, including former Facebook President Sean Parker, that provides a cash infusion to Stability AI. (Link)
🖼️ Coders spotted Midjourney references in Grok's standalone version
It hints at a possible collaboration between xAI and Midjourney for image generation. Grok might use Midjourney models via API to generate images. This could mean that the standalone Grok might let users upload images. All this could be part of the Grok 1.5 rollout. (Link)
🚫YouTube fights AI fakes with new policy
YouTube expanded its privacy policy to cover AI-generated content. Now, users can request the removal of synthetic media that imitates their face or voice. However, privacy complaints alone may not guarantee content removal. While reviewing requests, YouTube will consider the potential for confusion and intent (parody/satire). (Link)
✨ OmniAI transforms unstructured enterprise data for AI and analytics use
Founded by Y Combinator alums, OmniAI syncs with your existing storage, prep the data, and lets you run your choice of AI models on the data. It offers integrations with popular models like Meta's Llama 3 and Anthropic's Claude. (Link)
🤖 The new Dot AI chatbot wants to be your companion
Dot chatbot aims to personalize interactions based on users' thoughts and feelings. It uses multiple AI models from OpenAI and Google. The app uses intensive Q&A to learn about users' interests and experiences. Dot offers a subscription model at $11.99/month for unlimited conversations. (Link)
New to the newsletter?
The AI Edge keeps engineering leaders & AI enthusiasts like you on the cutting edge of AI. From machine learning to ChatGPT to generative AI and large language models, we break down the latest AI developments and how you can apply them in your work.
Thanks for reading, and see you tomorrow. 😊