Meet OpenELM: Apple's Game-Changing Move in AI

Welcome to "The AI Pro Max" Daily Digest

LATEST FROM AI

Apple's Big Leap: OpenELM for On-Device AI

Apple has introduced OpenELM, a series of eight small yet effective language models designed for on-device usage, marking a significant step towards enhancing AI capabilities on iPhones. These models, which are significantly smaller than models like GPT-4, are aimed at empowering the open research community and could potentially lead to a more capable Siri in iOS 18. This development has sparked excitement among users and experts, highlighting Apple's commitment to bringing AI directly to devices without relying on cloud services.

AI Safety Board: Tech Leaders Unite

The Department of Homeland Security has recently established the Artificial Intelligence Safety and Security Board, aimed at ensuring the safe and responsible deployment of AI within the nation's critical infrastructure. The board includes prominent tech leaders such as Sam Altman of OpenAI, Satya Nadella of Microsoft, Sundar Pichai of Google, and over 20 other executives, academics, and policymakers. This initiative is a response to the rapid advancement of AI technologies and the potential risks they pose to critical infrastructure, including power grids and airports. The board's mission is to advise on the deployment of AI technologies in a manner that prioritizes safety and security. However, the selection of board members has sparked some controversy, with concerns raised about the inclusion of individuals with vested interests in AI's rapid, unregulated adoption.

ChatGPT's Memory Upgrade: Personalized AI or Privacy Invasion?

OpenAI recently introduced a memory feature for ChatGPT, allowing the AI to remember past conversations and provide more personalized responses. This update has sparked a range of reactions from users, with some expressing excitement about the potential for more relevant and customized interactions, while others are concerned about privacy implications and the AI's ability to form preferences based on previous chats. The feature is seen as both a marketing move and a step towards enhancing the user experience, with some users already testing its capabilities and sharing their experiences.

Meta's Llama 3 Outpaces OpenAI's Models

In a significant leap for open-source AI, Meta has introduced Llama 3, an advanced large language model that outperforms alternatives like vLLM by a notable 20% in training efficiency for its 8B and 70B models. This development has sparked excitement and discussion among AI enthusiasts, with some comparing Llama 3 favorably to OpenAI's ChatGPT due to its open-source nature. The rapid progress in open-source AI, exemplified by Llama 3, could potentially challenge the business models of proprietary AI companies like OpenAI and Google.

AI TOOLS TO CHECK OUT

Notta Showcase - An AI-powered video translator for content localization.

Magnific- The most advanced AI image upscaler & enhancer.

Julius AI- Analyze any form of structured data. Ask questions and have Julius AI perform the analysis for you.

Trickle - Trickle transforms your screenshots using GPT-4 Vision.

Rows - Access the power of AI to analyze, summarize and transform data. Build better spreadsheets, faster.

JOB HUNTING USING AI

YOU SHOULD KNOW

If you have an issue that a bot can't solve, you'll either waste a lot of time or simply give up on the issue entirely, which is often their goal.

A lot of companies nowadays have been using AI and chatbots more and more to do the jobs of human workers. One of those places is as customer service representatives. Amazon does it, Bank of America does it, I suspect that CitiBank does it, and I'm sure many others do it as well.

They will initially have you speak to a helper bot that acknowledges itself as such. The bot will eventually claim to be connecting you to a human agent, but it often isn't. Instead, it's connecting you to a slightly more advanced version of itself, usually a ChatGPT derivative with little, if any, actual power to help you.

This bot will have a human name and will claim to be human, but will often be frustratingly obtuse and unable to comprehend issues.

This is why I recommend, now more than ever, calling companies directly. That way, you can be guaranteed to get a human agent who can understand and has the power to do their jobs.

THATS A WRAP