​AI Tech Circle​

Welcome to your weekly AI Newsletter! Read and listen on AITechCircle:

This newsletter has become an essential resource for myself and countless others in the AI community, delivering practical, actionable insights you can apply immediately in your work or business.

Before diving into this week’s updates, do a quick favor and share these valuable insights with a friend or colleague who could benefit from them!

Today at a Glance:

• AI Discussions from GITEX 2024
• Generative AI Use cases in Legal Courts
• AI Weekly news and updates covering newly released LLMs
• Courses and events to attend

Common Questions Discussed at GITEX 2024: Trends, Insights, and What’s Next

This week was all about GITEX, the region's largest technology conference. The event concluded with a packed schedule and a testament to AI's significant impact.

The entire city was buzzing with AI solutions, and every aspect of technology was infused with AI. It was impossible to have a conversation without mentioning AI.

GITEX 2024 was a hub for discussions around AI’s growing influence across industries, from smart cities to advanced healthcare solutions. The focus was on AI-enabled automation, personalized experiences using generative AI, and innovations in AI-driven cybersecurity.

​

The attendees were curious about what is new in AI and how they can adopt or utilize AI in their businesses, processes, or careers.

​

I thought of summarizing my discussion by categorizing.

​

1. General Understanding of AI and Gen AI:
   • How does Generative AI integrate with existing IT ecosystems & business operations?
2. AI in Business Strategy:
   • How can AI provide a competitive edge in [industry]?
   • How can businesses leverage AI for cost reduction and efficiency?
   • What are the steps to build an AI strategy from scratch?
3. AI Implementation Challenges:
   • What are the critical challenges in deploying AI at scale?
   • How can businesses ensure data privacy when using AI models?
   • What are the common pitfalls when integrating LLMs into existing systems?
   • How do you choose between open-source LLMs and proprietary models?
4. Use Cases for AI and LLMs:
   • How can LLMs be used for [Industry]?
   • (this was the most common question, as the attendees were trying to understand how they could use AI)
5. Ethics and Responsible AI:
   • What are the best practices for responsible AI deployment?
   • How can businesses ensure that their AI models are unbiased?
   • How do we build trust with stakeholders when deploying AI systems?
6. Oracle AI Differentiation
   • What makes Oracle’s AI services unique compared to its competitors?
   • How does Oracle integrate AI capabilities directly into its databases and critical features of Oracle DB 23ai?
   • What role does Oracle’s AI play in simplifying enterprise AI deployments?
   • How does Oracle ensure the security and privacy of AI models in the cloud?
7. Oracle Architecture with Nvidia GPUs:
   • What Nvidia GPU options are available on Oracle Cloud?
   • How does Oracle ensure the high availability of Nvidia GPUs for AI workloads?
   • Can Oracle Cloud’s Nvidia GPUs be scaled up for large AI training models?
   • How do Oracle and Nvidia collaborate to optimize GPU performance for AI?
   • How does Oracle Cloud Infrastructure (OCI) handle multi-GPU setups for intensive AI tasks?

​

Please share highlights of your conversations...

Few other takeaways...

​

Trends:

At GITEX 2024, AI’s role in transforming industries was front and center, particularly emphasizing AI-enabled automation.

Discussions highlighted the evolution of AI-driven cybersecurity measures, such as threat detection and response, which are becoming crucial for modern digital infrastructures. The focus on AI’s integration with cloud solutions, like those provided by Oracle, illustrated the growing need for scalable, secure AI capabilities across various business sectors.

​

Insights:

A key takeaway from the event was the rising accessibility of AI tools and platforms, making advanced capabilities available to more organizations. For instance, Oracle emphasized integrating generative AI into cloud services, enabling businesses to harness AI without needing extensive in-house expertise.

Speakers highlighted the importance of understanding AI’s role in data privacy, emphasizing how organizations must balance innovation with responsible AI practices to maintain trust and compliance.

​

What’s Next:

Looking forward, AI’s evolution will continue towards greater integration with edge computing, enabling real-time AI processing closer to data sources.

There is a growing emphasis on sustainable AI practices, exploring how AI can help monitor and reduce carbon footprints across industries.

Weekly News & Updates...

Last week's AI breakthroughs marked another leap forward in the tech revolution.

1. Meta's new models, datasets, and code to support Advanced Machine Intelligence (AMI) link:
   • Meta Spirit LM: An open-source language model for seamless speech and text integration
   • Meta Segment Anything Model 2.1: An updated checkpoint with improved results on visually similar objects, small objects, and occlusion handling. Plus, a new developer suite to make it easier for developers to build with SAM 2
   • Layer Skip: Inference code and fine-tuned checkpoints demonstrating a new method for enhancing LLM performance.
   • SALSA: New code enables researchers to benchmark AI-based attacks to validate security for post-quantum cryptography
   • Meta Lingua: A lightweight and self-contained codebase designed to train language models at scale
   • Meta Open Materials: New open source models and the largest dataset to accelerate AI-driven discovery of new inorganic materials
   • MEXMA: A new research paper and code for a novel pre-trained cross-lingual sentence encoder covering 80 languages
   • Self-Taught Evaluator: a new method for generating synthetic preference data to train reward models without relying on human annotations
2. Swarm: An educational framework exploring ergonomic, lightweight multi-agent orchestration. link​
3. Perplexity for Internal Search: one tool to search over both the web and your team's files with multi-step reasoning and code execution. link​

The Cloud: the backbone of the AI revolution

• AMD introduced the Instinct MI325X, an AI chip with production starting by the end of 2024. AMD Instinct MI325X accelerators deliver industry-leading memory capacity and bandwidth, with 256GB of HBM3E supporting 6.0TB/s, offering 1.8X more capacity and 1.3x more bandwidth than the H200. The AMD Instinct MI325X also offers 1.3X greater peak theoretical FP16 and FP8 compute performance than H200. link​
• GenAI RAG using OCI Digital Assistant and OCI Generative AI Agents. link​
• Advancing Surgical Robotics with AI-driven simulation and Digital Twin Technology: Integrating robotic surgical assistants (RSAs) in operating rooms offers substantial advantages for surgeons and patient outcomes. link​

Gen AI Use Case of the Week:

Using natural language queries, LLMs can transform legal research by providing quick access to case law, statutes, and legal precedents. This capability allows lawyers and judges to find relevant information more efficiently, interpret complex legal texts, and explore case similarities.

Favorite Tip Of The Week:

Here's my favorite resource of the week.

• The policy primer by Cohere explores the “AI language gap challenges,” focusing on the disparities in AI language models across different languages and cultural contexts. It addresses the need for equitable AI systems to understand and serve diverse linguistic communities effectively. The report emphasizes the importance of inclusive training data and outlines strategies for policymakers to bridge these language gaps, ensuring fair access to AI technologies worldwide. For more details, visit Cohere Policy Primer​

Potential of AI

• A recent Anthropic Alignment Science team paper introduces evaluations to assess a model’s ability for subtle sabotage. The study focuses on four areas:
  1. Human decision sabotage: Evaluate if a model can guide humans toward poor decisions discreetly
  2. Code sabotage: Tests whether a model can introduce undetected bugs into software over time
  3. Sandbagging: Assesses if a model can conceal risky capabilities during tests, revealing them later
  4. Undermining oversight: Look at a model’s potential to manipulate evaluation or monitoring systems subtly.

Things to Know...

Adobe has launched the Adobe Content Authenticity web app, a free tool for creators to apply Content Credentials to their digital work. This app aims to protect content from unauthorized use and ensure creators receive proper attribution. It supports a range of digital assets, such as images and videos, providing transparency and trust through secure metadata.

The app integrates with Adobe Creative Cloud tools, allowing creators to control AI training preferences. A public beta will be available in early 2025. link​

The Opportunity...

Podcast:

• This week's Open Tech Talks episode 132 is "Navigating the Rise of Generative AI: What Every CTO Needs to Know with Kevin Surace." Kevin Surace, Chairman & CTO, Appvance AI. Globally recognized futurist, inventor, and innovation leader credited as the pioneer of virtual assistants

Apple | Spotify | Amazon Music ​

Courses to attend:

• 20-hour boot camp on Probability & Statistics from Steve Brunton, University of Washington.

Events:

• ​EUROPEAN Conference on Artificial Intelligence, Oct 19-24, 2024 Santiago de Compostela
• ​TED Conference on AI, October 17-19, 2024 | Vienna, Austria

Tech and Tools...

• ​Phidata: is a framework for building agentic systems
• ​LLM Finetuning toolkit is a config-based CLI tool for launching a series of LLM fine-tuning experiments on your data and gathering their results
• ​Citizen DJ: Make music using the free-to-use audio and video materials from the Library of Congress​

Data Sets...

• Human somatic label-free bright-field cell images. This cell images dataset is collected using an ultrafast imaging system known as asymmetric-detection time-stretch optical microscopy. Link​
• EmoReIQ (Emotion Recognition for Iraqi Autism Individuals) dataset is a specialized EEG dataset designed to capture emotional responses in individuals with Autism Spectrum Disorder (ASD) and Typically Developed (TD) link​

Other Technology News

Want to stay updated on the latest information in the field of Information Technology? Here's what you should know:

• AI deepfakes a top concern for election officials with voting underway, as reported by ABCNews​
• Dan Ives Expects $1 Trillion in Artificial Intelligence (AI) Infrastructure Spending in the Next 3 Years. Here's My Top Pick to Benefit, published by Yahoo Finance​

Join a mini email course on Generative AI ...

​Introduction to Generative AI for Newbies​

Earlier week's Post:

And that’s a wrap!

Thank you, as always, for taking the time to read.

I’d love to hear your thoughts. Hit reply and let me know what you find most valuable this week! Your feedback means a lot.

Until next week,

Kashif Manzoor