AIR#37 - April 10, 2024

Good morning, AI enthusiasts! As you sip on your morning coffee and gear up for the day ahead, today's edition of AIR: The AI Recon brings you a fascinating blend of breakthroughs and innovations that are set to redefine the boundaries of artificial intelligence. From the unveiling of Intel's Gaudi 3 AI Accelerator, promising a quantum leap in generative AI performance, to OpenAI's groundbreaking GPT-4 Turbo now boasting vision capabilities, the AI landscape is buzzing with developments that promise to enhance how we interact with technology on a fundamental level.

In the realm of academia and research, we're seeing an intriguing cross-disciplinary approach with a new study proposing an AI Mentor for social skills training, bridging the gap between computer science and psychology for more accessible learning. This initiative not only highlights the potential for AI to transcend traditional boundaries but also underscores its role in fostering human-centric advancements. Meanwhile, the tech giant Google is not far behind, with its latest AI Chip launch rivaling Nvidia and an Arm-based CPU set to challenge industry titans like Microsoft and Amazon, marking a new phase in the tech giant's hardware development journey.

As we delve into these stories and more, today's edition is a testament to the incredible pace at which AI continues to evolve, touching every aspect of our lives from education to hardware innovation. Whether you're a developer eager to explore new tools or a business professional intrigued by the latest AI applications, there's something in here to spark your imagination and perhaps even inspire your next big idea. So, grab that coffee, settle in, and let's embark on another journey through the fascinating world of artificial intelligence, where innovation meets practicality, one breakthrough at a time.

Business

🔥 Intel Unveils Gaudi 3 AI Accelerator: A Leap in Generative AI Performance and Flexibility
Intel's Gaudi 3 AI accelerator, unveiled April 9, 2024, triples down on generative AI, offering 4x compute, open software, and scalable systems.

Google Unveils AI Chip to Rival Nvidia and Arm-Based CPU Challenging Microsoft, Amazon
Google launches AI chip rivaling Nvidia and an Arm-based CPU to compete with Microsoft and Amazon, marking a new phase in the tech giant's hardware development.

OpenAI Launches GPT-4 Turbo with Lower Pricing and Enhanced Capabilities
OpenAI unveils GPT-4 Turbo: more power, fresher knowledge at a lower cost. Pay-as-you-go with flexible model options for developers.

AI's Massive Energy Footprint: A Looming Environmental Crisis
AI's energy use rivals small nations, doubling by 2026. Urgent call for greener AI and digital sobriety to combat climate crisis.

Ziggy: Revolutionizing Feedback with AI-Powered Conversations
Ziggy revolutionizes feedback with AI, enabling meaningful 1-on-1 conversations for instant insights and faster solutions.

Navigating the AI-SaaS Startup Seas: A Comprehensive How-To Guide
Embark on an AI-SaaS startup journey: From laying foundations, unearthing user pain points, to scaling and adapting. A guide to navigate and thrive in the tech startup seas. 🚀🌊

Canyon: Revolutionizing Job Search with AI-Powered Tools and One-Click Applications
Canyon revolutionizes job hunting with AI: one-click applications, tailored resumes, and interview prep on a single platform.

BuzzGeniusAI: Revolutionizing Marketing with AI-Driven Social Media Engagement
BuzzGeniusAI: The AI that boosts your brand on social media 24/7, simulating word-of-mouth growth for skyrocketing conversions.

Intel Launches Gaudi 3 AI Accelerator to Challenge NVIDIA in AI Market
Intel's Gaudi 3 AI Accelerator gears up to challenge NVIDIA with promising benchmarks, aiming to slice a bigger share of the AI market. Launches Q3 2024.

Engineering

[GitHub] AutoCodeRover: NUS APR's Tool for Autonomous Program Improvement
AutoCodeRover by NUS APR fixes GitHub issues autonomously, improving AI software engineering with a 22% issue resolution rate.

Ollama Blog Introduces Embedding Models for Advanced RAG Applications
Ollama Blog unveils embedding models for RAG apps, blending text prompts with data for smarter searches and AI responses.

🔥 OpenAI Launches GPT-4 Turbo with Vision Capabilities
OpenAI's GPT-4 Turbo now includes vision, enhancing AI's ability to understand images alongside text, promising more accurate problem-solving.

OpenAI Trains GPT-4 on Over a Million Hours of YouTube Content
OpenAI transcribes 1M+ hours of YouTube for GPT-4 training, navigating legal gray zones in AI's data hunger.

Google Launches Gemini 1.5 Pro Globally: Audio Understanding, JSON Mode, and More
Google's Gemini 1.5 Pro launches globally, offering breakthroughs in audio understanding, JSON mode, and more, now in 180+ countries.

Meta to Launch Llama 3 Open Source LLM Next Month
Meta gears up to launch Llama 3 LLM next month, aiming to power AI assistants with improved capabilities and catch up in the generative AI race.

[Paper] Ferret-UI: Enhancing Mobile UI Understanding with Multimodal LLMs
Ferret-UI revolutionizes mobile UI understanding with a multimodal LLM, outperforming GPT-4V in UI tasks.

🔥 [Paper] Google Research's ScreenAI: Revolutionizing UI and Infographic Understanding
Google's ScreenAI revolutionizes UI & infographic understanding, setting new benchmarks with its visual language model. #AIRevolution

[Paper] Griffin: A Hybrid RNN with Gated Linear Recurrences and Local Attention Outperforms Mamba and Matches Llama-2 with Fewer Tokens
Griffin AI outperforms Mamba and rivals Llama-2 with 6x fewer tokens, offering faster inference and efficient scaling.

[Chrome Extension] TabCrunch: Analyze and Categorize Your Tabs with LLMs
TabCrunch Chrome extension: Analyze, categorize, and manage your tabs securely. Prioritizes privacy.

Symbolica Bets on Symbolic AI to Revolutionize Efficiency and Reduce Costs
Symbolica AI, led by ex-Tesla engineer, bets on symbolic models to cut AI costs and boost efficiency, challenging the compute-heavy status quo with a $33M boost.

Academic

[Paper] Leveraging Large Language Models for Social Skill Training: A Cross-Disciplinary Approach
New study proposes AI Mentor for social skills training, bridging computer science and psychology for accessible learning.

[Paper] Ferret-UI: Enhancing Mobile UI Understanding with Multimodal LLMs
Ferret-UI revolutionizes mobile UI understanding, outperforming GPT-4V in tasks by magnifying details for better interaction.

None
None

[Paper] AI Consciousness is Inevitable: Insights from Lenore and Manuel Blum
Lenore & Manuel Blum argue in new study that AI consciousness, inspired by Turing and Baars, is an inevitable outcome of theoretical comp science.

Gary Marcus: Why Scaling Alone Won't Achieve AGI
Gary Marcus argues scaling isn't enough for true AGI, highlighting limitations in data generalization and reasoning. Calls for new approaches.