🦁 AI-Enabled PCs and GenAI Smartphones

The Tech Issue | February 8, 2024

Author

Qamar Zia
February 08, 2024

☕️ Greetings, and welcome to my daily dive into the Generative AI landscape.

In today’s issue:

  • AI spurs space sector innovation

  • AI-Enabled PCs and GenAI Smartphones

  • Google changes Bard's name to Gemini: unveils new Android app and advanced AI model

  • D-Wave and Zapata AI announced a multi-year strategic partnership

  • and more

🔔 Please forward this newsletter to your friends and team members and invite them to join. This will help me grow my reach. Thanks, Qamar.

🗞️ AI-Enabled PCs and GenAI Smartphones

Recent research from IDC and Gartner reveals a significant trend in the tech industry: a surge in AI-enabled devices, particularly PCs and smartphones. By 2027, it's anticipated that 60% of PCs shipped will be AI-native, a substantial increase from one in five PCs this year. This shift is driven by the demand for on-device generative AI (genAI) capabilities, moving away from the traditional cloud-based AI data processing due to concerns over privacy, security, and network connectivity issues. The move towards embedding AI directly into devices like PCs and smartphones aims to enhance performance, reduce reliance on cloud resources, and address the scarcity of specialized processors needed for AI tasks.

Highlights:

  • AI PC shipments are expected to rise from 50 million units in 2024 to over 167 million by 2027.

  • Gartner predicts a total of 295 million shipments of AI PCs and genAI smartphones by the end of 2024.

  • The transition to local AI processing addresses issues with cloud-based models, including high energy consumption and network latency.

  • Edge devices with AI capabilities are becoming more common, with manufacturers focusing on improving privacy, security, and cost-efficiency.

  • Future smartphones, like the speculated iPhone 16 with Apple GPT, will feature advanced AI capabilities, including neuro-processing units for tasks like photo optimization.

  • The market for AI-enabled devices is growing, with a 3.5% increase in PC shipments and a 4.2% increase in smartphone shipments expected in 2024.

  • The evolution of AI technologies in devices is not expected to drive immediate demand but will become a standard feature, challenging vendors to differentiate their products.

🗞️ TRENDS

OpenAI is developing two AI agents to streamline complex tasks. One agent can operate user devices to manage data transfers and fill forms, while another focuses on web tasks like data collection and booking tickets. These efforts aim to evolve ChatGPT into a versatile personal assistant, potentially tailoring its functions to individual users and tasks, including email composition and document creation. Integrating various GPT capabilities and the Assistant API mark steps toward creating general-purpose AI assistants, aligning with broader industry trends toward actionable AI support.

🗞️ IMPACT (Economy, Workforce, Culture, Life)

A cultural divide in AI development is becoming legal, affecting open innovation's future. The contention surrounds generative AI models, capable of complex tasks, and whether their "weights" should be public. Critics argue for control to prevent misuse, while proponents, including Stability AI's head of public policy, advocate for open models as crucial for transparency, competition, and grassroots innovation. Recent EU and US regulations introduce thresholds for AI model regulation, potentially stifling innovation by imposing burdens on smaller developers and risking a shift towards more restricted development environments.

🗞️ OPINION (Opinion, Analysis, Reviews, Ideas)

Sam Altman's vision for OpenAI as a provider of "magic intelligence in the sky" aims to democratize access to the accumulated knowledge of humanity, theoretically reducing the cost of intelligence to zero. Ironically, the prevalent use of this advanced AI, including generative AI tools by companies like Google and Meta, often diverges towards trivial or malicious activities such as creating spam, cheating, and generating disinformation, including deepfakes and fake content. Despite the potential for learning, creativity, and productivity enhancement, these negative uses dominate, challenging the integrity of various societal facets, including politics and individual reputations. This misuse reflects not on the technology itself but on the human inclination towards exploiting tools for lesser purposes, thereby, paradoxically, reducing the cost of producing and spreading "bullshit" rather than elevating intelligence.

🗞️ LEARNING (Tools, Frameworks, Skills, Guides, Research)

Creating custom AI chatbots has become significantly easier and quicker with OpenAI's Assistants and Fine-tuning APIs. This article offers a detailed guide, including Python code, for leveraging these tools to develop AI agents, differentiating between chatbots (basic conversation) and assistants (enhanced with functionalities like web browsing). It introduces OpenAI's Assistants API, illustrating its use with a YouTube comment responder example, and explains the process of making an AI more personalized through few-shot prompting, knowledge retrieval, and fine-tuning, using practical examples and code snippets.

🗞️ BUSINESS (Use Cases, Industry spotlight)

In Mountain View, Calif., the space sector's innovation is surging, largely due to artificial intelligence (AI) and machine learning. Amazon Web Services (AWS) customers are at the forefront, using generative AI to design and 3D-print spacecraft parts, leading to rapid and innovative iterations. Generative AI, complemented by machine learning and computer vision, is revolutionizing space companies by improving data utilization and operational automation. Startups like Cognitive Space automate satellite operations with AI tools, enhancing imagery search efficiency. HawkEye 360 benefits from AI in identifying and specifying radio frequency emitters globally. The panelists at the SmallSat Symposium highlighted the essential role of cloud computing and discussed the potential of space-based edge processing, including data centers in orbit, to reduce data latency and support remote operations, signaling a significant shift towards in-space computing and storage solutions.

🗞️ LATEST FROM THE WEB

Google saves your conversations with Gemini for years by default: Google warns against sharing sensitive information in conversations with its GenAI apps, Gemini, highlighting its data collection practices. Human annotators, whose affiliations are not disclosed, review conversations to enhance the service, retaining data for up to three years. Users can limit data retention through settings, but conversations are still stored for 72 hours for safety and improvement purposes. Google's approach mirrors the industry's struggle to balance privacy with GenAI model development, amidst increasing regulatory scrutiny and organizational concerns over privacy risks. Read more at techcrunch.com.

D-Wave and Zapata AI announced a multi-year strategic partnership: D-Wave Quantum Inc. and Zapata Computing, Inc. have announced a strategic partnership to combine generative AI and quantum computing for developing applications aimed at solving complex optimization problems and accelerating discoveries. This collaboration will leverage D-Wave's quantum annealing technology and Zapata's AI software, focusing on quantum generative AI models to discover new molecules. The partnership promises to accelerate commercial application development, with exclusive application availability through D-Wave's Leap quantum cloud service, marking a significant advancement in harnessing the potential of quantum computing and generative AI. Read more at ai-techpark.com.

Favorable Winds for AMD in the GenAI Chip Market: In the rapidly evolving field of Large Language Models (LLMs) and Generative AI, AMD stands out with its promising hardware offerings, characterized by performance and efficiency gains for AI tasks, and a robust software ecosystem. AMD’s GPUs, optimized for AI workloads, along with their open-source ROCm platform, are becoming increasingly attractive for AI development, fostering an ecosystem where entities like Lamini enhance LLM efficiency on AMD chips. With market trends favoring AMD, including shorter lead times and high-profile endorsements, AMD is poised to play a crucial role in advancing AI technology. Furthermore, innovative approaches like Secure Multi-Party Computation (SMPC) are emerging to address data privacy in AI, enabling secure collaboration without compromising sensitive information, thus broadening the application scope for LLMs in regulated industries. Read more at gradientflow.com.

Google changes Bard's name to Gemini: unveils new Android app and advanced AI model: Google's Bard AI chatbot is being rebranded as Gemini, aligning with the tech giant's latest large language model initiative. Alongside this rebranding, Google introduces a dedicated Android app, allowing Gemini to replace Google Assistant as the default on Android devices. A notable upgrade, Gemini Advanced, leverages Ultra 1.0 for enhanced reasoning, coding, and creative tasks, available through a new $20 monthly Google One Premium Plan. This move signals a shift in Google's AI and subscription strategies, aiming to boost its user base and offer comprehensive digital services. Read more at techspot.com.

Google Deepmind proposes ‘self-discover’ framework for LLMs, improves GPT-4 performance: Researchers at Google Deepmind and USC have introduced a 'self-discover' prompting framework to boost reasoning in large language models like GPT-4 and PaLM 2, increasing their performance on benchmarks by up to 32%. This method enables models to autonomously create reasoning structures for problem-solving, significantly reducing the computational resources needed. Tested on various tasks, the approach surpassed existing methods in both efficiency and accuracy, marking a significant step towards enhancing AI's problem-solving capabilities and potentially facilitating human-AI collaboration. Read more at venturebeat.com.

Join my community by subscribing to my newsletter below:

Reply

or to participate.