As we venture into 2024, the field of Artificial Intelligence (AI) will continue to evolve at an unprecedented pace. Our AI experts Mark Chen, Gordon Hart, and Skip Everling offer a comprehensive outlook on what we might expect in the world of AI this year. Their insights provide a blend of technological advancements, environmental concerns, legal challenges, and shifts in public perception.
Mark Chen, Machine Learning Engineer:
Accelerated Growth and Environmental Impacts
Chen emphasizes that machine learning (ML) shows no signs of slowing down. We are witnessing developments that are not only bigger, better, and faster but also carry significant environmental impacts. The increasing energy demands of larger models and the carbon footprint of AI are becoming critical considerations and will be a key topic in 2024.
Robotics and GenAI in Animation
Remarkable advancements in real-world robotics will continue in 2024, approaching the complexity of human sensory experiences and widespread automation across various industries. Additionally, the emergence of Generative AI (GenAI) for animation signifies a leap towards applications once deemed too futuristic.
Security and Quantum Computing
A unique angle from Chen concerns growing awareness around security issues, particularly in watermarking GenAI agents. The interplay between quantum computing and ML could also introduce new dynamics in AI security.
Gordon Hart, Chief Product Officer:
The Rise of Multi-Modality
Expect more multi-modality. Although these have been around for a while, 2024 is the year the "large language model" is out and the "large multimodal model" is in. Models capable of integrating various data streams — processing images, texts, and audio in more unified and sophisticated ways.
Compute Scarcity and Customization
Despite advancements, Hart points out that compute scarcity remains a bottleneck, with GPU supply struggling to meet the demands of ever-growing models. He also anticipates a move towards more personalized AI models, breaking away from the generic output commonly associated with large language models (LLMs).
Changing Public Perception
A potential decrease in broad AI hype, as the novelty of groundbreaking models like GPT gives way to a new normalcy, with people accepting and integrating AI into their daily lives.
Skip Everling, Developer Relations:
AI in Everyday Applications
Everling predicts that AI will become more embedded in everyday apps and experiences, as companies that have been integrating gen AI models in 2023 will start having release cycles in 2024.
AI on the Edge
A significant move towards running AI models on edge devices, like smartphones, marks a departure from traditional cloud-based processing. This shift could democratize AI access and reduce latency in AI applications.
Open-Source AI Advancements
Lastly, Everling foresees open-source AI models catching up to their closed-source counterparts in performance. This trend could lead to a more diverse and accessible AI landscape.
Conclusion
The predictions for 2024 suggest a year of significant advancements, challenges, and transformations in the AI sector. From environmental considerations and legal precedents to technological breakthroughs and shifting public attitudes, the landscape of AI is set for another exciting year of development and discovery.
.svg){kind=small}
