Topics In Demand
Notification
New

No notification found.

Beyond GPT-4: Navigating the AI Landscape of Tomorrow
Beyond GPT-4: Navigating the AI Landscape of Tomorrow

43

0

As the curtains fall on 2023, it's time to reflect on the tremendous strides and ethical quandaries that defined the year in AI. From the debut of advanced chatbots like Bing Chat and Google Bard, showcasing remarkable natural language abilities, to the creative prowess of generative AI models like DALL-E 3 and MidJourney V6 in image generation, the AI landscape witnessed monumental advances. Yet, the year was not without its concerns. The EU's groundbreaking AI Act aimed to restrain certain uses of the technology, while the Biden Administration issued guidelines for its development. Looking ahead to 2024, the burning questions are – Will there be another "AI winter" as reality tempers hype, or will breakthroughs bring about the technology’s widespread adoption across industries? Also, how will policymakers and the public respond?

Rise and Fall of Startups

Venture capitalists unveil a nuanced perspective on the future of generative AI. Masha Bucher, Founder and General Partner at Day One Ventures, predicts a shakeup, with many generative AI startups, especially the ones built on top of OpenAI, facing risks and funding dry-outs. However, she redirects attention to promising sectors like biotech, genome, climate, and industrial applications, emphasizing AI's potential to not only transform industries but also save lives.

Convergence of Data Modalities for Transforming AI Capabilities

Amit Garg, Managing Partner at Tau Ventures, sees the need among companies to think beyond the chat paradigm of 2023. The focus, he suggests, should shift to multimodal AI across input, training, model creation, and output, marking key areas of innovation. The overarching theme is clear — the realm of generative AI is evolving, and its impact extends beyond the confines of conversation.

Cathy Gao, Partner at Sapphire Ventures, also envisions the convergence of data modalities — text, images, and audio — into multimodal models. This transformation is expected to usher in improved decision-making and user experiences across industries such as manufacturing, e-commerce, and healthcare.

Rak Garg, Principal at Bain Capital Ventures, echoes this sentiment, emphasizing the shift towards multi-modal retrieval and inference in AI products. The goal is to create more expressive software that aligns with users across various modalities, from voice to video to code.

Democratization of AI through Open Source

Vivek Ramaswami, Partner at Madrona, and Sabrina Wu, Investor at Madrona, anticipate a surge in open-source models, with major tech companies contributing significantly. However, as Jimmy Kan, Partner at Anzu Partners, points out, there's a race-to-the-bottom in generative AI pricing among companies serving open-source models with most incurring losses on the existing hardware infrastructure. The emphasis should be on profitability and scalability, which will steer investments towards efficient models and value-added services.

Navigating the GPU Conundrum

The perennial GPU shortage looms large in discussions about the future. Rak Garg, Principal at Bain Capital Ventures, sees two solutions to this problem –new compute options (hardware) or new models/architectures that are more efficient with compute resources. Rak expects to see a lot of funding flowing into novel model architectures, and platforms built around diffusion models and liquid neural nets, as the key to overcoming this hurdle.

Chris Kauffman counters the popular narrative, highlighting the GPU shortage as a nuanced issue. He highlights improper utilization of existing infrastructure as a bigger challenge for starters, which can be resolved by fixing lower-level AI software. Until the companies realize this, he predicts more computation will seem like the only solution to GPU shortage, which will lead to NVIDIAs of the world experiencing backlogs and their rivals gaining some market share.

Jess Leao, Partner at Decibel VC, thinks that eventually, the market will converge to a handful of buyers and suppliers, and GPU providers will scale up to meet the demand.

According to TSMC, the long-term sustainability of AI in production won’t rely on general-purpose GPUs, but on hardware specialized for inference, rather than training. Innovations in NPUs will make them a cost-effective and sustainable solution capable of being deployed both on local AI PCs and data centers.

Apple and Google: The Giants Awaken

The narratives around Apple and Google, as envisaged by Jess Leao, Vivek Ramaswami and Sabrina Wu, align seamlessly with the broader conversations about the AI landscape.

Vivek Ramaswami and Sabrina Wu foresee significant releases from Apple in 2024, possibly introducing their own large language model. This could reshape the AI landscape and influence regulatory approaches, given Apple's substantial role in consumer device manufacturing.

Jess Leao, Partner at Decibel VC, shifts the spotlight to Google, anticipating that Google's investment in Gemini and its vast resources could position it as a major player, challenging the dominance of OpenAI and Microsoft.

Preparing for the Long-Term AI Shift

Chris Kauffman's insights about the shift towards chiplets and new architectural paradigms find resonance in the broader narrative of preparing for the long-term AI shift. As AI continues to advance, the industry grapples with the limits of Moore's Law and explores new architectural paradigms.

Jimmy Kan from Anzu Partners highlights the trend towards “edge-to-cloud” or "hybrid AI," integrating both cloud and edge devices. This approach, bridging the gap between centralized cloud processing and on-device capabilities, aligns with the overarching theme of scaling AI to meet global enterprise and consumer needs in the long term.

As we step into 2024, the AI landscape stands at a crossroads — a dynamic interplay of advancements, challenges, and the ever-present quest for ethical and responsible AI development. As the journey unfolds, the world will watch how AI shapes businesses and lives going forward.


That the contents of third-party articles/blogs published here on the website, and the interpretation of all information in the article/blogs such as data, maps, numbers, opinions etc. displayed in the article/blogs and views or the opinions expressed within the content are solely of the author's; and do not reflect the opinions and beliefs of NASSCOM or its affiliates in any manner. NASSCOM does not take any liability w.r.t. content in any manner and will not be liable in any manner whatsoever for any kind of liability arising out of any act, error or omission. The contents of third-party article/blogs published, are provided solely as convenience; and the presence of these articles/blogs should not, under any circumstances, be considered as an endorsement of the contents by NASSCOM in any manner; and if you chose to access these articles/blogs , you do so at your own risk.


images
Dhiraj Sharma
Principal Analyst

© Copyright nasscom. All Rights Reserved.