Redefining AI R&D with GPT-4o to Embrace a Comprehensive and Pragmatic Approach

The rapid evolution of AI technology challenges traditional R&D approaches. This week’s release of GPT-4o underscores the necessity of not only strategising, developing, and researching, but also effectively integrating these technologies to deliver real impact; which is the core DNA of ADC. Our CTO, Casper Rutjes, explains the significance of this release and how ADC is addressing it below. 

The release of GPT-4o, a model capable of real-time, multi-modal understanding (text, audio, vision), has once again set new standards. This model can interpret complex inputs and react nearly as quickly as a human conversation partner.

At ADC, this represents a leap forward from previous models which became obsolete faster than anticipated. It teaches us the importance of agility and foresight in technology adoption.

OpenAI has unveiled GPT-4o, its new model designed to revolutionise human-computer interaction by reasoning across text, audio, vision, and video in real time. Dubbed “omni” for its comprehensive capabilities, GPT-4o accepts inputs from any combination of text, audio, image, and video, and can generate outputs in text, audio, and images.

The model achieves rapid response times to audio inputs, averaging 320 milliseconds, comparable to human conversational speed. Since GPT-4o is the first model to combine all of these modalities into a single network, the extent of both its capabilities and limitations are unknown.

GPT-4o’s ability to process multiple modalities through a single neural network architecture also resolves long-standing issues like latency in real-time interactions.

This advancement aligns with R&D work we have done at ADC to overcome similar challenges with a multi agent approach. However, our developed solution again feels just as cumbersome and ready to be put in the code-graveyard.

ADC’s approach to R&D has evolved to focus more on curation, partnering, integration and strategic management of new technologies.

By learning from rapid advancements like those seen with GPT-4o, we prioritise a comprehensive but pragmatic application of AI, helping clients with real impact. This involves robust risk management, Responsible AI, dynamic change management, and a keen focus on the practical implications of AI that add value in businesses and organisations.

Interested in learning more about ADC’s approach to implementing AI in a safe and responsible manner? Reach out to our CTO, Casper Rutjes, for an informal chat.