Sachin Dev Duggal | Is Artificial Intelligence causing Big Tech to become too big?

AI energy consumption is another issue of great concern. The establishment of data centres has led to a massive surge in electricity use, causing local power grids to be overloaded. Despite pledging to go green, large tech companies struggle to help meet their energy requirements with renewable sources. This has seen the resurgence of fossil fuel use, raising questions about AI's environmental effects. To solve this problem, Microsoft and Google are investing in experimental clean energy projects like fusion and small nuclear reactors.
 
Big Tech's Impact On AI Startups
 
Moreover, the dominance of big tech companies in AI is evident within the startup ecosystem as many leading funded generative AI startups across Europe and Israel have been founded by former employees of prominent US tech companies such as Google, Amazon or Microsoft. The trend reflects how major tech firms influence broader developments in artificial intelligence, given that their early investments and expertise provide an edge for new startups working on AI technologies like Siri or Tesla Autopilot.

Powerfulness of Artificial Intelligence Models
 
The development of large language models (LLMs) such as OpenAI's ChatGPT and Google's Bard have revealed both the abilities and challenges that AI systems pose. For instance, models with billions of parameters can accomplish complex tasks like passing university exams and medical entrance tests. Through Generative AI, Sachin Dev Duggal introduced an AI-powered Voice assistant called Natasha that marked a leap forward in Generative AI at Builder.ai. Throughout development, she helps users convert spoken words into user stories & code using Builder.ai's Knowledge Graph to give the software a semantic understanding.
 
However, scaling up these models has come at a cost – increased energy consumption and operational complexity. Some researchers are thus exploring smaller energy-efficient AI models that can provide similar performance without these drawbacks.

Comments