Listen to this Post
In the ever-evolving world of generative artificial intelligence (GenAI), there has been an ongoing discussion about the level of differentiation between the major players in the space. Despite significant advancements, CEO of Perplexity AI, Aravind Srinivas, recently addressed a common concern about the current landscape of AI chatbots. He noted that, while competition is fierce, the core capabilities of many leading chatbots remain largely indistinguishable. In a conversation with Zerodha co-founder Nikhil Kamath, Srinivas also touched on potential differentiators, the future of AI chatbots, and how user preferences may shift in the coming years.
As AI technology rapidly progresses, one of the most pressing questions is: What sets apart today’s most popular chatbots, like ChatGPT, Gemini, Anthropic, and Perplexity? According to Srinivas, the differences between them are slim, particularly when it comes to the core tasks of generating text and answering questions. The AI models, despite the branding and innovation from companies like Meta, Google, and OpenAI, often provide similar outputs as they chase the same benchmarks.
Srinivas also shared insights into
Looking ahead, Srinivas believes the next major evolution for AI chatbots will center around the concept of “agentic behaviour.” Unlike the static text-based interactions we see today, future AI systems will be able to actively perform tasks and not just generate responses. This shift, he predicts, will be the next big leap for AI in 2025, as more sophisticated reasoning and action-based capabilities are developed. As AI chatbots transition from mere conversation tools to proactive assistants capable of handling complex tasks, the landscape of generative AI will continue to evolve.
What Undercode Says:
The current state of AI chatbots, as described by Aravind Srinivas, reveals a major truth about the generative AI space: differentiation is still lacking. Despite the heavy competition, leading models from major tech companies have not yet demonstrated marked qualitative differences. They are all relying on similar evaluation metrics and benchmarking, which leads to similar outputs for most user queries. This is perhaps the most significant takeaway from Srinivas’s remarks. The AI industry seems to be stuck in a cycle where only the surface-level aspects, such as branding and certain features like speed and accuracy, separate these models.
This raises an important question about how users are choosing between different platforms. If the answers across all AI models are largely similar, then the choice of AI may come down to small differences, such as the incorporation of sources in the responses or the ease with which users can interact with the platform. But the real question is: Are these nuances enough to keep users loyal to one platform? Or are we seeing a shift toward standardization where users simply rely on whichever platform is available or most convenient?
As Srinivas points out, fact-checking and research might be the primary area where significant differentiation exists at present. However, this could change in the near future, as more advanced AI systems develop the capability to perform tasks beyond just answering questions. The evolution toward “agentic behaviour” is what sets the stage for the next major leap in AI technology. While answering questions is the current norm, future systems will need to offer richer, more dynamic interactions, such as generating charts, graphs, and even carrying out complex tasks on behalf of the user.
This potential shift reflects a broader trend in the AI industry: the need to move beyond simple conversation and toward AI models that can actively do things. Srinivas’s forecast for 2025 indicates that AI systems will increasingly be designed with sophisticated reasoning and active decision-making capabilities, opening up new possibilities for how these models can assist in fields like finance, research, and everyday personal tasks.
What this all boils down to is the notion that the future of AI chatbots isn’t just about answering questions more effectively—it’s about transforming AI from a passive tool into a proactive assistant. Users will no longer just ask questions and get answers; they will expect AI to perform tasks, generate real-time visuals, and act on behalf of the user in a much more tangible way. This is the trajectory that could set the leaders of the AI space apart, as the focus shifts from simply providing text-based responses to delivering actionable, multi-faceted outputs.
Fact Checker Results:
- Current Chatbot Similarity: It’s accurate that many leading AI models are producing similar outputs due to shared benchmarks and focus on comparable metrics.
2. Source Implementation: Perplexity
- Shift Towards Agentic AI: Predictions about AI moving from text generation to task execution align with ongoing advancements in AI reasoning capabilities.
References:
Reported By: timesofindia.indiatimes.com
Extra Source Hub:
https://stackoverflow.com
Wikipedia
Undercode AI
Image Source:
Unsplash
Undercode AI DI v2
