Google Deepmind’s next level AI: Gemini all set to surpass chatGPT
In a somewhat provocative new interview with wired magazine Demis Hassabis head of google deepmind is quoted as saying that Gemini which could be released as soon as this winter will be Surpassing OpenAI’s Chat GPT in capabilities, an exciting revelation unfolds. Efforts are underway to synergize the formidable strengths of Alpha- Go type systems with the extraordinary language capabilities of large AI models.
Before we look in to how that might work here is the context of the Gemini announcement from Sundar pichai”we are focused on building more capable systems safely and responsibly. This includes our next generation foundation foundation model Gemini which is still training. While still early we are already seeing impressive multimodal capabilities not seen in prior models.” Hassabis promises that ”we also have some new innovation that are going to be pretty interesting.” Gemini functions based on the concept of addressing diverse data and assignments without depending on specialized models. It guarantees the creation of unparalleled material that surpasses the limitations of its training data, representing a remarkable advancement in the domain of artificial intelligence. Many people will dismiss this as all talk but remember deepmind was behind not just Alpha go but also Alpha Zero which can play any two player full information game from scratch.
So Let’s not underestimate Deepmind but back to Gemini we hear from the information recently that the multi modality of Gemini will be helped in part by training on Youtube videos and apparently YouTube was also mined by openai. Off course that’s not just the text transcripts but also the audio, imagery and probably comments. I wonder if Google Deepmind might one day use YouTube for more than that. The initiative also adopts reinforcement learning, an iterative methodology in which the software consistently strives to accomplish objectives and improves its efficiency through feedback.
If you remember the planning bit from Deepmind’s earlier systems that reminded us of something else from Gemini’s introduction in which Sundar pichai told” Gemini was created from the ground up to be multimodel highly efficient a tool and API integrations and built to enable future innovations like memory and planning ”
Nevertheless, it is crucial to acknowledge that Gemini’s existing functionalities predominantly revolve around text manipulation, unlike GPT-4, which boasts the capacity to process images, audio, text, and video. Despite this constraint, Gemini strives to provide more inventive and innovative replies, with the objective of surpassing the constraints imposed by its training data and producing unforeseen content.
As Gemini continues to evolve, it emerges as a project that garners substantial attention. The prospective triumph of Gemini possesses the ability to redefine the AI sector and establish unparalleled standards for the capabilities of artificial intelligence.Let us patiently await the launch of Gemini and observe whether it will successfully break new ground in the field of AI or remain in its current state.
