- Sensibility.ai
- Posts
- Cappital | Sensibility.ai - Apple Intelligence Vs. Google's Gemini Nano (1)
Cappital | Sensibility.ai - Apple Intelligence Vs. Google's Gemini Nano (1)
Comparing AI Integration in Smart Phones.

The Progression of AI
AI started with a single text-to-chat model. The user would write something whether it would be a question or a statement and the model would send back an educated answer which at the time was incredible. That was tech innovation at its finest. One of the main reasons AI progressed so fast, especially that model is because of Open AI. OpenAI was founded in December 2015 by a group of prominent tech entrepreneurs and researchers, including Elon Musk, Sam Altman, Greg Brockman, Ilya Sutskever, Wojciech Zaremba, and John Schulman. The organization was established with the mission to ensure that artificial intelligence (AI) is developed and used in a way that benefits all of humanity. OpenAI was created as a non-profit research organization, with the founders committing over $1 billion to its establishment. The company was founded in response to the growing concerns about the potential dangers of AI if not developed with proper oversight and ethical considerations. By focusing on creating safe and beneficial AI, OpenAI aimed to lead the field in a direction that would minimize risks and maximize positive outcomes for society.
The development of ChatGPT and its underlying models has evolved significantly since it was created, reflecting rapid advancements in natural language processing and artificial intelligence. The Chat GPT began in June 2018 with the release of GPT-1, the first version of the Generative Pre-trained Transformer (GPT) model. With 117 million parameters, GPT-1 was primarily designed for natural language understanding and generation, laying a critical foundation for subsequent models. Its introduction marked the start of a new era in AI language modeling, setting a precedent for future innovations.
In February 2019, GPT-2 was released, showcasing a more powerful model with 1.5 billion parameters. Unlike its predecessor, GPT-2 was introduced in stages due to concerns over potential misuse; however, the full model eventually became available in November 2019. The increased capacity of GPT-2 significantly improved its ability to generate coherent and contextually relevant text over longer passages, demonstrating a substantial leap forward in the capabilities of AI models.
The release of GPT-3 in June 2020 marked another major milestone in AI development. With 175 billion parameters, GPT-3 became one of the largest and most powerful language models at the time. Its ability to understand context and generate human-like text was vastly enhanced, leading to its widespread adoption across various applications, including OpenAI’s API. GPT-3 also served as the basis for the first version of ChatGPT, a specialized product tailored for conversational AI, which optimized its capabilities for maintaining coherent and contextually relevant conversations over extended interactions.
In November 2022, ChatGPT was further refined using GPT-3.5, which improved its dialogue management and conversational abilities. This version of ChatGPT demonstrated a higher proficiency in generating relevant responses, making it a more effective tool for natural language interaction. The evolution continued with the release of GPT-4 in March 2023, which offered even greater advancements in AI capabilities. With enhanced understanding of complex prompts, better adherence to instructions, and improved generation of creative text, GPT-4 represented a significant step forward. Notably, it introduced multimodal capabilities, enabling the model to process both text and image inputs, making it the first in the GPT series to support image generation.
The integration of image generation into ChatGPT was finalized in October 2023, when OpenAI incorporated a specialized DALL-E model, allowing users to generate images based on text prompts directly within the ChatGPT interface. This addition greatly expanded the utility of ChatGPT beyond text generation, opening new avenues for creative expression and practical application.
Building on these innovations, OpenAI introduced video creation capabilities on February 15, 2024, with the preview of a text-to-video model named Sora. The preview included several high-definition video clips, such as an SUV driving down a mountain, an animation of a monster, people walking in Tokyo, and fabricated historical footage. Sora employs natural language processing to draw from a large database of information and create original content based on written text prompts, akin to other OpenAI tools like DALL-E and ChatGPT. The model begins with each frame as static noise and uses machine learning to gradually transform these frames into visuals that resemble the description provided by the prompt. Following this preview, OpenAI collaborated with filmmakers to test Sora, including a partnership with the Tribeca Film Festival. By March 2024, OpenAI had released new videos created with Sora, such as the short film "Air Head" by Shy Kids. In May 2024, a music video generated by Sora for the song "The Hardest Part" by Washed Out was also released.
Where are we Now?
Now where are we? With the release of Sora and the high-definition videos that Open AI has shown it can create I didn’t know where AI could go from there. Was AI at a plateau and could it do anything better? Then I came across an AI company called Exists.AI. Exist .ai is an AI-based gaming company that has the goal of making indie game creation just that much easier. They are in the process of developing an engine that uses AI to create digital worlds and or landscapes based on user text prompting. When you really think about it this development in AI advancement will benefit so much more than just game creation. When boiled down into lamens terms this engine could be used as a text to 3D Model which could benefit Landscapers, architects, and so much more.
This Week in AI:
Elon Musk is building an AI Training Supercluster in Austin, Texas, marking a bold step in his quest to lead in the artificial intelligence space. This new facility aims to create one of the most powerful AI training environments in the world, leveraging state-of-the-art infrastructure and vast computing power to accelerate the development of advanced AI models. With the Supercluster, Musk plans to enhance AI capabilities across his companies, from Tesla's self-driving technologies to potential applications in robotics and energy solutions. Positioned in Austin, this initiative takes advantage of the region’s growing tech presence, abundant resources, and strategic location, allowing Musk to push the boundaries of AI research and development in pursuit of future technological breakthroughs.

Sensability.AI is a weekly newsletter to keep you up to date in the ever-changing landscape of AI by the team at Cappital.co
Is there a tool you want us to look into?