The AI app ChatGPT is trained on the entirety of the web—a staggering dataset currently exceeding 1,200 petabytes (1 petabyte equals 1 million gigabytes). To put that into perspective, imagine each gigabyte in this dataset as a book.
In simpler terms, the indexed web resembles a vast library containing over 1 trillion books—approximately 200 times larger than the combined collections of all academic libraries in the United States. This enormous dataset fuels the engine of artificial intelligence and its Large Language Models, serving as the lifeblood that powers the machine.
High velocity change
The sheer volume of data within the AI universe isn’t the only factor unsettling our equilibrium and challenging our capacity to adapt. Equally critical is the rapid pace of its evolution, which demands careful consideration of how we navigate and respond to this unprecedented change.
I have been in the tech industry for 40 years and I have never seen this pace of change. In just two years, we have created a $1 Trillion industry. And $600 Billion has poured into AI investments including data centers, GPUs and software development and the pace isn’t slowing but accelerating.
Why this matters
Being lost in the clutter of all this change needs us to stand back and get perspective.
To see the big picture and find what matters to us individually both as humans and entrepreneurs.
Trying to find opportunity in chaos is sometimes an art and sometimes dumb luck.
This mammoth universe of data contains our human creativity, intelligence, information and wisdom. It used to be found in books and in the verbally shared wisdom of our elders that was revealed in stories and parables around the tribal fire and from the church pulpit. Making sense of that and finding signal and wisdom in the noise is complicated.
So how do you get to the information and distill all that noise into commonsense and wisdom that guides us forward into a world that makes sense? The answer is AI. It isn’t artificial intelligence but human wisdom captured by a tool we created. Humans are also not designed to consume and make sense of all that data. Our DNA and our genes are wired for tribal and primal motivations.
We have now escaped our natural habitat built over hundreds of thousands of years and most of us are confused and maybe a bit frightened as we now sit in a landscape that our tribal roots don’t recognize. ‘In the blink of an eye’ in the context of our human evolution we have gone from a hunter gatherer and then to an agrarian economy, to industrial and then to a knowledge and digital world and now to an AI economy.
So where are we now and where are we going with AI? These 7 phases provide some answers and perspective.
7 phases of ChatGPT (AI) since 2022
Amidst this rapid evolution and overwhelming volumes of data, it’s essential to pause and gain some perspective. Here’s a breakdown of what has unfolded over the past two years since ChatGPT’s launch on November 30, 2022.
Each phase highlights a key application along with a usage example for context and clarity.
Phase 1: Introduction of ChatGPT (November 30, 2022)
ChatGPT was launched by OpenAI on November 30,2022, making powerful language models accessible to the public, primarily for text-based interactions. AI now has an easy to use interface. AI has now been democratized.
You put in a text prompt and it provides an answer gleaned from the entire web that has all human knowledge and information sitting in data centers that have backbones of over 1 mile.
Simplicity distilled from the orgy of human complexity
Key App: ChatGPT
Its user interface couldn’t be simpler. And sort of reminds me of Google’s search box.
Usage example: ChatGPT is now used by over 1.5 billion users a month. It is used for answers to any question and it only took 8 weeks to reach 100 million users. AI was democratized.
Phase 2: Rise of single-mode applications (Early 2023)
AI was then seized upon by text platforms such as Grammarly and it continued to be utilized primarily in text-based applications. Single mode use was first. It helped people write articles, blog posts, create emails and much more text based applications.
Key App: Grammarly (AI-enhanced)
Usage Example: Grammarly uses AI to provide real-time writing assistance, offering suggestions on grammar, tone, and style to help users improve their writing across emails, reports, and social media posts.
Phase 3: Expansion to multimodal Models (Mid 2023)
Text was the simplest first step after decades of AI development and now richer and more complicated media beckoned. Images and video.
AI now started to expand into multimodal capabilities, integrating text, images, and video. Developments in video included AI-driven animation, real-time video editing, and enhancements, setting the stage for more complex video applications.
Key App: DALL-E
Usage Example: DALL-E is utilized by graphic designers and content creators to generate unique images from textual descriptions, streamlining the creative process and enabling the creation of bespoke visuals for marketing and advertising.
Phase 4: Introduction of AI-Driven animation and synthetic media (2023)
This period saw the rise of synthetic media where AI began creating video content from textual descriptions, including deepfakes and AI-generated characters, raising both opportunities and ethical concerns.
Key App: Synthesia
Usage Example: Synthesia creates AI-driven video content where users can generate videos with AI avatars that read scripts in multiple languages, used extensively for creating scalable, personalized video messages in customer service and marketing.
Phase 5: Integration and API Accessibility (Late 2023)
ChatGPT is what we would call a horizontal AI. Generic and broad based. Allowing access to all that information that is human intelligence distilled in a general sense is one thing. But making it usable for all and specific for our individual needs is another.
With the opening of its API to allow startups to access its Large language models (LLM’s) and data we started to see the building of specific apps for vertical niches.
APIs for various AI models, including those capable of video processing, were made widely available, allowing for greater integration into custom applications across different sectors.
Key App: OpenAI API
Usage Example: The OpenAI API is integrated into various business applications to automate tasks like data extraction from documents, language translation, and sentiment analysis, enhancing efficiency across operations.
6: The ChatGPT App Store (2024)
The introduction of specialized applications and plugins based on the ChatGPT model extended into the video domain, enabling more sophisticated video functionalities such as live moderation and interactive content generation.
This is a ChatGPT app store that mimics to some degree Apple’s app store and so its potential revenue is still to be realized. To put some perspective on this potential, the annual revenue in 2023 for Apple from its app store is $89.6 billion.
Key App: Descript
Usage Example: Descript offers advanced video and audio editing capabilities, where users can edit media files as easily as editing a text document. It’s used for podcast editing, video content creation, and producing transcriptions automatically.
Phase 7: Emergence of AI video agents (2024)
AI agents capable of acting more autonomously included video agents that could interact in real-time, conduct interviews, and even host shows, indicating a move towards more dynamic and engaging AI applications in video content.
Key App: Replika
Usage Example: Replika, initially a chatbot, now includes video interaction features where the AI can serve as a virtual companion. Users interact with AI through both text and video, using it for social companionship and basic mental health support.
The next phase of AI (The Future)
The next emerging phase of AI is where ideation and content creation can be acted on by AI. And it is called “Agentic AI.”
According to the Nvidia blog “Agentic AI systems ingest vast amounts of data from multiple sources to independently analyze challenges, develop strategies and execute tasks like supply chain optimization, cybersecurity vulnerability analysis and helping doctors with time-consuming tasks.
This has exciting possibilities. As humans we have many ideas and what stops most of us is acting on them. Procrastination and fear of failure often stops us from taking ideation to creation. Producing an idea and then acting on it. The real magic happens when we create something and release it into the wild and see what unfolds. What if you could come up with an idea and ask AI to act on it? That is an Agentic agent.
One app that is chasing that promise is the Sales force AI agent branded “Agent Force” that says it is the next wave of the AI evolution. Taking ideation and information to autonomous action.
Most of us are good at coming up with ideas but most of us are also good at procrastination. Ideas devoid of action.
That is where an AI agent can help. It is early days but there is nothing wrong with dreaming.
The bottom line
High velocity change driven by AI is the new normal. So be prepared to have your life and business both threatened and amplified by AI. The dystopian threat we fear will fade after we realize that Hollywood’s version of the world is a dramatic facade to sell more tickets and make more money. As the newspapers discovered. “If it bleeds it leads”
Utopia is also a distant nirvana that we all dream about but its promise will always escape our grasping hand. The truth lies somewhere in the middle. The AI adventure continues.
We live in amazing times.
The post Blink and You’ll Miss It: The Unstoppable Rise of High-Velocity AI appeared first on jeffbullas.com.
* This article was originally published here
No comments:
Post a Comment