From the course: What Is Generative AI?

Productivity enhancements in large language models (LLMs) through APIs and real-time interactions - DALL-E Tutorial

From the course: What Is Generative AI?

Productivity enhancements in large language models (LLMs) through APIs and real-time interactions

- The usage of Large Language Models, also referred to as LLMs has evolved significantly since it first became mainstream. We went from standalone LLMs like ChatGPT, to the dynamic world of LLM APIs and the latest GPT-4o, which enables real-time conversations and solutions through voice and picture commands. OpenAI made the new model of GPT-4 available. A more versatile model that is also multimodel and is better, much better in handling complex queries. GPT garnered 1 million users in the first week of its launch, ramping up to a hundred million users in two months, proving the demand for a text spaced AI assistant. All the other technology companies, of course, jumped right into the gold rush, and started coming out with their own models. Google's Deep Mind launched Gemini, X formerly known as Twitter launched Grok, Meta launched Llama, and Nvidia launched Megatron Turing. What a great name. In December 2022, the public availability of GPT provided a glimpse into the potential of LLMs. However, the recent development of LLM APIs, I will get there, has unlocked a new realm of accessibility and integration. I started hearing you LLM, API, API. What is an API? An API is short for Application Programming Interface. Imagine it as a waiter in a restaurant. When you order, you don't go into the kitchen with your own feet to make your own food. Instead, you tell the waiter what you want from the menu. The waiter takes your order to the kitchen and then brings back the finished dish back to you. While similarly, an API is the waiter for the software application. It takes our prompted requests like ordering food to the main LLM engine, the kitchen, which processes the request and sends back the desired output, the finished dish. This way, different software applications can communicate and interact with each other without needing to know the complex details of what happens behind the curtains. Since GPT made API access available to their main engine, they allow developers to integrate the power of GPT models into their own applications, their own products, their own services, of course, after custom fine tuning GPT to the unique needs of their products. For example, Stripe customer support uses GPT APIs that are fine tuned with the in-depth information about Stripe. Other renowned companies that use GPT, API are Zapier, Jasper, Duolingo, and Shopify. In May 2024, OpenAI introduced GPT-4o that allows real time conversations, just like in the movie "Her." If you have not seen it, please go see it. It's amazing. This news marks a monumental leap in human AI LLM interaction because it facilitates problem solving through voice and image commands mimicking like real world conversations between two humans. This mimicry of organic communication brings a completely new level of immediacy and intuitiveness to our AI interactions. Of course, I'm going to say that as I always do, with this, "Best thing since sliced bread" awe, that we feel initially for this development, which is really amazing though comes my necessary caution. The more our interactions with AI will become organic, the more we will risk the danger of forgetting that AIs are not humans. There's a very big difference. They are far from that. AIs are tools trained on a finite limited data set, meaning if the information we are seeking for is not in their data set, we risk being misinformed. If we were to believe that LLMs are the ultimate source of truth, our opinions would end up being manipulated by the makers of the AI tools. Well, consciously or unconsciously. That's why this year and in the next years when working with LLMs, we will learn to develop a healthy excitement coupled with caution. Actually, my second LinkedIn Learning Course, "Build AI Aptitude in your Organization as a Leader," is exactly about that. It's about developing the skill sets to consciously integrate advanced technologies in our workforce. Check it out if you have not. I promise, I promise not to bore you.

Contents