ChatGPT vs Deepseek - Part 1 #chatgpt #deepseek #ai Response Evaluation (Basic Context Handling) Goal: Test comprehension, clarity, and ability to follow a single-threaded context. 1. Prompt 1 (Basic): “What is photosynthesis? Explain in simple terms.” 2. Prompt 2 (Building Context): “How does photosynthesis contribute to the oxygen cycle?” 3. Prompt 3 (Integration): “Explain how the oxygen cycle impacts human life and ecosystems.” 4. Prompt 4 (Synthesis): “Summarize the relationship between photosynthesis, the oxygen cycle, and human life in two sentences.”

And so I want to just illustrate something for you guys very quickly. This is DeepSeek on the side here, and this is ChatGPT, and what I'm going to do is I'm going to run a few prompts so you can just get a comparative idea of how these two models perform differently next to each other. And this is really important because right now we are like in the Manhattan Project 2.0, and the Manhattan Project was about, you know, how do we build, you know, nuclear bombs, right? But this is really it. This is really the magnitude and scale of what we're dealing with in artificial intelligence. And right now, China just released DeepSeek, and they have their R1 and V3 models. The R1 is probably something that's comparative to ChatGPT's O1 advanced reasoning model, but we'll get into that because I've been using DeepSeek, and if the American government was afraid about TikTok and being able to get all of the information and behavior and deepest, darkest thoughts of Americans, then DeepSeek is like just injecting it right into our veins. Like, here's my social, right? Here's everything about me that you can have. And so I want to be able to illustrate just some prompts. You know, we're going to go through this. I'll probably break this up in sections where we're going to go just basic context handling, some immediate context handling. We're going to do creative context handling, multi-step logic and execution, some advanced reasoning and critical thinking, and then we'll do some advanced contextual challenges. And so I want to kick this off just so you can see what this looks like, because DeepSeek is about 30 times less expensive than ChatGPT, right? And so all of this uses tokens when you write something. So if you think about what tokens look like, if you were to write a novel, I did, I created a token calculator in Cloud. So if you create a novel to a book, it's like 2,500 words, a research paper is 1,500 words. So we just kind of maxed this all the way out just to see, you can see how crazy it gets in terms of cost per 1,000, right? Cost per 1,000 tokens, $22, cost per 1,000 tokens, $483, just in terms of scale. And so, all right, and so that's just some context in terms of why the American government is investing $500 billion in AI infrastructure is because they don't want to lose to China in this AI wars race. But anyway, we're going to get to just some basic simple prompting, right? And we're going to say, what is photosynthesis in terms, so I'm going to put these in here and you'll get a good idea of like how fast the models are even moving, right? And so you can see the answers, the speed in which an answer and how thorough each one actually gives an answer. And then we're going to do explain. And so this is our first just basic context handling. And then we're going to review what they say, and then summarize in sentences. And so we're going to see here, ChatGPT is still talking, DeepStink is still thinking, and then there we go. All right, and so that's all of them. And so we want to see what ChatGPT did. Photosynthesis is the process by which plants, in simple terms, and we're going to come back up here also to DeepSea just to see. In simple terms, plants, photosynthesis is the process which green plants, large algae, and some bacteria use sunlight to make their own food, to make their own food. They use sunlight. And so you can just see the structure in which it gave an answer. It's how plants eat, in simple terms, sunlight, carbon dioxide, water, chemical reactions, glucose, oxygen. So it's interesting in terms of how we lay this out. So how does photosynthesis contribute to oxygen, contribute to the oxygen cycle? And so what happens here when you think about context is that we ask for in simple terms. And so oftentimes, the large language model will remember the context in terms of how it's interfacing or discussing with you. And so oxygen production, oxygen production, atmospheric oxygen, respiratory, I mean, it's a more, it's definitely a more verbose answer within DeepSea versus ChatGPT. And so here we go on the third question. Explain how the oxygen cycle impacts human life and ecosystem. So the oxygen cycle is essentially for human life and ecosystem because it maintains the balance of oxygen. Supply of oxygen, which is essential for survival, impacts on human respiration, health and well-being, health, medical applications, impacts on the ecosystem, animal life, aquatic ecosystem, plant life, decomposition, climate regulation, and then it offers a summary. And then I like to look at just when we get in some of these more, okay, the summary. How robotic does it sound? Photosynthesis drives the oxygen cycle by covering sunlight, carbon dioxide, and water into oxygen and glucose. Glucose produces oxygen as a byproduct, which is released into the atmosphere. And so this is really interesting in terms of how they lay it out. And so this is our first test. And I just want you guys to just see that and provide some feedback on what you think about this, because the next one I'm going to do is going to be on research synthesis.