Why did I delete ChatGPT after 3 years of use? The complete guide to Claude that no one else is showing you

While ChatGPT competes with other neural networks, Claude, an artificial intelligence created by former OpenAI employees and funded by Google, is quietly but steadily growing. The brother and sister assembled a team of top OpenAI engineers and released Claude with Constitutional AI. This means that while ChatGPT simply tries to please people but doesn't always understand why something is right or wrong, Claude goes through a special filtering process before giving an answer, namely through a constitution with more than 50 clear principles.

Because of this, the answer becomes much better than in Chagi Piti. But is everything perfect in Claude? How is it better than Chagi Piti? And how is it worse? I analyzed all the life hacks, and in this article, I will only talk about what is really useful for everyone, not just programmers. We will figure out which models you need for your task, how to create assistants within Claude, how artifacts help you create applications, documents, games, and generally simplify your communication, and I will also share secret life hacks that allow you to use Claude to its full potential.

At the end of the article, I will explain why you should not use Claude for all tasks. As always, all the information from this article is available on our website. After you log in to the Claude website, the familiar chat window opens.

But there are a couple of distinctive features here. First, we can switch between the Claude, Sonnet, Opus, and Haiku models. We see a note about Sonnet saying that it is ideal for everyday tasks and is the default model.

This is true because this model provides a balance between speed, quality, and integration cost. Overall, this is the only model I use for everyday tasks, analysis, and working with files. Opus is officially recognized as the best for programming, but it is also suitable for more advanced analysis.

The third model is Haiku. It has low quality compared to other models, but it is the fastest and cheapest. Therefore, it is often used in chat bots, for stores, for comment moderation, for quick code or text corrections without in-depth analysis. 

The cost of using models for integration is also available on the website. The code has a larger context window than many versions of chat.gpt. What does this mean? The context window is the size of the dialogue within your chat with artificial intelligence, during which all the information you wrote within a single dialogue is remembered. In chat.gpt, you have probably encountered a situation where you write a task to it, taking into account the previous context and message, but it responds with general information unrelated to your context, or, conversely, writes something that is not what you need.

The reason is that you have reached the context window limit, and the previous part of the context in the dialogue has simply been deleted, and a new context window has opened, and you are wasting time retelling the context and clarifying the task again. So, Claude remembers more information during the dialogue, and if we exceed the context window limit, it simply pauses the chat and does not allow us to continue writing in this dialogue. In this way, Claude protects us from stress and wasted time re-explaining the context.

Here's an interesting life hack. If you think you've gotten enough valuable info in one chat, I suggest making a document with the key ideas and context by just asking Claude to do it. You can use this text document in new chats or just save it in your note-taking app, like Obsidian.

Artifacts in Claude: Why It's Killing Canvas from ChatGPT

The documents that Claude creates are called artifacts. By the way, Claude's work with artifacts is much better implemented than working with Canvas for gpt. Firstly, as a screenwriter or student, I often come across the fact that gpt does not accurately determine the number of words or characters of messages.

For example, I often ask him to reduce the size of my text to 3,000 words, but instead I get a 700-word text cut. But I didn't see such a problem in Claude. He does not create and format the text as an artifact, and all these artifacts are stored in a separate place.

On the left side of the panel, we can see my artifacts. Here you can store the key ideas of all the dialogues. Secondly, you can use artifacts to create interactive games and applications.

And it works much better than we did in the gpt chat, because Claude writes code much better. For example, let's ask him to create a game based on a self-assessment book. If we try to upload the book directly to the chat, it may give an error stating that the file exceeds the allowed size. 

To do this, we simply upload this file to our Google drive and open this file using documents. Then we open access to it and share the link to this file in a chat with Claude. And we write him this prompt, which will also be on my website.

And he creates an interactive game that will help us not only remember the material of the book we read, but also gradually learn to implement the principles in life. Thirdly, artifacts do a great job with documents. He can create checks and receipts that are very realistic, but naturally as a sample due to the principle of law-abiding.

However, if you need to translate the document into another language, while maintaining all the structure, data and design, then it will handle it perfectly. He also makes minimalistic presentations with animations and graphs based on your data, for example, he will analyze your financial condition and give you recommendations. Moreover, I like that when a large response is planned, he transfers it to a separate artifact so that it is easier for me to read it.

And this artifact can not only be downloaded as a text document, but you can also share it with a friend via a link. Claude accepts text documents, photos, and can connect to third-party services. It may not accept text files, such as PDF books, if it exceeds a certain size, for example 100 pages.

And if it exceeds the size, then you can upload it to Google drive, as we did earlier, or upload it to projects, about them a little later. Claude can also receive and analyze information from mail, Google Drive, and calendar, just like Gemini does. For example, I sometimes ask him to make a summery of the last letter in the mail and save it in an artifact.

Recently, connectors were delivered to Claude, that is, tools for connecting Claude to popular services in one click. You can connect Notion, Figma, Canva and other tools and use Claude as a no-code suggestion to create designs and texts that will immediately appear on your respective platforms. Speaking of no-code, we recently published an article on our website where we explained what it is and how it can be used to automate your life.

If you are interested in how Claude helps automate your life using third-party services, write about it in the comments, I will definitely prepare a separate material about it.

Response Styles and Extended Thinking: Secrets That Change Everything

In Clod, you can customize the response style depending on your mood and desire. If in Chaji 5 we just specified the general style in the settings and he followed this style in all the dialogues, even when I just needed to fix the door handle at home, he gave me motivational phrases for some reason.

So in Clod, you can set up a situational response style. For example, to respond to an email message in English, we create a specific style, and to write articles and texts, we also use a separate writing style. Many users create their own styles to maintain a unified brand voice when creating advertising texts and posts on social networks.

To create styles, click on the tools button, select use styles, and by default there are styles that I also use. Normal, Konkraise, Formal, and Explanatori. Normal is the standard answers, no more, no less.

Conkrais are short answers, purely to the point. Formal answers are more scientific and formal. Explanatori is my favorite educational answer, which is well suited for learning new information.

Here we create our own style, click on the create and edit styles button. Here we can simply upload pre-prepared instructions on how Claude should respond to us, or simply insert a file with the text of your favorite blogger or other source and Claude will also respond as in this file. Or we can create a response style using Claude's hints.

I will choose just the last clue and write that it will be a psychological style. And I want to know the name of the psychological problem, the ways to solve it, and let him ask 5 clarifying questions. Click on create style, then I change the name of the style to my own, then save and use it in the following chats.

We will send him a message that we often lie instead of real actions and we see that he responds as our psychologist. By the way, such promptas are already available on my website in the prompta section, you can take them from there and improve them for yourself. Also, Clod has an extended thinking feature, which I have by default in every request I make.

She allows me to see the logic of Claude's reasoning for solving my question. It fits well with the exploration mode and the explanator style. With these three settings, the research is thorough, with up-to-date information and with a clear presentation.

Unlike gpt-o3, where the thinking process is shown rather blurrily, Claude puts each step on the shelves and we see exactly how he came to this conclusi

Projects in Claude and Honest Restrictions: What is Being Hushed Up in Other Reviews

In my opinion, projects in Clod work much more efficiently than in gpt. They are ideal for repetitive tasks, for example, I have a blog on my website, psychological analysis, website optimization and other other projects.

Here you install the system instructions for the specified project, add files, books, texts, and attach files that are on your Google drive or github. I like that you can access projects just from a regular chat, you don't need to go anywhere, just click on the plus icon and select the appropriate project and write your request to it. An interesting life hack, you can automate the input of large promptings using applications such as weave text or raycast.

You write one word and a long prompt is inserted. To do this, you need to configure it in the weave text application, enter the full text and write a substitute word. This is especially useful when a routine is repeated within the framework of a single project, for example, a study was conducted on this topic. The only serious problem is the limitation on the free rate.

On the free plan, you can write 4-5 requests per day, in general, there is little you can experiment on the free plan. However, the cost of a paid subscription is the same as in chat gpt, $ 20 per month. But even paid subscriptions have a limit on the number of requests.

The max subscription almost removes all restrictions, but it costs a lot, $ 200 per month. Therefore, I use a combination of claude and chat gpt using artifacts, which contains the key ideas of the dialogue. Claude, like other neural networks, is not perfect or universal yet.

Despite the large context, he can still forget the details of the dialogue and, in general, sometimes even gets confused in short dialogues. Therefore, artifacts with key ideas of the dialogue are saved. I can't recommend it for all tasks, but it does a great job with deep analysis, programming, and education.

And in general, with the help of writing styles, you can create personal assistants right inside Claude, if you want a psychologist, if you want a lawyer, or if you want an analyst. Personally, I started using it to teach programming, because the games and the material it creates are more informative than chat gpt does. Working with files is convenient, especially if you connect Google, nowshen, Figma and other know-how pieces.

Thus, I use Claude for advanced analysis, for teaching, and for writing large texts. Chat gpt for fast and everyday tasks. Gemini for interacting with the Google ecosystem and working with media content.

Stay tuned for updates on the site, in the following articles we will analyze in detail Claude's interaction with other services.