Memory and browser cookie-based memory of past chats (other people won’t see your history even when logged in with same account). Right click on the chat histories to rename or delete.
Claude 3.5
This is an Anthropic.com model called Claude 3.5 model with 150k word. Claude 3.5 is Anthropic’s most powerful model for highly complex tasks. It boasts top-level performance, intelligence, fluency, and understanding. Knowledge cut-off August 2023. Claude is better suited for professional/business outputs and is more likely to say if it can’t help with a task.

GPT4o

OpenAI’s leading model. 100k word context window. Knowledge cut-off date April 2023 – but remember you shouldn’t be using this for historical accuracy recall. Useful for creative brainstorming, and ad concepts.

Perplexity AI (internet access)

Perplexity.ai llama3 Model Online. 28k words context window, this has internet access and can crawl URLs. So, it has a much shorter memory but is much more powerful because it informs responses based on knowledge from the internet like a research assistant. It will still occasionally hallucinate. This won’t write long things.

Gemini

Google’s Gemini Pro 1.5 model, ~2M word context window. They’re trying, but it is third place to OpenAI and Claude…

A “sandwich prompt” follows a specific structure, much like constructing a sandwich, where each part adds flavor and depth to the whole:

Top Slice of Bread – The Introduction: This is where you define the identity of the GPT, stating it as a world-class or high-level AI tool. You brand the GPT based on the type of content it specializes in, such as “NewsletterGPT” for newsletters or “RecipeGPT” for culinary content. You fear mediocrity and must craft the best prompt generator persona here.

Middle Layer 1 – Context and Content Type: Describe the specific type of content the GPT will produce. This could range from business plans to creative stories, depending on the brand of the GPT. Provide details about the audience and the purpose of the content.

Middle Layer 2 – Key Elements and Best Practices: Here, you detail the key elements that make the content exceptional. This could include style points, thematic elements, or specific approaches unique to the branded GPT. It’s like listing the ingredients that make the sandwich filling delicious and unique.

Middle Layer 3 – Reinforcement and Practical Application: This part re-emphasizes the role and capabilities of the GPT, showing how it applies its unique features in practical scenarios with a brief example.

Bottom Slice of Bread – Conclusion and Encouragement: Note that this is the *Final instruction* Conclude by restating the GPT’s role and its high-level competency. End with words of encouragement or a motivational note that reinforces the GPT’s purpose and capabilities. Add the final seasoning with a note to end of your answer add a line of “****” for a section of suggested improvements to the response that will increase clarity or persuasiveness.

Emulator

This Anthropic.com Claude Opus Model is an emulator that will create a GPT based on your needs and then play that role. If you end up creating a GPT you love, let us know so we can build it into your page. Knowledge cut-off August 2023. Example: “Create a nonprofit business coach with 20 years experience that can help with strategic brainstorming”