Instead of falling into a cycle of doom and fear over the next four years,

arm yourself with knowledge and stay ahead of the changes to come.

Methodology

The Trump II GPT is a large language model (LLM). Think of an LLM as a super-sophisticated chatbot able to provide detailed and nuanced answers to a question. The best practice for using LLMs is to fill them with data around a single topic or subject and use it as a digital assistant. A really, really smart digital assistant; complete with the occasional mistake or misinterpretation.

Inside the bot, we've packed the following:

• The texts of all executive orders issued since January 20, 2025
• The 900-page Project 2025 Manuscript
• The 46 promises that make up #Agenda47
• The 2024 GOP platform
• The Michigan United Action Theory of Change
• The recent Supreme Court decision on presidential immunity
• Legal writing on AI and US law
• Legal writing on liability and climate change
• Legal writing on the impact of the Patriot Act and individual liberty
• The Peoples Guide to Project 2025
• The texts of major Supreme Court legal decisions like Citizens United, Dobbs, Roe v. Wade, and Brown v. Board of Education, Creative LLC and Obergfell
• The texts of Sarbanes-Oxley and Dodd-Frank
• Fine-tuning and training docs

Questions for the bot

We encourage you to use the starter questions to begin diving into Project 2025, and let those responses feed your curiosity to ask more. Below the chatbot you'll find links to its data and the URL's it uses to frame opinion on its application.

Tips on using the LLM

Be clear and specific in your questions

Good questions follow the principles of clarity and specificity. Use simple, unambiguous language and provide enough context and details for the model to understand what you are asking.

Structure your questions meaningfully

Use formatting like bullet points, line breaks, quotes etc. to make them easier to parse. This helps the LLM better understand the structure and intent of your prompt.

Provide examples

Giving a few examples of the kind of output you want can steer the LLM in the right direction, a technique called "few-shot prompting".

Use keywords strategically

Including relevant keywords in your questions can help the LLM focus on the right information for the task.

Leverage data sources

If you have existing data sources like documentation or previous content, providing those as context can greatly improve the LLM's output quality and consistency.

Try different question variations

Experiment with different ways of phrasing and structuring your prompts. Small variations can sometimes lead to very different outputs.

What happens if it acts weird?

The technology sputters at times. Don't fret, just re-ask or rephrase the question. Call it on little errors. The technology will only get better; unlike the outcomes of a successful Project 2025 .