J&P
Sverige / USA
J&P

ChatGPT – Introduction

A page intended for the absolute beginner who has never tried this exiting, powerful, and potentially very damaging tool.

WARNINGS

  1. NEVER enter any personal, personally identifiable information (PII) or business sensitive information!
    • All data you enter becomes part of the behind-the-scenes system, used to further extend the functionality (how AI tools are “learning”, evolving positively or negatively, is all by data entered into the system, including everything you throw at it. At least assume all this until the opposite has been proven (like if you set up and train your own AI system locally on your own computer).
  2. NEVER trust responses without both your own sanity checks and follow-up questions as applicable.
    • This is why I J continuously emphasis that ‘education [of kids, people in general] is the critical path forward to our society not going under’, and ‘not only related to AI, but climate, politics, fake news, and everything’. ,
    • ALWAYS ask yourself ‘Does this answer seem plausible?’
      • Answers CAN be (more towards) correct but also 180 degree wrong. (Just think of a small ‘not’/other inverting word completely changes all meaning of an answer.)
      • And even if looks plausible and true, MAKE SURE TO verify important details with other sources.
  3. Keep in mind ‘the devil is the the details’ and think carefully around more complex questions, and probably have more of a discussion around more complex subject matters. Ask follow-up questions, argue, question responses, try get references.

While on Warnings – some cautionary tales:

Get Started with ChatGPT

  1. Go to https://chatgpt.com/
  2. And type in a question, phrase, you want to answer or discussion around.
  3. That’s It. As a start, for trying out things. See more in Account and Plans page.

I.e. one doesn’t have to (well – ‘you might be able’..’in guest mode’) create an account to get started but even the Free plan does provide benefits.

Good to Know about AI, AI Tools

AI Solutions, Tools

“AI, or artificial intelligence, is the simulation of human intelligence in machines designed to perform tasks such as learning, reasoning, problem-solving, and understanding language.” [from asking ChatGPT: ‘explain ai in one sentence’]

Training

Just like a human, an AI-solution, like the human brain, must learn something befire it can make any reasoning e tc.

  1. Tools are trained on data, lots of data, LOTS. Which take also lots of time and compute resources.
    1. Quality of responses depend on quality of data in, so ‘Lots of Shit In => Lots of Shit Out’
  2. As training / learning takes time (think more in terms of months than weeks or days or more) so be aware that tools are better at older topic and can be more toward worthless on latest, most current, topics.From 2023-11-07 zdnet.com/article/chatgpt-is-no-longer-as-clueless-about-recent-events/
    “For GPT-3.5, the cutoff date has been increased to January 2022 from September 2021.  And the cutoff date for the new version of GPT-4 dubbed GPT-4 Turbo is April 2023. These new timeframes mean you won’t bump into as many unanswered questions and apologies when asking the chatbot about the recent past. ChatGPT’s responses about relatively recent events should also be more accurate and precise.”I.e.

    1. You will be more successful with good answers on “older”, more mature topics, like food recipes, older programming languages with syntax, programming logic, et c.
    2. To less successful to get good, or any useful what so ever, answers of what happens today or yesterday.

More on Cutoff Dates

Search latest using (e.g.) google.com/search?q=chatgpt+models+cutoff+dates. From one source – A comprehensive list of Large Language Model knowledge cut off dates – llmo.ai/articles/list-of-large-language-model-cut-off-dates, last updated May 5, 2025, per Oct 6, 2025 when retrieved at time of updating this section:

ChatGPT’s knowledge cutoff date has changed over time:

  • September 2021: When ChatGPT was first released in November 2022, the chatbot’s large language model (LLM) was trained on data up to September 2021. This meant that it couldn’t access more recent information through internet searches without the use of plugins.
  • GPT-3.5: January 2022
  • GPT-4: April 2023: OpenAI CEO Sam Altman announced that ChatGPT would be trained on information up to April 2023.
  • GPT-40: October 2023
  • GPT-4.1, GPT…. (many variants per llmo.ai/articles/list-of-large-language-model-cut-off-dates
  • GPT-5: October 01, 2024 – public release August 7, 2025, some 10 months after cutoff-date.
    ChatGPT is an AI language model that was trained on a large amount of text from a variety of sources, including books, Wikipedia, news articles, and scientific journals.