J Rice

In a previous post, I demonstrated using ChatGPT to estimate a software development project integrating an existing Django project with the Jira API. And it gave some very reasonable answers. Today, I am prompting ChatGPT to estimate the same project in a very meta way. Now the estimate is to

In a previous post, I poked fun at ChatGPT for its confidently incorrect understanding of an abbreviation for a single word. That example was ChatGPT-3.5. Now with ChatGPT-4 the understanding has improved. It explains its reasoning and now gives correct answers if you ignore the Latin. Link to the

Continuing some table-turning with ChatGPT...it seems cooler if I call it table-flipping, though. (╯°□°)╯︵ ┻━┻ I decided to have OpenAI's ChatGPT-4 ask me the questions, and the result is impressive. Even compared to ChatGPT-3.5. I had it ask me a series of questions to estimate the time it

Have you wanted to turn the tables on ChatGPT and it asks you the questions? Well, now is your chance. Just check out my following example. Given the correct prompt design, you can get ChatGPT to ask you a series of questions it thinks needs answered to derive the answer

The following is a fun demonstration of the many unexpected capabilities of ChatGPT. This is a classic problem, especially with children, of very few ingredients to make dinner and needing more time and energy for a last-minute grocery trip. In this example, I give it a contrived list of random

In a previous post, I demonstrated how a simple prompt in ChatGPT could be a great way to kick-start and facilitate orthogonal thinking. Here is another fun example of applying the cooperative nature of honeybees to the overall strategy for an online retailer. Original Post on LinkedIn: John Rice on

(Subtitle: Napolean goes holiday shopping) In continuing to demonstrate some exciting abilities of ChatGPT I asked it to apply the strategy of one thing to something completely unrelated. This ability is impressive since this feature can facilitate orthogonal and cross-domain thinking to acquire new insights. Original Post on LinkedIn: John

If you have seen some of my prior posts you have seen me poke fun at ChatGPT, but now I will show an example of how powerful ChatGPT is at inferring, and acquiring context, from a potentially ambiguous dataset. If you have a background in equities data, you will recognize

Another funny example of ChatGPT blissfully unaware yet oddly aware of its limitations. When I stumbled into this, I was asking ChatGPT for an example of how to generate synthetic time series data. Original Post On LinkedIn: John Rice on LinkedIn: Another funny example of ChatGPT blissfully unaware yet oddly

This is a follow up post to this previous post: ChatGPT3.5 The Probability Calculator (tinselai.com) This demonstrates an amazing aspect of ChatGPT. After the post linked below where it shamelessly gave incorrect answers I gave it some training via few shot training. The difference in minor training via