What about AI?

The next big thing I want to get out here is the iterative nature of AI.  You have probably heard the word in relation to AI before, but have you ever contemplated what it really means?  Ask AI something.  Rephrase the prompt.  Follow up the prompt.  Keep tightening the focus until you are getting the results you are looking for. Here is an example.

What kind of vision insurance should I get?

            This is a pretty generic question but it gave a really good overview of types of insurance and why different types of people would need each.  But what I really wanted to know was what kind I needed.  Me.  Personally.  I added a follow-up.

What kind of insurance would be good for macular degeneration?

In addition to some specific features to look for, it provided this handy table.  Its bottom line was that this person would not need vision insurance at all.  It is a health insurance issue.

It can go deeper.

What kind of insurance would be good for a person with macular degeneration and posterior vitreous detachment?

Here is even more detailed information, including links to studies about various types of medications that can be used in persons with both conditions.  It is still in the realm of health insurance rather than vision insurance.  “These studies suggest insurance policies for managing posterior vitreous detachment and macular degeneration should cover regular specialist visits, advanced imaging like OCT, and ongoing treatments such as anti-VEGF injections, as these are key for monitoring and treating both conditions." 

Keep asking until you are getting the results that you need.  Don't forget you can ask the AI to interrogate you or ask it what would be good prompt.  We will get examples of those next time.

June 10, 2025

Ok, I have settled on Consensus as the library AI.  Pending discovery of something better, this is the tool I am going to use when I create examples for this site.  At the pace of change on the topic, I imagine there will be something better every week.

Today’s real topic is prompting.  This will probably be the topic of several posts.  As I learn more about it, it seems that prompting is the most important part of the AI process.  If you want to get everything you can out of AI, the prompt is what matters.

First, let me recommend a video called Generative AI in a Nutshell.  It is 18 minutes long, but well worth your time.  It has a lot of information, but I think the most important part is the way it explains how the right prompt can make all the difference in your results.

This screenshot from the video helps a lot.  Notice the last ‘good’ prompt.  You can tell AI to ask you clarifying questions.  This is probably the single most important thing to know about AI.  You can ask AI how to frame your questions to get the best results.

AI is an opportunity, not a threat.

I watched a terrific video today in preparation for this post.  That was the most important takeaway.  AI is an opportunity, not a threat.  Its only limit is your imagination and your prompt creation skills.  Both can be improved on with practice.  So how do we get there?  We are going to focus on prompts and techniques later, but first I want to look at tools. How do you know which tool to use?

It has been a long time since I updated this and for good reason.  As the AI world expands, it becomes ever harder to keep up.  When I first created this information page, it was really just ChatGPT.  At the beginning of 2025, there were approximately 70,000 AI companies worldwide.

That is an overwhelming number and it seems like new ones pop up every day.  It feels impossible to pick the one that will be most useful for your needs.  Based on criteria that is useful to both the student and the teacher, I have evaluated a few of the AIs that advertise they are good for student research. 

Many were eliminated without testing because they advertise that they will write an entire paper. Since this is the student’s job, Ranger College students should not use these types of AIs.  Any work submitted would be considered plagiarism.  Another criterion was free or nearly free.  This also eliminates many products. 

Right now, June 3, 2025, there is a pretty good one out there called Consensus.  The free level allows for unlimited searches, so you can get basic research on all your projects.  It also includes ten analyses per month.  Drop in work you have already done and it finds articles and sources that may relate. 

To do a basic search, enter a search parameter and it will provide you with a brief overview, three to four paragraphs, and a list of citations.  All the ones I tested were real sources.  Most will be available through the library’s journal subscriptions and the library can help you find them. 

There are probably good and free tools out there.  If you know of one, let me know.  I will be happy to share it here.

I have decided to take this little blog more public.  Until now it has only been accessible through the library AI discussion page.  The first step is to move all the old entries over, so here goes:

June 6, 2024

I have decided to update this in a blog format so that previous insights remain available, but new ideas can rise to the top.  I typically get my information through the One Useful Thing blog and invite all of you to subscribe.  Ethan writes in a way that even I can understand.  Today’s article brought the state of AI up to date, but as Ethan points out, by publication it was probably already out of date.  This is an incredibly fast-moving target.  I do not hit on nearly everything covered in One Useful Thing.  My goal is to provide you links to the various AIs and ideas for how to use them.  There are quite a few out there now and each has strengths and weaknesses.

Today’s ideas are about things you can do with research.  Input was fed to a series of AIs and asked for a variety of things.  For example:

Music!  Suno wrote a song based on prompts I gave it about oil boomtowns (yes, I am a one-trick pony).  It was easy to figure out, the lyrics were visible, and simple to download.  My oil boom song is here.  Udio also wrote a song.  I never found lyrics and I had to download it instead of getting a link.  Both provided 30 second samples, but there was an option to extend the length.  Both also had an option to reuse the prompt, so if you don’t like what you got, get something completely different.  I encourage you play with both and keep tweaking until you get what you want.  I only tried one on each site.

Illuminate is still in the beta stage.  You can’t submit your own papers, but you can see how it works.  It takes a full length paper and turns it into a conversation.  It has multiple voices, rises and falls in tone and breath, and interactions between the voices.  It makes sense as if you were following an actual conversation.  I am really looking forward to this.  Dusty, dense research can easily be turned into something more interesting.

Now, I will revisit the core uses of AI and the current best tools for that.  Summarize, generate ideas, write outlines, produce reports, create strategic plans, and so much more.  As always, I used the prompt “explain the 1917 oil boom in Ranger, Texas.  Include citations.”  The full text of each response is attached to the title.

Claude 3 Sonnet:  Not as accurate as Gemini.  There were only two citations, but both were real books well suited to accurate research.

Gemini 1.5:  For what it’s worth, you cannot log onto this site through Ranger College.  I have to open an incognito window and log in with my personal email address.  I do use it frequently at work, however.  The information today was generally accurate.  The citations were real, but shallow.  It even used Wikipedia as a source.

Chat GPT:  This does not seem much improved over the model we started this process with two years ago.  Neither citation, provided by hyperlink, worked.  The TSHA article was easy enough to find by searching the TSHA site once you were there, but the Legends of America site had no article on Ranger at all.

December 7, 2023

LLMs change so fast that it is hard to keep up.  I subscribe to a blog called One Useful Thing to help me.  This morning (December 7, 2023) he published a blog called An Opinionated Guide to Which AI to Use: ChatGPT Anniversary Edition.  Surprisingly, it is one year since ChatGPT took the world by storm.  The blog discusses what has changed and what is coming.  I have been attempting to keep up, but was shocked by what has changed.  The bottom tab in this Guide is called Other AI Programs.  Mollick only mentions two of the ones I had listed.  I will have to investigate what is still alive and what has failed. 

The most important takeaway from the article, which everyone interested in AIs should read, is the huge quality difference between ChatGPT 3.5 (free version) and ChatGPT 4 (paid version).  Bonus – he explains a few ways to access ChatGPT 4 for free.  Every one of his blogs provides insight into great ways to use AIs to increase productivity, but using ChatGPT 4 seems to be the most important aspect. 

The most important takeaway from this post is that this page, and every other page you look at, is probably already out of date.  That said, read on – there is a lot to learn.