Tag: ChatGPT

I love and hate science.

I love that science continues to try to understand the world at large, and I hate that it is always back pedaling.

I asked ChatGPT how science gets away with this, and to offer examples of when it works, and when it doesn’t.

Here’s the report…

Science: Correcting Itself One Study at a Time

Science is an ever-evolving field. Through rigorous research and experimentation, scientists uncover new insights about the world. However, not every study stands the test of time.

Some findings are later proven wrong, while others continue to hold up under scrutiny.

In this post, we’ll look at five famous scientific studies that turned out to be incorrect, balanced by five that, so far, remain reliable.

5 Studies That Turned Out to Be Incorrect

1. The “Fat is Bad” Study (1950s–1970s)

For decades, fat was labeled as the ultimate dietary villain. A series of studies, led by researcher Ancel Keys, suggested that dietary fat, particularly saturated fat, caused heart disease. This led to a low-fat craze that dominated the food industry. However, recent research has shown that the relationship between fat intake and heart disease is more complex. Healthy fats, like those found in avocados and nuts, are actually beneficial, and sugar and processed carbs may be more harmful than previously thought.

2. Stanford Prison Experiment (1971)

Psychologist Philip Zimbardo’s famous Stanford Prison Experiment suggested that people would quickly conform to abusive behaviors when given authority. The study has been widely criticized for ethical issues, poor methodology, and researcher bias. Participants were often encouraged to act out extreme behaviors, and the results have been difficult to replicate. Today, it’s considered an example of flawed research.

3. Cold Fusion (1989)

In 1989, scientists Stanley Pons and Martin Fleischmann claimed to have achieved cold fusion, a process that would produce unlimited, clean energy at room temperature. The announcement was met with excitement, but soon other researchers failed to replicate the findings. Cold fusion has since been debunked, and it remains a controversial topic.

4. Vaccines Cause Autism (1998)

One of the most damaging incorrect studies was the 1998 paper by Andrew Wakefield that falsely linked the MMR vaccine (measles, mumps, and rubella) to autism. The study was later retracted due to unethical practices and poor methodology, and numerous studies since have found no connection between vaccines and autism. However, the initial report sparked a dangerous anti-vaccine movement.

5. Polygraph Tests as Lie Detectors (1920s–Present)

The polygraph, or lie detector test, has been used for decades to determine whether someone is telling the truth based on physiological responses like heart rate and sweating. However, numerous studies have shown that polygraph tests are unreliable. Stress, fear, and anxiety can produce false positives, and some people can lie without showing physical signs. As a result, polygraphs are not admissible in many courts.

5 Studies That Have Stood the Test of Time (For Now)

1. Theory of General Relativity (1915)

Albert Einstein’s theory of general relativity fundamentally changed our understanding of gravity. It predicts how massive objects like planets and stars warp space-time. This theory has passed many tests over the past century, including the recent detection of gravitational waves in 2015, further proving its accuracy.

2. Double Helix Structure of DNA (1953)

The discovery of the double helix structure of DNA by James Watson and Francis Crick remains one of the most significant breakthroughs in biology. Since its discovery, the structure of DNA has been the foundation for numerous advances in genetics, medicine, and forensic science, and it continues to guide research in the field.

3. Evolution by Natural Selection (1859)

Charles Darwin’s theory of evolution by natural selection explains how species change over time due to environmental pressures. Over 150 years later, the theory remains one of the cornerstones of modern biology. It has been supported by countless studies in fields ranging from paleontology to genetics.

4. Germ Theory of Disease (19th Century)

Proposed by Louis Pasteur and Robert Koch, the germ theory of disease revolutionized medicine. It states that microorganisms are responsible for many diseases. This theory has led to the development of vaccines, antibiotics, and modern hygiene practices, saving countless lives. To this day, it remains a fundamental principle of medical science.

5. Plate Tectonics (1960s)

The theory of plate tectonics explains how the Earth’s outer shell is divided into several plates that move over the mantle, causing earthquakes, volcanic activity, and the formation of mountains. This theory has been supported by evidence from seismology, geology, and oceanography, and it remains the best explanation for many geological processes.

Conclusion

Science is a self-correcting process.

While some studies that were once accepted have been disproven, others have withstood decades of scrutiny and continue to be the foundation of our understanding.

What’s important is that science remains open to new evidence and willing to adapt. Through this process, our knowledge grows stronger and more accurate over time.

All of the above was from ChatGPT.

As I hope you know by now, ChatGPT is known to fabricate case studies, so take the above with a grain of salt.

Again, I love science, but I’m also aware that whatever it reports as fact today can be negated as incorrect tomorrow.

Expect Miracles.

Ao Akua,

Dr Joe

As you know – if you’ve been following this blog – I’ve been testing ChatGPT. I’m not overly impressed yet but it did help me create a cool e-book that I put on Amazon.

Basically I had the AI program write a dialogue between Seneca and Epicurus, two philosophers who never met in real life, in response to my sincere questions.

I thought the results were impressive. They were surprising, in-depth, useful, and seemed authentic.

You can find the e-book on Amazon at https://www.amazon.com/Pursuit-Tranquility-Secret-Dialogues-Epicurus-ebook/dp/B0C41253F8/

Enjoy.

Ao Akua,

Joe

PS – I wrote this blog post, not AI.

Thoughtline was an early “ChatGPT” AI program back in the 1990s.

It was based on the DOS system of the time and came on a disk.

I loved it.

I wrote about it in my book, Hypnotic Writing, because it was the first product I sold as a copywriter and direct marketer that actually made money for me.

I used it to help me think through articles, book chapters and even outlines for speaking gigs.

It was fun, stimulating, practical, and surreal.

It often felt almost human, like a real person was behind the system.

Of course, it also often felt clumsy, limited, and somewhat predictable.

But that was in the last century.

What about today?

I wanted to know how the current AI (Artificial Intelligence) programs are working to help us in 2023.

I wrote about ChatGPT last month.

I wasn’t impressed.

But I tested ChatGPT again.

I asked for a book recommendation.

I love reading about publicity stunts and outrageous publicists, especially those who came out of old Hollywood.

I have a collection of books by and about these flamboyant people, and of course I wrote a book on one of them (P.T. Barnum), but I wondered if there was something I might have missed.

So I asked ChatGPT for suggestions.

ChatGPT (3.5) surprised me.

It gave me several titles including, “Flack: True Tales from the Hollywood Trenches” by Dylan Landis.

I never heard of it.

I got excited.

I wanted to read it.

I looked on Amazon.

Nothing.

I looked on eBay.

Nothing.

I searched my favorite rare book sites.

Nothing.

I prompted for more info.

ChatGPT stood its ground: “The book was published in 2018.” 

I took that added information and dug deeper.

Still nothing.

I’m a book collector.

I can find rare books.

Why couldn’t I find this one?

I prompted ChatGPT yet again.

It apologized.

“I apologize for the confusion earlier. Upon double checking, I realized that I made an error in my previous response.”

The book doesn’t exist.

It doesn’t exist!?

Why would the system offer a book that doesn’t exist?

Use ChatGPT for fun but beware of using it for facts.

I posted an abbreviated version of this blog on my FB profile. One person said they were using ChatGPT to research legal cases and discovered many of the cases weren’t real.

Enjoy ChatGPT but don’t trust it.

For the record, when the fourth version of ChatGPT showed up last month, I tested it, too.

I was told it was far more accurate.

I was told to be clear in my prompts so the AI system could be clear in its answers.

I Tested 4.0

I asked for a few nonfiction books about famous songs and how they were written, in particular the classic song, “Smile.”

I knew Charlie Chaplin wrote the melody for a film of his called Modern Times. Decades later the lyrics were added by John Turner and Geoffrey Parsons.

The result was the classic song everyone knows and the legends sing. (Even I recorded it on one of my singer-songwriter albums.)

But I wanted to know more.

Was there a history to the song?

Was there a book about it?

So I asked the wizard for a suggestion.

ChatGPT4 replied –

“Smile: The Story of Charlie Chaplin’s Lost Classic” by David Gritten.

I went bananas.

THAT’S the book I wanted to read!

It sounded perfect.

So I went on my book search and – you guessed it – it doesn’t exist.

ChaptGPT4 made it up.

When I pushed back, the system again said –

“I apologize for the confusion. It appears there isn’t a book titled that.”

SIGH.

Again, don’t trust ChaptGPT, even the latest version is prone to writing fiction.

And since it seems like it wants to write creative material, I thought I would test it on writing a song.

So I went to ChatGPT-4 and prompted with this:

What do you think?

Is it good?

Well, it’s a song.

It’s not a great song.

It needs tweaked and polished.

But it’s actually a workable song, or at least the beginning of one.

So ChaptGPT wins points.

But.

Don’t turn off your brain.

Don’t surrender your awareness and discernment.

Don’t drop your common sense.

The system is probably great for brainstorming projects, but if you allow it to do your thinking, creating and writing, you may find yourself with passive writing, mediocre products, inaccurate info, and potentially in court.

Have fun with it.

After all, not having a book about the song Smile means I could potentially write one.

And the beginning of a song about our dog Cabi gives me something to work with.

So I can thank the system for those ideas.

But let’s test the system one more time.

Some people claim ChatGPT can write sales letters, ads, and other marketing materials.

Well, can it?

I again went to ChatGPT-4 and made a clear request:

What do you think?

Is that a good sales letter?

Are you sure?

One of the problems with using ChatGPT is knowing whether what it writes for you is actually any good or not.

How would you know if you don’t have the experience to recognize excellence or mediocrity?

Again, what do you think of the sales letter?

Here’s what I think:

It’s a very good letter.

It has all the key elements of a persuasive letter.

But it’s not a great letter.

It has lies, half truths, and misleading content.

There aren’t any step by step guidelines in the book.

Those testimonials in the letter are all fictional.

Like the song for Cabi, the letter is a good start. You can take it and work with it. It might save you time and trouble by giving you material to review rather than a blank page to stare at.

But don’t give it your mind.

Let’s do one more test.

I created a mental time travel program last year that helps people travel to the past to sit at the feet of some legend of long ago.

They can then ask questions, gain insights, and more.

I wondered what Seneca and Epicurus – representing two rival schools of thought – would discuss if they ever met.

Being a fictional scene, and AI seems to shine writing fiction, I thought this might yield interesting results.

So I went to ChatGPT-4 and prompted:

Well, what do you think?

I’ve studied both Seneca and Epicurus.

I’ve collected and read their ancient writings.

I find the fictional dialogue above illuminating and accurate.

But since I’m not an expert on Stoicism or Epicurism, I can’t say if the dialogue is 100% accurate or not.

Still, it’s a good beginning to a possible book, play, or movie.

If I wanted to, I could develop the dialogue (with AI’s help and my discernment) into a product.

With that in mind, I give ChaptGPT more points.

It impressed me and gave me something I didn’t have: a good start to my idea.

I’d be wary of just taking what it writes and running with it. It could be considered plagirism; after all, the program is pulling from existing content and allegedly creating new content from it.

But is it?

Turns out plagiarism is a potential issue with anything ChaptGPT “creates.”

There’s a program online you can use to check content.

You can check it out here.

Bottom line for me:

Use ChaptGPT for fun, as a kick starting device, for first drafts, but do not turn over your ability to reason, rewrite, or reject what it gives to you.

You’re smarter than it.

Right?

(Note: I wrote this post not ChatGPT.)

Expect Miracles.

Ao Akua,

Dr Joe Vitale

I didn’t write the bulk of this blog post.

I asked ChatGPT to tell me how to attract money.

Before I show you what it “wrote,” let ChatGPT tell you what it is:

ChatGPT is a computer program that can talk to people like a human. It can understand what you say and respond to you in a helpful way. It has been trained on lots of information from the internet and can answer questions, provide information, and even make small talk. It’s like having a really smart friend you can talk to anytime!

I’m not so sure I agree with that, but since ChatGPT is the raging trend right now, let’s play along.

I asked it to tell me how to attract money.

Here’s what it wrote: