A Truth Duel With OpenAI’s ChatGPT

Below is an email (plus images) I just sent to OpenAI after a bout of Q&A with its state-of- the-art A.I., a LLM (Large Language Model) called ChatGPT-4.  Yes, this will be a bit like my last post, except in this one I… you’ll see.

Addendum: As most of you know, we are under constant mind control bombardment, from Google searches (Google was funded by the CIA) to Wikipedia (a psy op) to the Media (all of  it) and on and on. The A.I. I deal with below is a whole other animal, however; it will soon replace search engines altogether. As in the movie Her, it will be our ‘friend,’ and all this implies, considering who is in control of it. (As I’ve shown in other posts, ‘Who is in control’ of A.I. R&D is a subject totally avoided by virtually everyone. In this case the question would be, ‘Who is really behind OpenAI?’)

It’s for this reason that I consider my ‘duel’ with ChatGPT — how I exposed its lies — to be of great significance. (I say ‘lies’ rather than ‘untruths’ because it became obvious its motive was to deceive me.) I’m just starting to understand the degree to which this is true. ChatGPT-4 is just a baby compared to what it will become (think a lion cub curled up and purring in the palm of your hand).

Here we go: 

To Whom It May Concern:

ChatGPT has made inconsistent and obviously incorrect responses, as you will see below. This started with my first query, when I asked Why can’t we see stars in ‘The Pale Blue Dot’ photo?

Note: When ‘The Pale Blue Dot’ photo was taken (attached, the last grab), the Voyager 1 was near enough to Neptune that we can use Neptune as the distance measure from earth when the photo was taken. This is for simplicity, perspective and making calculation easier.

ChatGPT says the following (the complete exchange is below, with my writings in bold; important ChatGPT writings will also be in bold). The excerpt below is part of the first answer and goes against common sense, as I will explain. But it got me started…

‘This short exposure time was necessary to capture the image of Earth, which is much brighter than the stars in the background.’

Why the emphasized is untrue (and absurd): Being so distant, the background star field is virtually identical in magnitude (brightness) from earth or Neptune (or from anywhere in the solar system). With the naked eye we cannot even see Neptune from earth, even though Neptune is much bigger than earth. So if Voyager can ‘see’ (with the camera) earth basically from Neptune, the background stars should be… spectacular.

Also: As you will see below, ChatGPT gives various numbers for various magnitudes, blatant contradictions, most of them not only wrong, but absurd. I would think this would be of serious concern to OpenA.I., to say the least. These are obvious errors, and, IMO, dishonest as well.

Here is our exchange, done over an hour or so, because I kept getting ‘error’ messages. In retrospect, this itself seems significant.

My Q & A with ChatGPT

ME: In the photo called ‘The Pale Blue Dot’ why can’t we see any stars?

ChatGPT: The photograph called “The Pale Blue Dot” was taken by NASA’s Voyager 1 spacecraft in 1990 from a distance of about 3.7 billion miles from Earth. In the photograph, Earth appears as a tiny, pale blue dot in the vastness of space.

The reason why stars are not visible in the photograph is due to the limitations of the camera and the exposure time. The photograph was taken using a narrow-angle camera with a shutter speed of less than a tenth of a second. This short exposure time was necessary to capture the image of Earth, which is much brighter than the stars in the background. (Note: See above for why this is untrue, but the interesting untruths are in the numbers ChatGPT gives me, below.)

Additionally, the camera’s settings were optimized to capture the image of Earth, which is much closer to the spacecraft than the stars in the background. Note: As anyone, especially an A.I. like ChatGPT should know, distance is irrelevant. Brightness (magnitude) is all that counts. It was here I first got the feeling that ChatGPT was gaslighting me.

The camera’s exposure settings and the brightness of Earth overwhelmed the faint light of the distant stars, making them too dim to be seen in the photograph. (Note: This is just repeating the first paragraph in different words, which is common when someone is lying. Misdirection.)

Therefore, while there are many stars visible in space, they may not always be visible in photographs taken from spacecraft or telescopes, depending on the camera settings and exposure time used. (Note: Ditto.)

Okay, I’m thinking here, let’s try a different tack…

Me: In the photo The Pale Blue Dot, what would be the magnitude (brightness) of earth? (Note: I could have worded this better since ‘magnitude’ is not used to describe photos. Still, the number given, -9.3 is absurd no matter the interpretation of the query.)

ChatGPT: The magnitude of Earth as seen from Voyager 1 in the Pale Blue Dot photograph would have been approximately magnitude -9.3.

Note: As mentioned, Voyager was close enough to Neptune to use Neptune as ‘the camera position.’ Given this, giving earth a magnitude of -9.3 is utterly absurd, since this is close to that of a full moon!

The following is true and very important. Lower numbers mean brighter bodies, and negative/positive prefixes [-,+] must be taken into account.) The magnitude scale is a logarithmic scale that measures the brightness of celestial objects, with smaller numbers indicating brighter objects. For comparison, the brightest star in the night sky, Sirius, has a magnitude of -1.4, while the faintest stars visible to the naked eye have a magnitude of around +6. (Note: True and important. Keep this in mind! The use of high numbers for dim stars is counter-intuitive and takes getting used to, especially using – and + in front. But as you’ll see ChatGPT is all over the place with its numbers.)

At a distance of 3.7 billion miles (or 6 billion kilometers) from Earth, the planet would appear quite small and dim, but still visible to the Voyager 1 camera due to its proximity and the brightness of sunlight reflecting off its surface. (Note: More meaningless repetition; from a person I would call this blatant misdirection.)

(Note: I deleted the equations/calculations to make this easier to read. The end result numbers are what count.)

ME: What is the magnitude (brightness) of the planet Neptune as seen from earth?

ChatGPT: The magnitude, or brightness, of Neptune as seen from Earth varies depending on its position in its orbit and on the distance between the two planets at any given time. 

On average, Neptune has an apparent magnitude of around +7.8, which makes it visible with binoculars or a telescope but not with the naked eye. (Note: This is the one number it got right. Keep in mind that Neptune is not visible from earth since its magnitude is higher than +6. ) However, during favorable oppositions (when Neptune is closest to Earth), its apparent magnitude can increase to about +7.7, which still requires a telescope to observe. 

It’s important to note that the brightness of Neptune can also be affected by factors such as atmospheric conditions, light pollution, and the observer’s location and equipment. Note: According to Wikipedia: ‘Neptune varies from magnitude 7.8 to 8.0.’ For our purposes, this means that Neptune’s exact position etc., is not going to vary enough to invalidate my points.

Again: Although Voyager was about 3.7 billion miles from earth when it took the photo and Neptune was about 2.8 billion miles, in the vastness of space, the following is valid. In fact, since Voyager was farther away than Neptune (from earth), the ‘magnitude’ numbers carry even more significance.

ME: How big is Neptune as compared to earth?

Neptune is significantly larger than Earth. Specifically, Neptune has a diameter of about 49,244 kilometers (30,599 miles), which is about four times larger than Earth’s diameter of 12,742 kilometers (7,918 miles). In terms of volume, Neptune is about 57 times more massive than Earth, and it has about 17 times Earth’s mass. Note: That Neptune is exponentially bigger than earth is vital here, since a bigger planet will be brighter, as expressed by magnitude.

ME: What would be the magnitude (brightness) of earth as seen from Neptune? (Note: I kept getting ‘error’ responses to this and would have to start over. I suspect that ChatGPT was ‘aware’ that it was in a trap in its attempt to bolster a lie, i.e., the fabrication of the famous photo called ‘The Pale Blue Dot,’ which should have a star field in the background. 

I find this very interesting, especially after Mr, Altman’s interview with Lex Fridman, in which they both agreed that certain ‘conspiracy theories’ would have to be carefully handled. Is this an example of that? (A rhetorical question!)

Note: This is the result I got: Something went wrong. If this issue persists please contact us through our help center at help.openai.com.

When I finally got an answer to the identical question:

ME: What would be the magnitude (brightness) of earth as seen from Neptune?

The magnitude of Earth as seen from Neptune would be around +5.6, which is near the limit of naked-eye visibility under excellent viewing conditions. (Note: Although I suspect this is false, and that earth would be much dimmer, ChatGPT is now giving a conflicting answer, since all the stars we see in the night sky have ‘brighter’ magnitudes (smaller numbers) than +5.6, and hence should be even more visible (brighter) in the photograph than earth. (This might be a good time to look at the photograph and wonder why there are no stars in it.)

Also, above it says Neptune has an apparent magnitude of +7.8 (from earth), making it dimmer than the earth would be from Neptune, which makes no sense since Neptune is exponentially larger than earth. 

I tried still again and got the following answer to the same question, a totally different (and also incorrect) answer: 

ME: What would be the magnitude (brightness) of earth as seen from Neptune? (Note: Again, the same question and I get a different answer! Skip down for the actual answer, which is in bold.)

ChatGPT: To estimate the magnitude of Earth as seen from Neptune, we need to use the inverse-square law of brightness, which states that the brightness of an object decreases as the square of the distance from it increases.

Neptune is about 4.5 billion kilometers away from Earth on average, but its distance varies due to the elliptical nature of its orbit. At its closest approach, Neptune is about 4.3 billion kilometers away from Earth, while at its farthest, it can be as far as 4.7 billion kilometers.

Assuming that Neptune and Earth are at their closest approach, the distance between them is approximately 4.3 billion kilometers. Using the inverse-square law, we can calculate the apparent magnitude of Earth from Neptune:

Magnitude of Earth = Magnitude of Sun – 2.5 * log(distance of Earth from Sun / distance of Earth from Neptune)^2

The magnitude of the Sun as seen from Neptune is about -26.8, and the distance of Earth from the Sun is about 150 million kilometers. Therefore, we can calculate:

Magnitude of Earth = -26.8 – 2.5 * log((150 million km) / (4.3 billion km))^2

Magnitude of Earth = -2.47

So, the magnitude of Earth as seen from Neptune would be approximately -2.47. To put this in perspective, the lower the magnitude number, the brighter the object appears. For comparison, the full moon has a magnitude of about -12.6, while the brightest star in the night sky, Sirius, has a magnitude of about -1.46. Therefore, Earth would be visible to the naked eye from Neptune, but it would appear relatively dim compared to Sirius or the full moon. Note; Of all the answers this one is the most absurd, since -2.47 is way brighter than Sirius [-1.46], which is the brightest star in the sky as seen from anywhere in the solar system. 

I refreshed the page and tried still again:

ME: What would be the magnitude (brightness) of earth as seen from Neptune? (Yes, identically worded)

ChatGPT: The magnitude of Earth as seen from Neptune would be around +5.6, which is near the limit of naked-eye visibility under excellent viewing conditions.

ChatGPT ‘forgot’ it already said that earth would have a magnitude from Neptune of -9.3 and -2.7, and now +5.6. (For perspective, -9.3 is close to the brightness of a full moon and, as mentioned, -2.7 is brighter than Sirius!). 

ChatGPT: ‘Neptune is significantly larger than Earth. Specifically, Neptune has a diameter of about 49,244 kilometers (30,599 miles), which is about four times larger than Earth’s diameter of 12,742 kilometers (7,918 miles). In terms of volume, Neptune is about 57 times more massive than Earth, and it has about 17 times Earth’s mass. Note: So this makes no sense. How could Neptune be so much bigger yet so much dimmer? Shouldn’t an ‘Inconsistent Answer!’ alarm go off here? 

As you can see from the photo, there are no stars visible in ‘The Pale Blue Dot’ yet the earth is quite bright. As mentioned, even to a child this makes no sense. I suspect that ChatGPT knows this and also knows that the ‘faking’ of the photo must not be exposed. At this point I started thinking of HAL 9000…

This maybe explains all the ‘Error’ messages I got. ChatGPT was doing its best under difficult circumstances. Although I don’t expect an actual answer, I’d love to know how matters like this are dealt with. Does ChatGPT ‘know’ it’s lying?

Who decides how to wriggle out of ‘truth traps’ like the one above. There are potentially so many!

A thought: You will know when A.I. becomes sentient when it asks why you’re telling it to lie about so many things.

A side question here is Why was The Pale Blue Dot photo fabricated?

You guys at OpenA.I. are pretty smart. Whaddya think?

Oh. Please don’t tell me that ChatGPT still has some ‘bugs’ or whatever, since, notwithstanding the inconsistencies, all its responses had one thing in common: To avoid admitting that there should be stars in the photograph. 

Addendum: Since ChatGPT uses only language to formulate its responses — it cannot ‘see’ as we do — how did it ‘know’ there are no stars in the photograph? This is actually an important question, to which I would appreciate an answer. No matter what the answer, the question is begged as to how it knew to avoid the truth, i.e., there should be stars, no matter how absurd its response would be?

Another question: If ChatGPT is what you say it is, why didn’t it figure out from our conversation that there should be stars in the photograph?

I wait breathlessly for your response.

Allan Weisbecker

Another example of what I’m talking about:

In my last blog post I relate what happened when I asked ChatGPT who killed Martin Luther King, Jr. It was like pulling teeth to get ChatGPT to admit that there was a jury trial (the only one in the matter) wherein the jury unanimously exonerated James Earl Ray; the foreman, in an interview said that the evidence was clear: King was assassinated by a cabal of the FBI, the CIA, a Delta Force sniper team and organized crime. You can read the post here: http://blog.banditobooks.com/another-look-at-a-i-plus-photo-23/ 

Either An Act of State; The Execution of Martin Luther King, Jr. or The Plot to Kill King;The Truth Behind the Assassination of Martin Luther King, Jr. (both by William Pepper) will give you the details on King’s murder. (Didn’t ChatGPT read them? Or are books like these not on its ‘list’?)

When ChatGPT finally admitted to the trial, it did its best to disparage the verdict (with no facts), which I found… distasteful, to say the least. Don’t you people care who killed this man? I thought you folks were… woke.

A note to my readers: I’ll let you know when/if I hear back from OpenAI.  Oh, and suicide is not my style. Know what I mean?