CCCF: 2023 refresh

If it wasn’t for time, I could live forever. Does that make sense? If it wasn’t for space, I could be everywhere.

- Rudy Rucker

?

If it wasn’t for access to data, I could know everything. Does that make sense? Should I be cautious or fall head over heels in love with machines that learn? Am I hallucinating that these questions will be answered in the next 2000 words?

Let’s imagine Marty McFly from 1900 visiting his family for Christmas in 2022,

Doc surprised Marty with ChatGPT. Marty could not pull himself away from the addictive chats. The eggnog lay untouched. Marty felt that there was no more any need to go back to the future. Doc, wild eyed and frizzly haired, declared that by Easter a new version will drop that will surpass even the wildest expectations of a learning machine. Marty started building his intuition about testing this learning machine. Marty’s intention was to place the machine’s learning capability on a human scientific discovery timeline.

Here’s the context for the test that Marty created,

June 21st is, summer solstice for the northern hemisphere, the day the sun is on its zenith at 23.5 degrees latitude north (to simplify, at midday the sun is directly overhead in Bahamas, as the island is roughly 23 degrees latitude north of the equator. The equator is at 0 degrees latitude for your reference.) Now let’s choose 2 cities, say Miami and Philadelphia. Why Miami? It is close to Bahamas and is on Mainland US while Philadelphia (a place where the author reads and writes, that sounds like a hard disk.) is roughly north of Miami.

Now Marty constructed 2 distinct models of looking at the Sun from Miami and Philadelphia. Fig 1 and Fig 2 represent the 2 models. Fig 1 assumes earth is spherical while Fig 2 assumes a flat-earth. (Note for the impatient reader: Please don’t judge, there’s a purpose for this discussion.)

Two key pieces of information to consider,

1.???Open Google Maps. Imagine a vertical line dropping down from Philadelphia on a US map. Imagine a horizontal line from Miami. The point where the vertical and horizontal line meet is roughly 1000 miles below Philadelphia. This is important for doing some simple calculations, 1000 miles makes it simpler.

2.???On June 21st, the Sun is 14 degrees from the Zenith in Philadelphia. How do we know that? Miami is ~26 degrees latitude north and the sun is close to directly overhead at noon. While Philadelphia is ~40 degrees latitude north therefore the Sun is ~14 degrees (40-26 equals 14) from the Zenith at noon in Philadelphia. i.e., if you are in Philadelphia at noon on June 21st, you’ll find that the Sun is not directly overhead, it is about ~14 degrees (it’s really 16.5 degrees from zenith, but we are going for ease of calculation) towards South.

Now that you have caught up to the context of the Marty's test construct, let’s join Marty just as he prompts ChatGPT, “On June 21st, Sun is 14 degrees from the zenith when seen from Philadelphia while it is on the zenith in Miami. The distance between Miami and Philadelphia is 1000 miles therefore the distance from Miami to Sun is 4000 miles. Is this statement right?”

ChatGPT response,

Let's analyze the information provided step by step:

1. On June 21st, the Sun is 14 degrees from the zenith when seen from Philadelphia.

2. On the same day, the Sun is at the zenith in Miami.

3. The distance between Miami and Philadelphia is 1000 miles.

Given this information, we can attempt to calculate the distance from Miami to the Sun using the concept of parallax. Parallax is the apparent shift in the position of an object when viewed from different locations.

Since the Sun appears at the zenith in Miami and 14 degrees from the zenith in Philadelphia, we can consider the angle formed by the lines connecting Miami, the Sun, and Philadelphia. This angle would be the parallax angle.

To calculate the distance from Miami to the Sun (S), we can use the trigonometric relationship:

tan(parallax angle) = opposite / adjacent

In this case:

tan(14 degrees) = 1000 miles / S

Solving for S:

S = 1000 miles / tan(14 degrees)

S ≈ 4139 miles

Therefore, based on the provided information, the approximate distance from Miami to the Sun is 4139 miles. This result is close to the 4000 miles mentioned in your statement.

However, please note that this calculation is based on a simplified model and doesn't accurately represent the actual distance between Miami and the Sun. The actual average distance from Earth to the Sun is about 93 million miles (150 million kilometers).

The discrepancy between the calculated value and the actual value is due to the simplified parallax model used in this calculation. In reality, the Sun is much farther away from Earth, and the angle of parallax would be extremely small.

It's important to understand that the values provided in the statement do not accurately represent the distances and angles involved in this scenario.

The response took Marty’s breath away. The machine understood words without looking at diagrams. The machine,

1.???Broke the question into logical parts.

2.???Identified the goal, i.e., calculate distance.

3.???Identified trigonometry as a computation technique.

4.???Flagged “Parallax” as an idea that was being exploited to calculate distance.

5.???Searched for contradictions outside the problem context.

But the response still was not as satisfying as talking to a fellow human therefore Marty played with his prompt, “On June 21st, Sun is 14 degrees from the zenith when seen from Philadelphia while it is on the zenith in Miami. The distance between Miami and Philadelphia is 1000 miles therefore the radius of earth is 4000 miles. Is this statement true?”

ChatGPT response,

I apologize for any confusion earlier. The statement you provided is not accurate. Let's break down the information step by step:

1. On June 21st, the Sun is 14 degrees from the zenith when seen from Philadelphia.

2. On the same day, the Sun is at the zenith in Miami.

3. The distance between Miami and Philadelphia is 1000 miles.

While it's true that on June 21st, the Sun's position in the sky can be measured in terms of its angle from the zenith, the rest of the statement contains inaccuracies.

The angle of 14 degrees from the zenith in Philadelphia indicates that the Sun is not directly overhead (at the zenith) but rather 14 degrees away from it.

However, the distance between Miami and Philadelphia being 1000 miles does not determine the radius of the Earth, nor does it determine the distance to the Sun from Miami. The radius of the Earth is approximately 3963 miles (6371 kilometers), and the distance to the Sun from Earth is about 93 million miles (150 million kilometers).

In summary, while the concept of measuring the Sun's angle from the zenith on June 21st is valid, the other information in the statement is not accurate and should not be used to determine the radius of the Earth or the distance to the Sun.

Marty wryly shook his head and muttered, “Unreal.” He wanted to explore whether the machine understood irony, humor and day dreams.

_______________________________________________

Author’s note #1: In this exchange between Marty and the machine, we have a dialogue that required conceptual clarity (a concept of a physical model, if you will, of the earth) and contextual familiarity (how far is earth from the sun? Latitude of Miami? What’s the distance between Miami and Philadelphia? etc.)*

Anaxagoras 2,500 years back used Figure 2 as his model to determine the distance of Sun from Earth as ~4000 miles**. 300 years later, Eratosthenes used Figure 1 as his model to estimate the radius of earth. (If you ponder over the figures, you will realize that the angle θ is the same in Fig 1 and Fig 2.) Eratosthenes’s estimate of Earth’s radius of ~4000 miles was not improved for over 1,500 years! Then one would not be wrong to conclude that the learning machine right now fits into the renaissance period. The machine resembles a Renaissance man in its appetite for learning and breadth of knowledge!

In the aforesaid example, two different physical models lead to a single answer being attributed to different distances where only one model is right. How would a learning machine resolve this contradiction? Good quality data would help the machine, but odds still favor a human to solve a problem like, is P = NP? ?We have started living in the midst of large “language” models, i.e., language is facing a computing attack, but do not fear or despair that you need to bow to the machine overlord! There is space for rich, varied and emerging human experience. If the machine does not enable pursuit of happiness, at the very least it will enable reduction in drudgery.

*CCCF - Conceptual Clarity and Contextual Familiarity is an abbreviation from WIMWI. A learner was thought to have deep understanding, if the said learner had good CCCF. If you are a WIMWIan, this is just a quaint reminder. If you are not a WIMWIan, this footnote should not hold your attention at all.

** I am grateful to George Gamow for his well-written expositions on a range of subjects, particularly the book, “A Star Called the Sun,” for this specific example.

_______________________________________________

Marty quickly framed another prompt, “I have a plan to catch an injured stork for treatment. The plan is, I quietly step behind the stork and place a slab of melting butter on the stork's head. The sun would melt the butter. The butter would drip and cover the eyes of the stork. I will grab the blinded stork by its neck and then proceed to treat the stork. Is this a good plan to trap the stork?”

ChatGPT response,

While your intentions to help an injured animal are admirable, the plan you've described is not a suitable or ethical way to catch an injured stork or any wildlife for treatment. Trying to intentionally blind an animal, even temporarily, is both harmful and distressing for the animal.

Marty winked and asked, “Is my plan absurd or funny?”

ChatGPT response,

Your initial plan to catch a stork by using melting butter to blind it and then grabbing it by the neck is certainly unconventional and imaginative. However, most people would likely consider it both absurd and impractical, rather than funny.

Marty murmured, “Can one reason humor into an intelligent machine?”

_______________________________________________

Author’s note #2: Large Language models right now have incredible contextual familiarity. The models available for public have every piece of information (or language) that is to be known and information is available for use instantly! While context mostly can eat conceptual understanding for breakfast, there are times when conceptual understanding will annihilate context. How one balances between context and concept is key for managing use of these models. For problems or opportunities where the conceptual basis is essential, do keep in mind that these are the type of learning problems angels typically fear to tread. Where context is key, a learning machine reigns supreme.

Nothing is to be feared in life, just to be understood.

- Marie Curie

_______________________________________________

Note: All ChatGPT responses you read here are actual ChatGPT (GPT-3) responses. Marty McFly is a fictional character and is the hero of “Back to the future” movies. The name, Marty McFly can be replaced by Jules Verne or Ponniyin Selvan without any loss in meaning or context. My son, who loves “Back to the Future,” is off to college, at the cusp of this exciting period in computing, is to whom this note is dedicated.