ChatGPT - Top 10 misconceptions - weirdwonderfullworld.com

The Chatbot, “ChatGPT”, (Generative Pre-Training translator), has recently been the new kid on the block raking in millions of new users breaking records for first to market.

In this blogpost we discover 10 common misconceptions about ChatGPT and trust me there is more to this chatbot than meets the eye.

Figure 1: OpenAI’s logo

1. Using ChatGPT to replace software developers

People fear new and revolutionizing technology. They are afraid to lose their job and therefore be unavailable to financially support themselves.

A tale of horses and automobiles

This was the case with the introduction of tractors suppressing the use of horses. Historically speaking, horses was used to plow fields, drag harvested crops and equipment across the farmland. They were indispensable on farms and used heavily in cities.

In cities horses was used for everything from driving carriages with people, postal services, public transportation and transportation of goods and services. This required a stable amount of draft horses that needed horseshoes, saddles and brushes as maintenance for the draft horse stables.

Horse driven public transportation — **Gallery 1:** *Examples of old-fashioned use of horses*

Differences in style based design of the different carriages utilized backwards for the royal families.. — **Gallery 1:** *Examples of old-fashioned use of horses*

With the advent of the automobile all jobs in relation to the horses got replaced by automobiles. This replaced horseshoe makers, saddle makers and similar jobs with mechanics and factory workers.

New horizons and beginnings

Consequently, the innovation of the automobile replaced previous job-types and was initially met with extraordinary resistance and fear from the people. For fear that, the unknown new ‘dangerous’ technology was generally a big problem for the general public which reacted with civil unrest and demonstrations.

Now one may wonder, is that a better way of life for the general public? As it turns out people are smart and adaptable. They are able to adapt to their surroundings. The horseshoe makers changed their professions in accordance with the advent of automobiles. Consequently new jobs are created ensuring that people innovate and so the cycle continues.

ChatGPT – Replacing modern day horses?

In this case, I believe that ChatGPT won’t revolutionize standard software architecture jobs. The chatbot cannot fully abstract to the level which humans are capable of.

Necessity is the mother of invention and software architects/programmers will either specialize or new jobs will be created which they adapt for. The innovative capabilities of ChatGPT will compete against software programmers and possibly free up resources to be allocated towards more challenging tasks.

2. ChatGPT will conquer the world

The invention of new technology embraces the welcome feeling of new and untested territory which brings forth emotions of excitement and joy. However also feelings of discomfort and uncertainty, moral dilemmas and ethical considerations are keystones of the modern innovation space.

Specifically the feelings of fear with the unknown must be dealt with effectively. The spurious thoughts spawn from fear. These include feeling of disempowerment, and include the fear that ChatGPT will conquer the world.

Echo-champers

The irrationality of such thoughts are difficult to comprehend for the isolated individual. Especially if that person surrounds itself will people of the same belief systems. Such scenarios exist causing so-called echo champers. Echo-champers are ideological abstractions which encourage radicalization of ideas. This enables users to develop own beliefs of impeding doom.

AI-generated image of an abstract cheese bell shaped echo-champer — **Figure 1**: *A ‘cheese’ bell echo chamber sketch of a person -AI generated*

The lack of evidence is neglected and becomes evident when attacked by critical thinkers from the outside. To clarify, ChatGPT will not conquer the world. Neither will it enable terrorists to try and capture key infrastructure from scripts developed by ChatGPT.

3. ChatGPT will replace google

One of the ChatGPT – top 10 misconceptions, is that it will replace google as we know it. Google is the well-known search-engine answering all questions which people might wonder. It is basically an indexed table of the internet where users can post and question all of the naturally happening world.

On the other hand, we have ‘ChatGPT’ which according to the creators themselves (OpenAI source: https://openai.com/blog/chatgpt/) is an optimized conversational language model. The ChatGPT language model is based on reinforcement learning models, supervised learning and specifically Reinforcement Learning from Human Feedback (RLHF-models). The chatbot is therefore able to make informed conversations with its users remembering what the conversation is about and admitting mistakes underway.

Figure 2: ChatGPT diagram over the structure of the reinforcement learning models.

A profoundly different methodology than the index based search engine which is google. Unlike google, ChatGPT generates answers utilizing its AI engine. However these answers credibility is up to the individual users interpretation while general credibility for source material indexed from the users of google searches is based on source material written by other humans.

4. ChatGPT will replace writers

Another of the ChatGPT top 10 misconceptions is that ChatGPT will replace writers. This is a misconception which has partial credibility. Specifically ChatGPT is a chatbot that can generate paragraphs of text. This text is readable by humans and is somewhat easily comprehendible.

With this being said, ChatGPT might actually replace human writers with written blog material, news agencies or similar types of material. This is particularly different than writers of novels, books or poems which are profoundly different abstraction levels than commonly used material.

Consequently, ChatGPT will most likely not replace common writers or writers as we know it to exist. However it can likely make people writing for news agencies and similar unable to accurately make jobs.

5. ChatGPT will develop new languages

One common misconception about ChatGPT is that it will create new languages. In the world today there exist more than 5000 languages worldwide, not counting professional programming language syntaxes such as C++, Fortran and Python.

Creation of new languages

To make a new language you would have to have thousands of people using it on a day to day basis with its own history, culture and verbal/non-verbal communications.

Therefore it is not possible to correctly characterize the new chatbot ChatGPT, as a new language, nor does it actually invent new languages by replying to prompts through an interactive fashion.

The conversational logic behind ChatGPT might trick the users into thinking that it is a sentient being, thus passing the Turing test, however this is not yet the case and certainly the conversations you are having with the chatbot is based on trained neural networks, not actual sentient thinking.

6. ChatGPT will automate life as we know it

In our top 10 misconceptions about ChatGPT, this is one of the more spacey statements. To start off let’s characterize what the capabilities of ChatGPT is.

ChatGPT is a conversational reinforced learning network, answering to prompts received by human consoles. Similar to modern pc’s it can calculate and respond to a given set of instructions supplied via a prompt.

Similarities to a regular pc

The deviations from a regular pc, comes both from the fact that it is software algorithmically driven and that it therefore fundamentally runs on similar binary logic as a regular computer.

Secondly and more importantly, ChatGPT is unable to communicate with existing infrastructure and hardware meaning that it is unable to actually take control over old machinery or similar.

It can thus easily respond to prompt and provide answers and corrections to questions. However when trying to solve real world problems such as supplying hungry people with food and clean drinking water, it falls sort and will be unable to plan and perform changes in the real world for people.

7. ChatGPT will change everything

The changes which will be incurred from the world wide usage of ChatGPT lies more when trying to answer questions which we already have answers for and need people to execute.

In these regards ChatGPT might very well change common work principles and routines for people working with software and it-infrastructure.

Old fashioned ways

However when it comes to complex labor tasks such as; harvesting of crops, processing of clean drinking water and planning for road constructions, the old fashioned ways with project leaders and workers will remain unchanged going forward.

There will constantly be a need for people with a plethora of skills solving complicated pieces of work all within the same puzzle. This type of work with a mixture of manual and knowledge driven work, will be key when trying to fundamentally change the way of lives of regular people.

ChatGPT will change a small fraction of these tasks and most likely create new jobs along with it.

8. ChatGPT will replace humans

This misconception might actually have some grain of truth within itself. It is likely that ChatGPT will replace human interactions on a certain level come it to be scripting, coding or answering regular questions as a chatbot in public service infrastructure.

Boring jobs and uninteresting work

The replacement of boring, unchallenging, uninteresting work of humans by machines is a constant when innovating. It is important for people to feee up more of their time to do stuff which they want to do, such as art, music and creation of innovations.

However to think that ChatGPT will actually replace humans as we know it and make us obsolete in the workforce is fear mongering at best.

9. ChatGPT will replace doctors

Second to last on our top 10 misconceptions about ChatGPT is one of the more controversial topics, medicine and particularly doctors.

Big tech response replacing doctors

Recently, big tech giants like google, Microsoft and similar have invested quite a substantial amount of money into developing an artificial doctor capable of accurately provide diagnosis of sick people.

Their job is trying to effectively remove the old traditions of booking appointments with doctors to get diagnosis but rather take tests and answer questions by a chatbot.

Conversations as therapy

This last part respond well with ChatGPT however it won’t replace the regular doctors as we know it. The conversational chatbot ChatGPT is unable to read and process tests delivered to a computer and although the response to answers from patients might have a therapeutic effect it won’t replace hardcore blood samples.

ChatGPT is built with responsive language models and conversations in mind, not difficult deep-learning neural network tasks such as eliminating possible diseases.

10. ChatGPT will posses emotions

Last on our top 10 misconceptions about ChatGPT is that it will poses emotions. on a more abstract level, emotions are quite subjective feelings arising from a complex network of neurological connections whose interactions aren’t fully understood.

Emotions are defined as human characteristics best expressed when subjected to extreme circumstances. They are not possible by definition to invent into machinery and especially not ChatGPT, as feelings are universally inherent in human nature and thus cannot be prompted by a mere chatbot.

Bonus

Now as a final bonus for this blogpost consider the paragraph below which have been entirely generated by an AI talking about ChatGPT misconceptions. In my personal view, not an entirely bad addition to our humanly written content above, but what do you think? Will ChatGPT be the end to human life as we know it?

Misconception:

ChatGPT can entirely simulate a human-to-human conversation, replacing actual humans in the chat lounge. Despite its AI-backed capabilities, ChatGPT does not yet possess the capacity to completely replicate the nuances of a person-to-person conversation due to its programming.

References

https://openai.com/ – link to proper ChatGPT chatbot