GPT and the Consequences of Knowledge Friction

ChatGPT has democratized knowledge in ways that few innovations outside of the Gutenberg printing press, the internet, and search engines have done.  It not only finds content, but answers our specific questions with formatted explanations and analysis, and remembers our conversations at a later date.  

This month, the new AutoGPT is making headlines.  AutoGPT enables advanced reasoning capabilities and understands context and concepts effectively in a configurable and automated manner.  AutoGPT can provide valuable insights and recommendations, supporting data-driven decision-making and facilitating efficient business processes all automatically.  

 

The automation component of AutoGPT can help reduce knowledge friction caused by the lack of time.  If you don't have the time to study and research important and impactful topics, the lack of time becomes a source of knowledge friction.  Automating the research, analysis, formatting and distribution of knowledge is a powerful feature.

In business, knowledge friction hurts.  It often forces leaders to make decisions based on conjecture, rather than by data-driven decisions. In addition, it can have the following implications:

1. Information asymmetry: When one party in a transaction has more or better information than the other, it can lead to imbalances in bargaining power and market inefficiencies. For example, when sellers with low-quality products can exploit uninformed buyers.

2. Market failures: Knowledge friction can contribute to market failures when information is either scarce or costly to acquire. For instance, consumers might not be aware of the full range of available products or their quality, leading to suboptimal choices. 

3. Barriers to entry: Imperfect information can create barriers to entry for new businesses or innovations. Startups may struggle to convince potential customers of their value proposition or face difficulty acquiring necessary information to compete effectively with established firms.

4. Price discovery: Knowledge friction can impact the process of price discovery in markets, as buyers and sellers may not have complete information about supply and demand conditions. This can result in price volatility, inefficiencies, and the misallocation of resources.

5. Decision-making: In the presence of knowledge friction, individuals and businesses may face difficulties in making informed decisions, leading to suboptimal choices and potentially reduced economic efficiency.

6. Innovation and technology diffusion: Knowledge friction can slow down the dissemination of new ideas, technologies, and best practices, limiting the potential for innovation and technological advancements to drive economic growth.

7. Competitive advantage: Firms that can manage and reduce knowledge friction may gain a competitive advantage over their rivals. This can be achieved by investing in research and development, employing better information management systems, or developing a reputation for transparency and trustworthiness.

Knowledge friction plays a significant role in capitalistic markets by influencing market efficiency, competition, innovation, and decision-making. Reducing knowledge friction can lead to improved market outcomes, but it's important to understand that not every businesses wants to remove knowledge friction.  Some companies have built successful businesses in niche markets that thrive on knowledge friction.  They are unlikely to be as enthusiastic with artificial intelligence as others.

************************************************************************

Kevin Benedict

Futurist at TCS

View my profile on LinkedIn

Follow me on Twitter @krbenedict

Join the Linkedin Group Digital Intelligence

***Full Disclosure: These are my personal opinions. No company is silly enough to claim them. I work with and have worked with many of the companies mentioned in my articles.

The Amazing Potential of a ChatGPT and Human Pairing

I have been having all kinds of fun with OpenAI's Dall-e for art, and ChatGPT for answers, research, testing and jokes.  Let's talk about ChatGPT and jokes.  It is not very good at writing funny jokes.  It has an academic understanding of what jokes are, but it finds it difficult to deliver them.  Out of hundreds of attempts, here are a few of the best jokes ChatGPT could come up with:

• Why did the Luddite start using social media? So he could complain about it. 
• Why did the bioengineer create a new species of bacteria that can glow in the dark? To shed some light on the subject.
• When asked what he was working on the bioengineer answered, “I could tell you, but then I'd have to genetically modify you."
• How many Luddite farmers does it take to change a light bulb? None, they prefer candles.
• What do you get when you cross a Luddite with a Time Machine?  A trip to the past no one wanted. 

Here's the thing with the above jokes - they almost worked.  I had to tweak them just a bit to get them to work.  ChatGPT puts most of the right words together, but not necessarily in the right order to surprise and create humor.  ChatGPT is, however, a great idea generator, and generating ideas is immensely valuable.

I have found that if you ask ChatGPT to write some generic jokes it fails.  If you tell it to write some jokes with a combination of interesting characters such as a bioengineers, Luddites and a priest, you start getting material with some great ideas.  Again, ChatGPT mostly fails to be funny, but it's attempts provides some good material to get your creative juices flowing.

I have come away impressed with Dall-e and with ChatGPT.  They both make great human/AI pairings.  They help me produce better content, faster. 

I am now regularly producing humor from ideas generated by ChatGPT, and the illustrations generated by Dall-e.  Follow me on Twitter @krbenedict, or follow me on LinkedIn or Instagram@futurist_humor to see them.

It is clear to me that ChatGPT, and other AI platforms using large language models, can offer incredible value to most knowledge workers.  

I met with an engineer friend of mine last week, and he asked ChatGPT what it knew about some bleeding edge engineering topics. It produced an accurate summary, and he was impressed.  It could have written an executive summary for him.

I encourage you to test it.  Learn where it is strong, and where it is weak.  Use it.  Your competition will be. 

************************************************************************

Kevin Benedict

Futurist at TCS

View my profile on LinkedIn

Follow me on Twitter @krbenedict

Join the Linkedin Group Digital Intelligence

***Full Disclosure: These are my personal opinions. No company is silly enough to claim them. I work with and have worked with many of the companies mentioned in my articles.