LLMs and jobs good or bad? -

During the last months there is a great debate about the disruption of LLMs (large language models) and the impact on many fields.

One of these fields is jobs. You can read comments in basically 2 directions:

Pros: LLMs is a great opportunity!
Cons: LLMs are going to destroy million of jobs, it’s a disaster!

Ok, good, on this context, I do not how to react, I still have not a strong opinion about what is going to happen.

How do I see it?

I have been reviewing Simon Wardley’s book (again) reviewing notes about the climatic pattern: economy has cycles. This time my reading has been on Chapter 10.

Basically economy has cycles and now in the technology space (that is much rich than LLMs, social networks and other “popular” stuff) is continually evolving.

The table that better helps to gain perspective about the disruption of LLMs in this moment of uncertainty is this one:

So, let’s try to see how the context

Context about the labor market around technology

First of all some definitions and a warning.

Definitions

Technology jobs: The definition of technology is very wide and when I refer to “technology jobs” I mean to these jobs that are done more than 50% of the time in front of a computer.
Digital Transformation: I refer to all these works being done as move to the cloud, practices around workplaces (remote workplace), security (the great forgotten), move to cloud specialized services (from ad-hoc products), etc….
Users: I have drawn 2 users. The present user, with actual needs and that in general is asking him/her self “why do I need ChatGPT?” And the future user (red) that is going to have more clarity about what s/he wants.
User needs: there are many of them, I just have differentiated the needs of today (blue) and the future needs (red).
LLMs are now in it’s genesis, and they are going to evolve quickly. The LLMs existing in 2030, how they will be? I have no clue.

Warning

I draw this map alone, and I’m sure it’s wrong. Happy to be challenged on this map 🙂

First map: short term perspective

Right now the majority of the software industry is competing in the “Digital Transformation” space that to my view is in competitive state fo WAR.

You could have the same map 1 year ago and substitute “LLMs” by “Metaverse”. The mediatic disruption was the same. Where is the “metaverse”? Well, people working on metaverse continue with their work, they just have disappear of the newspapers. But they are still there: competing to make their projects successful.

At this point I want to remind a weird map that tries to show how uncertainty works in a competitive landscape:

It’s a reminder that entrepreneurs do not need, but it’s a good reminder for other people (like me): majority of initiatives do not survive to its genesis.

After this reminder, let’s go back to the map. LLMs right now are in their genesis: full of uncertainty, it’s seen as a marvel, nobody knows how to use it, it has a lot of potential value…

On the other side, we have the majority of the industry working on the “Digital Transformation” arena.

Let’s look at the map with more perspective.

Second map: more perspective

LLMs are not looking at the user of today, they are looking at the user of tomorrow (let’s say the user of 2030). The user of today does not know what s/he needs from these technologies.

In the following years we will see many companies investing to evolve the LLMs. Few companies will survive (it happened in the past, and the cycle will repeat again).

The amount of companies disrupting with LLMs will be small. When I mean disruption I mean real disruption, for instance: OpenAI is a disruptor, a company using OpenAI connected with Zapier offering a specific use case for a user… I’m sorry that’s not a disruptor.

What do we have between LLMs and the future user?

We have an empty space. Between these technologies and the future unsatisfied list of needs of the user there is an empty space.

What is going to happen? I don’t know, right now I only have questions:

Who is going to catch the more fish in that new herring?
What are the real needs of that future user?
What would be the name of the new practices related to the use of LLMs?
How ethics is going to be taken into account?
How aware is the average user about his/her privacy on social networks? How that average user will react on LLMs?
What are the real use cases that will survive more than 3 – 5 years?
How are citizens going to be protect against these threads on privacy?
How quickly are going the governments going to act on this space?

Takeaways

My unsolicited opinion: there is a sea of opportunities in front of us, we just have to look at the right perspective, run for them.

There’s a hype right now, it’s not the first one, and hopefully not the last one.