Navigating the AI-Driven Future: Skills for Success | by Greg Rog | The Startup | Feb, 2024

How to embrace the world where AI, adaptability, and ‘M-shaped’ skills thrive? Discover how these factors will redefine roles, transform corporations, and unlock unprecedented individual potential.

10 min read

·

14 hours ago

It is estimated that in 2024 we will generate 120 zettabytes of data. It’s difficult to even present this number in perspective, as it’s hard to comprehend that it’s approximately 328.77 quintillion bytes per day, or 328.77 million terabytes.

But that’s beside the point.

What’s important is comparing how much data we generate annually to previous years:

It’s somewhat easier to grasp this in the perspective that various sources report that from the dawn of humanity to around the year 2000, mankind produced a total of 5 exabytes of data. Nowadays, we generate 328.77 exabytes every day. This means that every day, we produce 65 times more data than the entire human race did in 2000 years.

On a side note, it’s fascinating that currently, we’re trying to pack all this data into one small data jar. Essentially, a large language model (LLM), such as the one that powers ChatGPT, which contains a substantial portion of the internet, is essentially two files. One file contains the program execution instructions, which aren’t overly complicated and at most a few hundred lines of code. The second file is the output file of all the magic and the result of training the model — in short, gigabytes of zeros and ones.

By the way, this is quite an expensive game. It’s exactly for this that Sam Altman from OpenAI is currently fundraising for a project that may consume, a trifle, $7 trillion dollars.

But what does all this have to do with the competencies of the future?

We’re slowly getting there. This means that stranded on a deserted island, with the knowledge of how to build a computer and a solar battery, and the aforementioned two files, you’ll probably be able to recreate most of humanity’s intellectual achievements.

Can you imagine discussing such topics with your grandparents, or even your parents?

I’m getting at the fact that most things that were relevant a few years ago in the context of education, or what is worth betting on as competencies that could potentially give us the opportunity for a well-paying job or social advancement in the future, are becoming obsolete right before our eyes.

It’s hard to advise someone on choosing a field of study these days. Because after all, studies that used to offer certainty and prestige, such as law studies, may soon be completely useless. First of all — absorbing vast volumes of legal knowledge and memorizing them seems, to put it mildly, a waste of time in the era of Perplexity, and soon also specialized models, trained for example on the entire database of judgments and legislation.

So how to cope in the rapidly changing technology space that surrounds us, which is changing faster than ever due to AI? What skills to acquire and what will actually be important in some time?

First of all — it is difficult to give a clear answer to this question. Will trends such as VR and AR, Blockchain, Cloud Computing, Data Science, which already seem ultra-modern and promising, stay with us for longer? We can’t predict that at the moment. It’s possible that AI will change our reality so much that even these clear trends will not survive the confrontation with a specialized model.

One thing is certain. The way we perceive our base knowledge will change significantly. This is about the possibilities of an individual who, with access to information with the help of AI, will be able to instantly reach most information from any field.

In this context, any textbook knowledge and memorization of detailed information seem completely unnecessary. It seems much more important to acquire information more superficially, but from different fields.

Until recently, we lived in times of narrow specialization. We are already seeing that the best experts are beginning to draw from different fields of knowledge and apply it to a specific case, going far beyond the area of their narrow specialization. In practice, specialized AI models, which will soon start appearing on the market, will be able to perform any specialized work better than humans. The more “processed” it is, the easier it will be for the model to navigate around specific rules.

That’s why an interesting way of testing models is to do it in an environment of closed rules. For example, AlphaGo, using a strictly defined set of Go game rules, was able to defeat any human after a certain number of iterations. The same goes for chess. Within defined rules, it’s easy to train a model that plays against itself, learning in this way, or to compare the effectiveness of two different models by inviting them to a joint game.

This is a good example of how specialization can translate into practical applications and implementations of AI in professions such as lawyers, accountants, architects, doctors, or builders. This also applies to professions related to technology and programming. And even primarily, as programming languages are even more natural environments for large models than the languages used by people. So, one can assume that soon models will be better than the best programmer in the world and will work with incomparable efficiency.

However, this does not mean that I predict the decline of the programming profession. The effects of such a transformation will simply be twofold:

Firstly, there will be a huge democratization of knowledge and technology. People who previously could not create a program, website, or mobile application will be able to do so with the help of interaction with the model. In this way, we will get enormous added value for the world, where only 4% of people are able to program anything. Soon this value will be close to 100%. Access to the internet and the ability to acquire knowledge, not where you were born, will be the determining factor of what you can achieve.

Secondly, specialists will gain incredible assistance at work, where AI will perform most of the tasks such as writing code, which currently had to be done manually. The programmer will use his specialization and knowledge of technology limitations to guide the model to create a significantly better application than one that would be created by a person without specialized knowledge. In addition, these will be experts who will create the infrastructure for completely new tools and technologies that the rest of humanity will use.

The time saved by AI, such a person will use to gain more general knowledge, for example about interface design, and in this way will be able to deliver more complex tools, programs, and products.

We are slowly entering a phase where an individual will be capable of implementing projects that until now were carried out by huge companies and corporations. Like Sam Altman I am also sure that a single person will be able to build a billion-dollar enterprise.

If you think about it for a moment, this means that such a person will have to find their way in all the processes that are currently carried out by specialized departments. In addition to operational functions, this person will manage marketing, sales, product, customer service, and all other processes — of course, with the help of artificial intelligence and tools, especially no-code and low-code.

But this directly suggests that the current structure of companies and corporations, where each of the employees has some designated role, is a product of a system that will soon become a thing of the past. The competencies of employees will intersect more, and companies will focus on skills that can be described as M-shaped.

However, there is a very important obstacle on the way — our mentality, beliefs, and habits. Everything that currently surrounds us at work is related to specialization. For example, we apply for a marketing position, work in the marketing department, and people from other departments are responsible for other things, such as programming.

The question is whether mentally we will be able to switch to thinking without limitations and, as a marketer, try our hand at programming or another field.

With the help of artificial intelligence, we will be able to effectively step out of our role and carry out tasks whose technical details we do not know and do not even need to understand. And within our specialization, we will in turn reach for new tools, also without the need to understand their mechanics, but only aware of the goal we want to achieve with their help and the general framework and limitations.

In this context, we will do things that we have done so far, but with completely different results, or on a completely different scale. A programmer, for example, will be able to use a language he does not know to program a new application in record time.

On the other hand, for example, a creator who creates videos on YouTube, will be able to create them more effectively and for a significantly larger audience and reach people on different platforms. Thanks to tools that are already able to generate video and content in different formats, distribution to social media or different media of the same original content will be childishly simple. What entire publishers or television stations are currently doing, one influencer will be able to achieve.

In short, all this means that it will no longer matter “how” to do something, and “what” to do will be much more important.

And this is where experience and industry knowledge come into play. The point is that people who understand a given industry well and are specialists within it will gain a huge advantage. However, understanding industry knowledge will be much more holistic and less focused on narrow specialization. Instead of learning new programming languages, the programmer will be able to understand how UI design and interface design work, in order to later use both these fields to create a complete application from A to Z. And in turn, the designer will know the principles of creating software and no-code and low-code tools, which will allow him to complete the project independently.

What will become a key competence can be described as the ability to search for and assimilate relevant knowledge — that is, knowledge that allows understanding of the problem but not necessarily its solution.

Considering how much data and information we generate, and in addition, how quickly this information can become outdated in the era of such rapid transformations in the field of technology, the ability to assess which data are valuable and necessary will become absolutely crucial.

This in turn comes down again to knowledge from many fields, the ability to make a proper assessment and also to ask the right questions. The ability to formulate them is also essential. In the future, most of the effects that we currently get from work with various tools and software are obtained from user interfaces. In the future, we will get these effects from the interaction with the model itself — this means that we will have to learn to issue effective commands and in this way we will receive not only answers, but also actions controlled by AI.

This concept is not at all a far-reaching, futuristic vision. We ourselves, right now, within the Alice application that we are developing, interact with tools and software in this way. For example, in this X, I show how Alice is already able to manage my calendar through conversation:

We are gradually granting access to Alice to Techsistence subscribers. We also plan a larger educational program that will help you understand the mechanisms of creating automations that allow Alice to interact with other tools. I will inform about this in the next posts.

In summary, the foundation of what we should consider as competencies of the future is a set of skills: firstly, adaptation and curiosity to discover new technologies and work with AI. Secondly, holistic knowledge, going beyond narrow specialization, but rather shaped in an “M-shaped” way, i.e., with the acquisition of specialist knowledge from various intersecting fields. In addition, accepting the fact that we do not need to know how something works underneath (basic competencies), but rather have an awareness of frames and schemes — limitations within which we operate. Finally, the ability to assess the truth, relevance, and search for reliable and necessary information, as well as conversation with the AI model — something we can safely bet on as key competencies in the coming years.