AI will not replace humans, but humans with AI will replace humans without AI.
During these months the trickle of news first about ChatGPT and then about GPT4 has been incessant. Beyond the astonishment about its capabilities and the evidence of the great improvement that GPT4 represents over the previous version, for example, the factual truth metrics have gone from a 5% error to 1%, two positions have been clearly outlined differentiated.
On the one hand, we have the voices that call for caution and regulation, although without specifying which one. Some of them even stand by the companies that develop these products. All this in order to preserve the veracity and avoid what some call an instrument of "mass intoxication". Some point out that in a few years we will go from an internet where most of the creation is human to another where the human will be residual.
On the other, we have a multitude of companies and some governments that have signed up to capture value and join the race.
For example, Bain, one of the large American consultancies, has long had an agreement with OpenAI to develop products that use ChatGPT in their consulting and auditing activities. PwC also announced an agreement for its legal department. Stripe and many others for use in communicating with customers. Also in pharmacological research, etc... or in teaching, especially programming with agreements with Khan Academy or languages with Duolingo. Along with these agreements that have been disclosed, it is well known that there are many that have not been disclosed because some companies have opted for a low-key attitude. And of course the big tech companies like Microsoft or Google are rushing to incorporate it into their generic applications like Microsoft Office, Gmail or Microsoft Dynamics.
Also some governments have already rushed to be in the race and not be left behind. As I write these lines I read in The Guardian that the British government is going to allocate 900 million pounds to equip itself with the supercomputing infrastructure and create the BritGPT and the Icelandic government has decided to use it in its e-gov platform.
The speed of the phenomenon is such that it has managed to capture the imagination and concerns of managers and companies, even placing it in first place, above the economic crisis, in a recent McKinsey report.
What will be the immediate future of this change?
The future that is already here are generic applications. Those of search, office automation, ... that Microsoft and in a certain way Google have already presented. It is an important issue because the redefinition of search eliminates links and advertising and, therefore, the business model that supports companies like Google.
Likewise, its incorporation into office automation tools redefines our way of working, in the same way that the incorporation of the spreadsheet redefined the work of accountants and the business world in general. The fastest to learn and reinvent themselves win.
The second level is the verticals, some are already practically here. One of them is programming, which will be with GPT-style co-pilots, greater productivity and fewer errors.
Something very similar happens with consulting, with similar but not identical projects that are difficult to automate. But perhaps the sector where the revolution will be most striking will be the legal one. For decades the work in law firms has not changed substantially. As in consulting, it is based on searching for precedents, jurisprudence, similar but not identical tasks, the ability to synthesize and combine… texts. Precisely what these models do so well.
Education is going to be in full swing. Khan Academy with its tutor Khanmigo based on GPT shows us what the path will be, but also Duolingo with its tutor that allows us to chat with a "native" in real situations.
And of course the app ecosystems that will use these models in the form of APIs. That is where we will see an explosion similar to that of the first iPhone.
But why is it so important not to be left behind? Will we not be able to adopt these technologies when they are more mature?
There is a big difference between generic technologies and more specific ones. General Purpose Technologies (GPT) have a great influence on the whole industry and society. Until now, the paradigmatic example was electricity that transformed work, schedules, social life,... basically everything.
Knowledge, the ability to handle language and concepts is even more generic than electricity. This is what we are up against.
It would seem absurd not to embrace electricity and not be experts in its use. A society that had followed that path would have been cornered by progress and history. Electricity had many problems, many accidents, but it made us not only more competitive but also better.
Technological changes are windows of opportunity that redefine the position of countries, their companies and their inhabitants. In these windows, the fastest to learn, create and reinvent themselves win. Once the technologies have been massively adopted and are common currency, it is difficult to find another window of opportunity that allows us to position ourselves.
Humans are not going to be replaced by Artificial Intelligence, but humans with AI will replace humans without AI.
