Skills and creativity are human traits that should not become rare

I have a gloomy feeling that software is getting better whilst human skills decline and creativity dries up. I work on software, but I also interact with dozens of customers and colleagues every day, and I notice less overall curiosity among my peers.

AI tools are indeed a type of software. They are an invention that has unlocked productivity by extending the limits of what human brains can process: large datasets, codebases or legal corpora can be analyzed and repurposed for different goals. I would celebrate this achievement unconditionally if I saw people in my circles gain more power and knowledge from it.

It seems that overall, we are going for convenience rather than empowerment. In my field, code can now be written by other software; products can be implemented, delivered and maintained by human-guided AI in a completely new way that's both better and worse. It's worse when product design is an afterthought, infrastructure design is a purely utilitarian choice, and there's no understanding or intentionality.

Watching this reality unfold keeps pushing me towards more consultancy work and fewer engagements focused on production work. I believe that my approach to product, software and infrastructure is driven by curiosity. AI teaches me new things every day, especially the value of discernment, speculative thinking and building skills that will support future decisions.

I have recently completed several projects ranging from finance to tech support where I built custom applications to perform deterministic and inferential analysis on large datasets, using private AI models for security, choosing data anonymization techniques and meeting complex requirements. I realize that what I did was the result of more than 20 years of experience and constant research. AI has boosted my ability to experiment and iterate, and in some cases shown that code is a means to an end.

Product design, market fit and distribution are much harder, unsolved problems. They relate to human perceptions and depend on much more than the underlying code. As such, digital products require creativity and skills that can't be replaced with model training alone.

I am going to focus on hard, human activities even more. Teams with poor adoption of AI can be empowered by the right amount and approach to it. Products already skewed by agentic decisions will need a fair touch of humanity. Ventures trying to make sense of tons of data will get the experience of data analysis powered by deterministic-inferential analysis.

I'm continuing to explore, and hopefully I'll witness a shift towards a more conscious and intentional use of AI and software in general. The best use I can make of my approach and experience is to advise teams on making the right choices, analyzing problems and tackling them with the right mindset.

The potential is there, but people should be given the skills and encouragement to develop their curiosity through experiments, supported by recognition, shared purpose and financial rewards.

I keep thinking we may need fewer people for each task. The truth is, we may need better informed and better trained people, in roles that go beyond the routine operational jobs companies still hire for.