The Tech-Domain Language Barrier Is Breaking – Now What?
LLMs and generative AI tools as a bridge between domain language and tech.
I have a thesis that drives much of what I'm focusing on right now:
Domain language and tech language is converging and LLMs are an acceleration force
This will shake up the role of the tech worker. It’s also an opportunity for domain experts to get the same leverage on their skills as programmers has had in the past.
Not a new trend
AI tools decrease the gap between domain language (often expressed with natural language) and technology (often expressed with code).
This is not a new trend. Compare the first computers with punch cards with modern programming languages like Python and further with the LLM-powered tools like lovable.dev and bolt.new. The punch cards were far from business language, Python is more accessible but still technical. LLMs however enable domain experts to express business logic in plain English and have that translated to a program without a large dev team.
Looking into the history, we can see that LLMs are not the initiators of this convergence . It is, however, an accelerator and I suspect that we’re at an inflection point where the accuracy of translation from domain language to a working program will increase fast over the coming years.
Looking forward
Extrapolating this trend and the exponential growth in the post-LLM era, we can see a few possible consequences:
New AI-first programming languages. Right now, LLMs typically translate domain language to code in languages like Python, JavaScript or SQL. But what says that these languages are the optimal bridge between domain language and machine? Maybe we can remove some constraints from existing languages when most code will not be written by humans.
Domain expertise, design and soft skills becoming more valuable. This also means pure coding skills and the ability to implement a product from a spec becomes relatively less valuable.
Product people and developers have a head start. By being able to adjust and correct code created by the current AI tools, developers are well-suited to be early adopters. On the other hand they have more to loose and might thus resist change.
More power to idea people. Domain experts with deep knowledge and unique insights will be able to realize ideas faster and cheaper without large development teams. Generative AI will benefit non-tech idea people by making it easier to quickly spin up things and iterate.
Summary
LLMs and the generative AI tools they enable will accelerate the convergence between domain language and tech. This will shake up the tech industry and the role of the programmer but also open new doors for domain experts without programming skills and resources.
Just knowing how to express a need in tech solutions won't be an edge anymore.
Next step
I'll keep posting more about this topic and other observations and insights from my daily life at the edge of this development.
Are you building in this area or a domain expert curious about AI? Hit me up and let's discuss it.
Leave a comment and subscribe for more on this topic. Over the coming weeks and months, I’ll keep sharing my thoughts and observations along with practical tips for both tech and domain experts.