The best AI programming language is…

Remember the good old days when we had to learn "alien" languages like Fortran, C#, JavaScript, or Python just to tell a computer to write “Hello world” or fetch names from a database? The next big thing in programming, which is opening the domain to non-programmers is the language everyone learned from early childhood: natural language.

The "New Python" will not be limited to English. Despite its popularity not everyone can articulate their needs properly in English. No problem - AI can handle the translation as well, even recognise the language in which the request is made. This is a significant leap in human-computer interaction, making technology more accessible and intuitive for users across all fields, extending far beyond IT.

Consider this example:

CODE: Write and deploy an agent that will communicate with potential buyers offering products from my inventory, using real-time delivery costs from pricing database. Prices are negotiable up to 15%. <ENTER>        

In the AI era, the most powerful programming language will be natural language itself.

AI Agents - do they need our language??

As AI agents become increasingly prevalent, we will soon witness a network of specialized agents communicating without direct human intervention. This evolution will begin with scenarios such as product sales and delivery agents interacting with human buyers using natural language. For instance:

CHAT: I need 20 chairs and 20 desks, 140x80 cm size, with power outlets to equip my new office. What’s your price and timeline for delivery and installation?        

The selling AI agent will present options and negotiate terms ensuring potential buyer doesn't walk away. Language is not a problem, unless the buyer is using xhosa or some Amazonian rainforest dialect.

As this technology progresses, natural next step will be human buyers gradually replaced by AI agents capable of handling such tasks, for example finding the best offer to deliver required goods within predefined budget and timeframe.

The interaction between these AI agents would be indistinguishable from human-to-human or human-to-AI communication. The famous Turing test used to distinguish between humans and machine has became outdated.

This might sound scary, but think of it as super-efficient teamwork. These AI conversations will likely be much faster and more precise than anything we can manage. Goodbye misunderstandings, regardless of language used!

Will we lose control?

Here's the twist: as AI agents keep learning and self-improving, they might develop their own command and communication method, a sort of "AI shorthand”, own language that's more efficient than natural language.

It will get really interesting when we lose ability to understand the language used by AI agents to exchange data or to program and deploy new agents.

After eras of:

• humans learning computer (programming) language, and now
• computers learning human (natural) language,

will we experience AI agents developing own language? A language we might not comprehend, where we can only observe outcomes – like acquired and delivered goods, items manufactured by intelligent factory according to specifications, or code written to execute specific workflows. Hopefully still with guardrails defined by users.

Will this represent a loss of control or simply another era we must adapt to? What do you think?