From BBC
'Vibe coding' sounds pretty scary to me, in fact: what could possibly go wrong? We'll know when a plane falls out of the sky.
As for the other words, they are generally covered by existing words or not that useful (e.g.: coolcation, because a vacation is a holiday centered on time you hope to be spending with interesting bovines...).
My favorite would be taskmasking, which is as old as work itself, though. As for micro-retirement, I prefer macro-retirement, which is what you do when you actually, well, retire - something every 20-year-old in the West seems to crave, nowadays, as they are so tired, bored and blasé already, even before they have even started working.
25 years ago, an Italian researcher, developed a code generation system that used speech as its input.
20 years ago, Microsoft SQL Server Database Management System supported an English language query mechanism, in effect, taking English statements and converting them into Structured Query Language queries.
IBM has a Rule Engine product that can be programmed via English-like Syntax.
As long as one asks for something that can be generated from the training set that has prior solutions, system can generate a schema, the code, and all that goes with it.
But once you are off the beaten path, one is back to puzzling out the requirements of the system that one desires to build.
Consider the UK legislation for MPG and Emissions standards as the requirements for a reporting system that, say Land Rover, must have to report the MPG and Emissions of its vehicles to UK government. Can the text of that legislation be fed into an LLM system and out would come a GUI, codified business rules, the database, and the queries that would furnish to UK government the mandated Clean Air Data?
I think not.
It takes real analysis and transformation by actual human beings to take that legislation into Programming Requirements. And no one can write those software requirements at the level that can be fed to LLM and out would come an entire system.
LLMs would need to have been trained on training sets, in any case, which include numerous such systems.
For limited domains, small in scope (class, method, database schema) this can be done. And even then, the trouble begins when the LLM-generated system needs to evolved. Since Computer Programs are mathematical objects, one is back to needing people with the mathematical aptitude for the changes and maintenance of the system.
If a real system with 1 million lines of code could be generated by LLM, then you have a breakthrough, without a doubt.
Another example would be a simulator for Intel 8080 chip; can one ask an AI code-genrator to produce one?
We will see.
(I could not get ChatGPT or Copilot to give me the correct code for creating a specific geometric structure in FreeFem++...)
I suppose some will say that this point will come, and perhaps sooner than we think.
The problem is that one would need to check there were no mistakes in the coding, I suppose, as machines also make mistakes, don't they?
_____________
If you've ever wanted to create your own computer program but never learnt how to code, you might try "vibe coding".
Collins Dictionary's word of the year - which is confusingly made up of two words - is the art of making an app or website by describing it to artificial intelligence (AI) rather than by writing programming code manually.
The term was coined in February by OpenAI co-founder Andrej Karpathy, external, who came up with the name to represent how AI can let some programmers "forget that the code even exists" and "give in to the vibes" while making a computer program.
It was one of 10 words on a shortlist to reflect the mood, language and preoccupations of 2025. [...]
No comments:
Post a Comment