How so? A Large Language Model is usually a transformer based approach nowadays, right (correct me if outdated)?
AI is artificial intelligence, which has been used and abused for many different things, none of which are intelligent right now (among others used for machine learning).
Machine learning is based on linear algebra like linear regression or other methods depending what you want to do.
An algorithm is by definition anything that follows a recipe so to say.
All of these things, bare transformers and newer in development approaches like spiked neural networks or liquid neural networks are fairly basic, no?
If something uses a lot of ifelse statements to do stuff like become a “COM” player in a game, it is called an Expert System.
That is what is essentially in game “AI” used to be. That was not an LLM.
Stuff like clazy and clang-tidy are neither ML nor LLM.
They don’t rely on curve fitting or mindless grouping of data-points.
Parameters in them are decided, based on the programming language specification and tokenisation is done directly using the features of the language. How the tokens are used, is also determined by hard logic, rather than fuzzy logic and that is why, the resultant options you get in the completion list, end up being valid syntax for said language.
Now if you are using Cursor for code completion, of course that is AI.
It is not programmed using features of the language, but iterated until it produces output that matches what would match the features of the language.
It is like putting a billion monkeys in front of a typewriter and then selecting one that make something Shakespeare-ish, then killing off all the others. Then cloning the selected one and rinse and repeat.
And that is why it takes a stupendously disproportionate amount of energy, time and money to train something that gives an output that could otherwise be easily done better using a simple bash script.
I am not talking about what it does, I am talking about what it is.
And all tools do tend to replace human labor. For example, tractors replaced many farmhands.
The thing we face nowadays, and this is by no means limited to things like AI, is that less jobs are created by new tools than old destroyed (in my earlier simile, a tractor needs mechanics and such).
The definition of something is entirely disconnected from its usage (mainly).
And just because everyone calls LLMs now AI, there are plenty of scientific literature and things that have been called AI before. As of now, as it boils down all of these are algorithms.
The thing with machine learning is just that it is an algorithm that fine tunes itself (which is often blackbox-ish btw). And strictly speaking LLMs, commonly refered to as AI, are a subclass of ML with new technology.
I make and did not make any statement of the values of that technology or my stance on it
I’m saying that code completion does not constitute AI and certainly isn’t LLMs.
I then provided an example of why that isn’t the case.
You decided to respond to this by pointing out that some LLM may be involved in some code completion. Although you didn’t provide an example, so who knows if that’s actually true, it seems sort of weird to use in LLM for code completion as it’s completely unnecessary and entirely inefficient, so I kind of doubt it.
I just want to point it out for a minute, because it’s sort of feels like you don’t know this, code completion is basically autocomplete for programmers. It’s doing basic string matching, so that if you type fnc it also completes to function(), hardly the stuff of AI
I’m sure everyone has always explained this to you given the number of down votes, but algorithms aren’t equal to AI.
Ever since the evolution of AI people seem to have lost the ability to recall things prior to 2019.
I mean doesn’t it heavily depend what you refer to as AI?
ML algorithms, come very close to LLMs and have been back in the day refered to as AI. They are also used in code completion.
Also both of these are algorithms, but with weights defined by data input.
You seriously misunderstand what the acronyms you’re using refer to. I’d suggest some reading before commenting, next time.
How so? A Large Language Model is usually a transformer based approach nowadays, right (correct me if outdated)?
AI is artificial intelligence, which has been used and abused for many different things, none of which are intelligent right now (among others used for machine learning).
Machine learning is based on linear algebra like linear regression or other methods depending what you want to do.
An algorithm is by definition anything that follows a recipe so to say.
All of these things, bare transformers and newer in development approaches like spiked neural networks or liquid neural networks are fairly basic, no?
EDIT: typos
If something uses a lot of
if elsestatements to do stuff like become a “COM” player in a game, it is called an Expert System.That is what is essentially in game “AI” used to be. That was not an LLM.
Stuff like
clazyandclang-tidyare neither ML nor LLM.They don’t rely on curve fitting or mindless grouping of data-points.
Parameters in them are decided, based on the programming language specification and tokenisation is done directly using the features of the language. How the tokens are used, is also determined by hard logic, rather than fuzzy logic and that is why, the resultant options you get in the completion list, end up being valid syntax for said language.
Now if you are using Cursor for code completion, of course that is AI.
It is not programmed using features of the language, but iterated until it produces output that matches what would match the features of the language.
It is like putting a billion monkeys in front of a typewriter and then selecting one that make something Shakespeare-ish, then killing off all the others. Then cloning the selected one and rinse and repeat.
And that is why it takes a stupendously disproportionate amount of energy, time and money to train something that gives an output that could otherwise be easily done better using a simple
bashscript.No because AI replaces a human role.
Code completion does not replace a human role, that’s like saying that spell check is AI.
I am not talking about what it does, I am talking about what it is.
And all tools do tend to replace human labor. For example, tractors replaced many farmhands.
The thing we face nowadays, and this is by no means limited to things like AI, is that less jobs are created by new tools than old destroyed (in my earlier simile, a tractor needs mechanics and such).
The definition of something is entirely disconnected from its usage (mainly).
And just because everyone calls LLMs now AI, there are plenty of scientific literature and things that have been called AI before. As of now, as it boils down all of these are algorithms.
The thing with machine learning is just that it is an algorithm that fine tunes itself (which is often blackbox-ish btw). And strictly speaking LLMs, commonly refered to as AI, are a subclass of ML with new technology.
I make and did not make any statement of the values of that technology or my stance on it
But these tools are not mere algorithms or ML products, they are LLM backed
Emmet has been around since 2015. So it was definitely not LLM backed.
My friend, nobody says all of them are LLM backed, but some are
I’m saying that code completion does not constitute AI and certainly isn’t LLMs.
I then provided an example of why that isn’t the case.
You decided to respond to this by pointing out that some LLM may be involved in some code completion. Although you didn’t provide an example, so who knows if that’s actually true, it seems sort of weird to use in LLM for code completion as it’s completely unnecessary and entirely inefficient, so I kind of doubt it.
I just want to point it out for a minute, because it’s sort of feels like you don’t know this, code completion is basically autocomplete for programmers. It’s doing basic string matching, so that if you type
fncit also completes tofunction(), hardly the stuff of AIhttps://learn.microsoft.com/en-us/visualstudio/ide/visual-studio-github-copilot-extension?view=visualstudio