On the D Blog: Language Plasticity is More Important Than Ever

[monkyyy]

On Monday, 10 August 2026 at 22:35:46 UTC, Adam Wilson wrote:
> On Monday, 10 August 2026 at 13:41:59 UTC, monkyyy wrote:
>> If that where the numbers come from, they are suggesting 
>> deleting whitespace; grouping 4-spaces as 1 token is a "gpt-2" 
>> level technology.
>
> There has been a lot of work done at the harness and ingestion 
> layers to reduce input token usage precisely because 
> whitespace-sensitive languages are so token hungry. So yes, 
> they've gotten the input token usage down, but by improving 
> tooling, not because the language itself is less token 
> expensive. And there is evidence, albeit in early stages, that 
> such schemes are increasing the amount of mistakes that the LLM 
> makes on output and thus creating more iterations over the 
> problem. Which increases total token usage.

Python fucking sucks and ai labs are stupid for putting do much 
python context into their magic strings. But your comparing d to 
a strawman, 4 token per indent python as far as I know never 
existed as 2019 chatgpt would`nt`ve; and they could go further 
and encode it with the exactly the same tokens `new line; tab++` 
`new line; tab--` they havnt yet, but they can at which point it 
would compete with `{` with removed white space

If they put in "allot" of work adding in their `8 space` token 
taking years, they are shit at it(they are shit engineers so 
totally possible); what they put allot of work is changing the 
names every year and marketing a new way to concat strings.

This however is not nearly good enough to make d competitive; 
with every possible language. *Ive never seen a good line of 
template code from an ai*, the main reason to use d; purely 
hypothetically someone could be well positioned to take my 
algorithms and data structure code fragments and spend tokens 
making them cohesive into a nice neat package maybe someone could 
figure out how to get good template code out of ai with such an 
auto research task.