Based on that extensive (NYT?) article from a few weeks ago, it doesn't look like Mexico will ever be able to pull itself out of this mess. The article suggested that most of the govt and police force was either in bed with the various cartels, or too terrified of them in order to do anything about it. It's a very lucrative arrangement for everybody but the poor bystanders who might become civilian casualties in the endless turf wars.
It feels like the only way to fix this is to tweak the "demand" variable, perhaps through legalization.
IMHO the reason behind the response to Arc is "underwhelming" - because the anticipation has been built for 5 years, but it turns out the implementation is a "just" set of macros. And it certainly don't fit the "iterate rapidly" concept.
Now - this is not a criticism of PG's work - on the contrary, I believe his design matches up with his philosophy in On Lisp very well. Looking past how the process, people now actually start to play with the language and make use of it.
What I'm mostly interested though - are the learning lessons if PG is willing to share them. Why did it take as long as it had? Is it because this is not his day job, or are there additional design changes or problems that he has to solve along the way, etc. Getting a sense of that will shed light for other aspiring developers.
I for one tend to try to shoot for just the right things and have scratched quite a few prototypes so I can go for the perfect one, so would sure love to hear from PG on his experiences.
It feels like the only way to fix this is to tweak the "demand" variable, perhaps through legalization.