If I had one shot at explaining my language I would never advertise it with a non differentiating punchline like "A Language for AI Era" which tells you nothing about the language.
If the title was
"Safescript - A Turing incomplete subset of JavaScript, designed for easy static verification"
I would be more interested. The front page wasted a lot of space on fighting a sandboxing rationale that someone looking for a new programming language probably doesn't even have in their head.
There's also a general fundamental misunderstanding what LLMs are good at, when I see someone reference that in their programming language design. The big thing about LLMs is that they don't get tired and unmotivated and operate at a text level where generating text is low cost, but understanding the unwritten context is much more difficult. This generally means being more explicit about everything on the text level is good even if it bloats up the source code and would be a chore if a human had to do it by hand.
This part probably should have been at the top, but it's at the bottom for whatever reason:
>safescript is not a general-purpose language. You can't write a web server in it or sort a list. There's no recursion, no unbounded loops, no dynamic dispatch. It's a language for writing agent skills that interact with APIs and inputs in a way that can be formally reasoned about.
I don't know if I can trust someone who gets something this basic wrong. You know, writing the introduction at the end...
The landing page definitely needs a BLUF pass, I agree. With that said, I think it’s a really interesting project and probably quite useful for a number of agentic use cases. The fact that it’s a DAG means these programs could be embedded as the leaf nodes in a much larger process diagram.
If I had one shot at explaining my language I would never advertise it with a non differentiating punchline like "A Language for AI Era" which tells you nothing about the language.
If the title was
"Safescript - A Turing incomplete subset of JavaScript, designed for easy static verification"
I would be more interested. The front page wasted a lot of space on fighting a sandboxing rationale that someone looking for a new programming language probably doesn't even have in their head.
There's also a general fundamental misunderstanding what LLMs are good at, when I see someone reference that in their programming language design. The big thing about LLMs is that they don't get tired and unmotivated and operate at a text level where generating text is low cost, but understanding the unwritten context is much more difficult. This generally means being more explicit about everything on the text level is good even if it bloats up the source code and would be a chore if a human had to do it by hand.
This part probably should have been at the top, but it's at the bottom for whatever reason:
>safescript is not a general-purpose language. You can't write a web server in it or sort a list. There's no recursion, no unbounded loops, no dynamic dispatch. It's a language for writing agent skills that interact with APIs and inputs in a way that can be formally reasoned about.
I don't know if I can trust someone who gets something this basic wrong. You know, writing the introduction at the end...
The landing page definitely needs a BLUF pass, I agree. With that said, I think it’s a really interesting project and probably quite useful for a number of agentic use cases. The fact that it’s a DAG means these programs could be embedded as the leaf nodes in a much larger process diagram.