In Chinese we call it "wash the ground with missiles", and many Chinese and Taiwanese say this constantly as a threat or concern. It's amusing that many people think this is a practical thing to happen, and keep bringing it up.
My experience is that Codex's auto review is extremely costly, with $20 on both sides, I can run CC with auto mode for longer than with Codex's auto review enabled. Also in my own experience Claude's usage is actually bigger than Codex, but I am not sure if that's due to I stick to 5.5 with Codex while keep Sonnet as the default to orchestrate other models in CC.
This kind of services that requires the user to share a seed/code to the recipient always seems kinda awkward to me. The code is not simple/short enough to be verbally communicated; If I can send the code, I usually can just send the file.
This does kind of smell like the wrong way to use it. Not trying to self-promote here, but the experiences you shared really made me think I headed the right direction with my prompting framework ("projex" - I once made a post about it).
I straight up skip all the memory thing provided by harnesses or plugins. Most of my thread is just plan, execute, close - Each naturally produce a file - either a plan to execute, a execution log, a post-work walkthrough, and is also useful as memory and future reference.
In my prompting framework I have a workflow that the agent would scan all the artifacts in my closed/ folder and create a yyyymmdd-archive artifact which records all artifact name and their summaries, then just delete them. Since the framework is deeply integrated with git, the artifact can be digged up from git history via the recorded names.
This is a bit concerning. Did you skip everything besides "yogurt delivery", or you don't agree someone talking to you regularly is counter-loneliness?
The question labeling a whole ethnic can't understand English rubbed me the wrong way, that's about it. This is a much better comment for understanding your rationale.
The points you brought up all are valid. Lexer, parser and general concepts are not language-specific, yes, and I wasn't talking about how the implementation is different.
When I said "you can tell they sometimes get confused and have trouble to comply to the foreign language spec and design", I was thinking about the many times they just fail to write in my language even when provided will full language specs. LLMs don't "think" and boilerplate is easy for LLMs because highly similar syntax structure even identical code exist in their training data, they are kind of just copying stuff. But that doesn't work that well when they are tasked to write in a original language that is... too creative.
I totally agree, and I was fully aware of how common people make language for fun when I replied.
But I feel like the rationale would still stands: Considering LLMs' natures, common boilerplate tasks are easy because they can kind of just "decompress" from training data. But for a new language design, unless the language is almost identical to some other captured by the model, "decompression" would just fail.
In the past 2 months I've been using all the SOTA models to help me design a new DSL for narrative scripting (such as game story telling) and a c# runtime implementation o the script player engine.
The language spec and design is about 95% authored by me up to this point; I have the LLMs work on the 2nd layer: the implementation specs/guidelines and the 3rd layer: concrete c# implementation.
Since it's a new language, I consider it's somewhat new/novel tasks for LLMs (at least, not like boilerplate stuff like HTTP API or CRUD service). I'd say, these LLMs have been very helpful - you can tell they sometimes get confused and have trouble to comply to the foreign language spec and design - but they are mostly smart enough to carry out the objectives, and they get better and better after the project got on track and has plenty of files/resources to read and reference.
And I'd also say "prompt better" is a important factor, just much more nuanced/complicated. I started with 0 experience with LLM agents and have learned a lot about how to tame them, and developed a protocol to collaborate with agents, these all comes from countless trial and errors, but in the end get boiled down to "prompt better".
I'll just disagree with an example: Codex on Windows.
They are known to be very inefficient using only Powershell to interact with files, unless put in WSL. They tend to make mistakes and have to retry with different commands.
Another example is Serena. I knew about it since the first day I tried out MCP but didn't appreciate it, but tried it out again on IDEs recently showed impressive result; the symbolic tools are very efficient and helps the agents a lot.
In Chinese we call it "wash the ground with missiles", and many Chinese and Taiwanese say this constantly as a threat or concern. It's amusing that many people think this is a practical thing to happen, and keep bringing it up.