We have been working hard on the past two months implementing agents for different tasks. We started with Claude code, as we really liked working with hooks. However being vendor locked and having usage limit problems, we ended up implementing our own "runtime" that keeps the conversation structure while having hooks. Right now it only works with OpenAI but its designed to be able to incorporate Claude / Gemini down the road
We ended up open sourcing that runtime if anyone is interested:
While I like the self reflection from this article, I don't think his methodology adds up (pun intended). First there are two main axis where LLMs can make you more productive: speed & code quality. I think everyone is obsessed about the first one, but its less relevant.
My personal hypothesis is that when using LLMs, you are only faster if you would be doing things like boilerplate code. For the rest, LLMs don't really make you faster but can make your code quality higher, which means better implementation and caching bugs earlier. I am a big fan of giving the diff of a commit to an LLM that has a file MCP so he can search for files in the repo and having it point any mistakes I have made.
Exactly, I had the same issue when I was younger. I thought if I read code I could not understand was because the other dev was a legend. Now its the opposite, I am amazed by code that does its job well, its understandable and has low complexity.
Claude Code is my favorite way to use LLMs for coding.
However I feel what we really need is to have an open source version of it where you can pass any model and also you can compare different models answers.
(Aider and other alternatives really doesn't feel as good to use as Claude Code)
I know this is not what anthropic would want to do as it removes their moat, but as a consumer I just want the best model and not be tied to an ecosystem. (Which I imagine is the largest fear of LLM model providers)
Stars are just a signal. When I am looking at multiple libraries that do the same, I am going to trust more a repo with 200 starts that one with 0. Its not perfect, but I don't have the time to go through the entire codebase and try it out. If the repo works for me I will star it to contribute to the signal.
Where did you get the data that there has never been a profitable one? Not calling you out, but curious of where you are getting this data.
I would expect that there have been multiple nuclear power plants that provide a net positive return, specially on countries like France where 70% of their energy is nuclear.
The comparison isn't even close. TikTok's relationship with the Chinese government is well-documented, not "supposed". They are legally required to share data under China's National Intelligence Law. The Chinese government has also a track record of pushing disinformation and find any way to destabilize Western democracies.
Douyin (The Chinese Tiktok version) limits users under 14 to 40 minutes per day and primarily serves educational content, while TikTok's algorithm outside China optimizes for maximum engagement regardless of content quality or user wellbeing.
US tech companies pursuing profit at the expense of user wellbeing is concerning and deserves its own topic. However, there is a fundamental difference between a profit driven company operating under US legal constraints and oversight, versus a platform forced to serve the strategic interests of a foreign government that keeps acting in bad faith.
We recently applied to the current YC batch and got rejected. Seeing this just made me die a little inside.
We are a small SaaS that has very happy paying customers and a huge market. We solve a boring problem, with boring technologies and we are not the next OpenAI or Stripe. Yet we have easily a 10,000 X potential.
I feel like YC now prioritizes funding things that can be hyped more instead of actually funding things that can be solid software businesses.
I strongly disagree with this statement, Golang is excellent for building backends and just the orchestration of all the services you are calling is better done in a typed/compiled language with great multi threading support and fast compilation times that is exactly what Golang excels at.
While from a first instance, this package seems a bit overkill, I think the idea is interesting and is something that can be improved for Go.
I also felt that Go errors where too bare-bones, so I developed a small package (https://github.com/Vanclief/ez) based on an awesome post that I saw here once. I use this package in all Golang code I touch.
Why should namecheap subsidize the cost? That makes 0 sense and I say it as a namecheap customer. I don't see how its fair for them to cover the cost and I don't want them to do so because I am a happy customer and don't want them to go out of business or lower their product quality.
> The customers of those users absolutely do not want this. If I ever find out a company uses something like this, I'll no longer be doing business with them.
I second this 100%
This for me is a major red flag. It could maybe be mitigated if OP would require customers to run their own instance of cc.dev.
I don't think that its bad for the motivated engineer, its just a trade off between expressiveness and simplicity.
I personally find myself choosing go because I feel I am way more productive with the language. By keeping me constrained, I have less choices to make on how I solve the problem, and focus more on solving the problem.
We ended up open sourcing that runtime if anyone is interested:
https://github.com/Vanclief/agent-composer