781–790 of 793 posts
Earlier quoted context omitted.
You are not allowed to discuss human rights until your government stops its unprovoked war on Iran, issues a full repudiation of the policies that led to it, and makes full reparations to the entire rest of the world whose energy prices you have driven up for the benefit of a handful of wealthy inside traders. Edit: Parent poster wasn't American. I incorrectly assumed their nationality.
Ad hominem is good for the heart, I hear.
Earlier quoted context omitted.
Yeah. IME, LLMs actually introduces a lot of friction when you want to improve the design. They are 100% biased towards the status quo.
Not my experience. It does depend on the scale of the project, but for tens of thousands LOC, it's really good at taking instructions for architectural redesign. LLM will take stuff at face value. If you tell it it's good, it'll gladly abide. Similarly if say it's crap it'll so the complete opposite.
My latest experience was towards a C++ to Rust migration and it failed miserably because the existing codebase enforced patterns I didn't want.
Earlier quoted context omitted.
Not my experience. It does depend on the scale of the project, but for tens of thousands LOC, it's really good at taking instructions for architectural redesign. LLM will take stuff at face value. If you tell it it's good, it'll gladly abide. Similarly if say it's crap it'll so the complete opposite.
Yes, it's good at taking instructions for architectural redesign, but when it's not doing an actual architectural redesign task, it will be heavily biased towards the status-quo, despite AGENTS.md and prompts trying to steer it. My latest experience was towards a C++ to Rust migration and it failed miserably because the existing codebase enforced patterns I didn't want.
It's important to do a first pass cleaning up and preparing code before unleashing agents onto it, similar to how Working Effectively with Legacy Code advocates approaching consolidated codebases. After that the agent will actually pick up the new pattern and start propagating that to the rest of the code.
Earlier quoted context omitted.
> the "open" part of this and K3, etc. is of limited use depending on how/where you can actually use it yourself. ... from servers owned and operated in the US by US providers? Go check out the list of providers on ex openrouter for ex the latest deepseek model. Some of them even fall under the zero data retention agreement.
I'm not saying that nobody can use it or consider it... I'm only saying a lot of people literally cannot/should-not. For example, I'm currently working for a local government office on elections management software... It's a no-go from the get-go for me... AI, in general is not allowed, though a couple groups in the city do have copilot access. Even if it's running on US hardware, without Chinese comms, that doesn't…
Earlier quoted context omitted.
Yes, it's good at taking instructions for architectural redesign, but when it's not doing an actual architectural redesign task, it will be heavily biased towards the status-quo, despite AGENTS.md and prompts trying to steer it. My latest experience was towards a C++ to Rust migration and it failed miserably because the existing codebase enforced patterns I didn't want.
Ah yes, I understand, but that's expected and actually desirable, it's preferable than having inconsistent project patterns across the project introduced by multiple LLMs. Imagine the mayhem. It's important to do a first pass cleaning up and preparing code before unleashing agents onto it, similar to how Working Effectively with Legacy Code advocates approaching consolidated codebases. After that the agent will actua…
> It's important to do a first pass cleaning up and preparing code before unleashing agents onto it
That’s a good idea in general, but the whole point of the rewrite was that Fable wasn’t capable enough of fixing the old codebase which we were using as a reference.
Another problem I had in another project was with having “sample” code purely written for testing being used as the architectural guidance, despite comments and repeated memories asking not to.
In the end, by the way, the solution was to simply blocking access to specific files, in the Claude settings.json.
But still Claude tried to cheat by running ‘cat file.cpp` a few times, so…
Earlier quoted context omitted.
Ah yes, I understand, but that's expected and actually desirable, it's preferable than having inconsistent project patterns across the project introduced by multiple LLMs. Imagine the mayhem. It's important to do a first pass cleaning up and preparing code before unleashing agents onto it, similar to how Working Effectively with Legacy Code advocates approaching consolidated codebases. After that the agent will actua…
I believe it’s desirable as a default, but not when the prompting/steering asks not to. > It's important to do a first pass cleaning up and preparing code before unleashing agents onto it That’s a good idea in general, but the whole point of the rewrite was that Fable wasn’t capable enough of fixing the old codebase which we were using as a reference. Another problem I had in another project was with having “sample”…
TL;DR: Qwen3.8-Max-Preview is already good enough that I cancelled my Anthropic subscription. There are other factors too, but Qwen is on a great path.
As a European, this is choosing between the plague and cholera. Trusting US companies is as troubling as trusting Chinese companies. But I get more bang for the buck from the qwencloud.com Token Plan than from any US AI lab.
Earlier quoted context omitted.
lol the pelican benchmark is basically the only review process I trust at this point. kinda wild that alibaba of all companies is making it hard to give them money tho, you'd think they'd want prominent devs testing their stuff. openrouter usually picks these up pretty fast, hopefully it shows up there soon. You can also just download the qwen app and do this in the chat interface using their MCP tools for local dev.
> kinda wild that alibaba of all companies is making it hard to give them money tho Okay so this is entirely off topic but, are there any options to subscribe to these companies without having a credit card? (I am Brazilian if this makes a difference) It's kind of wild that a Chinese AI lab will require something like.. a credit card with visa or mastercard label, and very little else.
China both in robotics and in AI is currently leapfrogging US innovation, not they are not distilling US models, its an argument only the loosing side adopts to have an excuse, reality is the chinese engineers are very freaking good.
I am from Europe, Denmark to be more precise, and I for sure notice the shift that has heppened the last few years in AI from the Eastern part of the world.
You can try to deny it all you want, but China is leading the AI race currently, and what a freaking brilliant move to OpenSource it, it shatters the business model of team Yankee completely.
I wrote up my experience here: https://dly.to/erfRBMyJhSC TL;DR: Qwen3.8-Max-Preview is already good enough that I cancelled my Anthropic subscription. There are other factors too, but Qwen is on a great path. As a European, this is choosing between the plague and cholera. Trusting US companies is as troubling as trusting Chinese companies. But I get more bang for the buck from the qwencloud.com Token Plan than from…