Rlhf helps. The current one is just coming out of someone with dementia just like we went through in the US during bidenlitics. We need to have politics removed from this pipeline

You just need to hear the guy stance on china open models to understand They not the goods guys.

Agentic coding is a card castle built on another card castle (test time compute) built on another card castle (token prediction) the mere fact that using lot of iterations and compute works maybe tells us that nothing is really elegant about the things we craft.

you can be better and decide not to do it because you dgaf

max payne 1 is what triple AAA(-?) gaming should always be.

No you are not. I tried all gemini models. They are slop.

its just a form of structured output. you still need an env to run the code. secure it. maintain it. upgrade it. its some work. easier to build a rule based workflow for simple stuff like this.

for me its clear OpenAI and Anthropic have a lead. I dont buy Gemini 3 being good. it isnt. whatever the benchmark said. same for meta and deepseek.

yeah lol. the interview is braindump on leetcode and sysdesign. two ways to pass it. do a lot of exercise/ learn the patterns or be an excellent programmer. there is 0 doubt he would have a full HIRE loop

The thing is, being an excellent programmer has nothing to do with leetcode and sysdesign (which is actually a back-end CRUD systems) questionnaire.

it has something to do. an excellent programmer knows how to solve leetcode. maybe he never seen the problem but with he would ask a few questions that would help him get to the solution. and sys design is not a back end crud system questionnaire. it depends on the role but could be much more challenging.

this is actually the opposite. all big model providers lobotomize their models through left leaning RLHF