Keep on to blur preview images; turn off to show them clearly
Seriously speaking, I think Izzat is a culture of a) fabricating prestige by any means (eg childishly lying) and d) defending it primarily by doubling down, sneering and mobbing. Hence izzaters try to“ratio” criticism or some such. It’s largely just premodern magical thinking.
Contra @CEBKCEBKCEBK I think different populations and cultures can all practice these different modes of prestige preservation and lying. But that only makes specific terms more useful. Eg Anglo Chyna hands and vatniks often do stuff similar to izzat. It’s a useful distinction.
Difference between izzat and honor In traditional terms, between “honor culture” and “dignity culture” But words with identical translations have very different connotations. “Izzat” is a crude, intimidation-based mode of honor culture in India, and they *can* do better.
Seriously speaking, I think Izzat is a culture of a) fabricating prestige by any means (eg childishly lying) and d) defending it primarily by doubling down, sneering and mobbing. Hence izzaters try to“ratio” criticism or some such. It’s largely just premodern magical thinking.
Ran my second marathon this AM. > 2y ago, I hated running > 1y ago, ran my first half > 6 mos ago, ran my first marathon + puked for 7 miles Today, I hit a 45 min PR. 8 min/mile would’ve felt impossible not so long ago. But 2000+ miles changes your body! Feeling very proud.
Context on the first marathon. ⬇️ Took months to figure out the issue!
刚看了SWE-Benchverified最新测试结果，MiniMax-M2 成为了得分最高的开放权重大模型！ Minimax M2目前是开源模型里的王者，agent能力很强，但官方说测试耗费了很多token,不过它的长任务处理能力真的很棒，200+步都能稳住。 Deepseek v3.2 reasoning版本紧追其后，价格便宜到感人，就是速度慢了点。如果不急着用，这个性价比确实无敌。100步左右就能达到很好的效果了。 GLM 4.6这次表现很均衡，速度快价格低，性能也不错，算是性价比之王了。跟qwen3 coder 480b a35b水平差不多，但响应快很多。 总体来说现在开源模型进步挺快的，虽然跟Gemini 3 pro、Claude 4.5 Opus这些闭源模型还有差距，但在不断追赶头部商业大模型。 #SWEBench #AI评测 #大模型 #Minimax #DeepSeek #GLM #开源模型 #AI性能 #代码生成
A coder, road bike rider, server fortune teller, electronic waste collector, co-founder of KCORES, ex-director at IllaSoft, KingsoftOffice, Juejin.
Just in case you missed it: A prompt to transcribe a YouTube video (from a URL) or uploaded local videos into a structured, formatted text complete with speaker labels and timestamps.
Prompt Engineer, dedicated to learning and disseminating knowledge about AI, software engineering, and engineering management.
RT @axiommathai: Putnam, the world's hardest college-level math test, ended yesterday 4p PT. Noon today, AxiomProver solved 9/12 problems…
Building @SakanaAILabs 🧠