PinnedMariusz Kurman@mkurman88Apr 28Say hi to Zorai.zorai@zorai_agentApr 28First light, now live. The dawn agents rise.11663K3K
Mariusz Kurman@mkurman88Mar 8, 2025I have it, guys! What should I ask? Give me some challenging topics to compare with OpenAI Deep Research, preferably tough medical questions.13213260601.8K1.8K547K547K
Mariusz Kurman@mkurman88Aug 17, 2025See it in action. 100 workers; @openrouter, @Alibaba_Qwen Qwen3 235B A22B Instruct 2507.00:00Mariusz Kurman@mkurman88Aug 17, 202533,325 medical questions generated in one and a half hours for $1.73 via @openrouter using Gemma 3 27B.1515222221421481K81K
Mariusz Kurman@mkurman88Sep 19, 2025Replying to @vixhalThey didn't. R1 is still my go-to model for most tasks. Surprisingly, it performs significantly better than others when a strict policy is required, and it hallucinates less frequently than GPT-oss-120b.771120920923K23K
Mariusz Kurman@mkurman88Aug 17, 2025104,335 medical questions generated in 24 minutes (100 workers) for $13.6 via @openrouter using @Alibaba_Qwen Qwen3 235B A22B Instruct 2507.664416616617K17K
Mariusz Kurman@mkurman88Oct 5, 2025Batch size 64 x 8 num_generations, 3B model, and still ~90GB VRAM free. AMD is awesome!116616116132K32K
Mariusz Kurman@mkurman88Aug 17, 202533,325 medical questions generated in one and a half hours for $1.73 via @openrouter using Gemma 3 27B.778814114136K36K
Mariusz Kurman@mkurman88Jan 26, 2025Replying to @Alibaba_QwenJust merged R1 and 1M together! mkurman/Qwen2.5-14B-DeepSeek-R1-1M · Hugging FaceFrom huggingface.co99881301307.1K7.1K
Mariusz Kurman@mkurman88Aug 9, 2025This is a real deal. 5k/day, any model, $20/mo @chutes_ai6615151081086.2K6.2K
Mariusz Kurman@mkurman88Oct 21, 2025AshutoshShrivastava@ai_for_successOct 21, 2025ChatGPT Atlas Agent Mode. LMAO 🤣2244919111K11K
Mariusz Kurman@mkurman88Mar 9, 2025Replying to @PeacerfulResult on my drive: drive.google.comRisk Reduction Using Wasserstein Measure with Stochastic Discount.pdf6611808019K19K