Updates and recent posts about DeepSeekMath-V2..

Posts
Description

Activity

@adrian_schmidt started using tool Express , 3 weeks, 5 days ago.

Activity

@adrian_schmidt started using tool AWS Lambda , 3 weeks, 5 days ago.

Activity

@adrian_schmidt started using tool Amazon Web Services , 3 weeks, 5 days ago.

Activity

@adrian_schmidt started using tool Amazon SES , 3 weeks, 5 days ago.

Activity

@adrian_schmidt started using tool Amazon S3 , 3 weeks, 5 days ago.

Activity

@adrian_schmidt started using tool Amazon EC2 , 3 weeks, 5 days ago.

Activity

@adrian_schmidt started using tool Amazon Cloudfront , 3 weeks, 5 days ago.

Activity

@adrian_schmidt started using tool Amazon ALB , 3 weeks, 5 days ago.

Story

@laura_garcia shared a post, 3 weeks, 5 days ago

Software Developer, RELIANOID

𝗕𝗲𝘁𝘁 𝗕𝗿𝗮𝘀𝗶𝗹 𝟮𝟬𝟮𝟲

📍 São Paulo, Brazil 📅 May 5–8, 2026 𝗥𝗘𝗟𝗜𝗔𝗡𝗢𝗜𝗗 is heading to 𝗕𝗲𝘁𝘁 𝗕𝗿𝗮𝘀𝗶𝗹 𝟮𝟬𝟮𝟲 — 𝘵𝘩𝘦 𝘭𝘢𝘳𝘨𝘦𝘴𝘵 𝘌𝘥𝘛𝘦𝘤𝘩 𝘦𝘷𝘦𝘯𝘵 𝘪𝘯 𝘓𝘢𝘵𝘪𝘯 𝘈𝘮𝘦𝘳𝘪𝘤𝘢. 🚀 46,000+ professionals 💡 270+ companies 🌍 One shared goal: transforming education Let’s talk about secure, scalable, and high-performance digital learning. 👉 See you at Expo Cen..

Link

@koukibadr shared a link, 3 weeks, 5 days ago

Mobile Developer, Nventive

LiveData vs StateFlow

LiveData and StateFlow both stream data reactively, but differ in two key ways:

Initialization — LiveData needs no initial value; StateFlow requires one.

Lifecycle — LiveData is lifecycle-aware by default; StateFlow is not, so you need to wrap it in repeatOnLifecycle to avoid memory leaks.

DeepSeekMath-V2 is a state-of-the-art mathematical reasoning model built on the DeepSeek-V3.2-Exp-Base architecture with 685 billion parameters. Unlike conventional math-focused language models that optimize only for correct final answers, DeepSeekMath-V2 introduces a self-verification framework where the model generates, inspects, and validates its own mathematical proofs.

This approach enables rigorous, step-by-step reasoning suitable for theorem proving, scientific research, and domains requiring high-integrity logic. The model is trained through a generation-verification loop involving a dedicated LLM-based verifier and reinforcement learning optimized for proof correctness rather than answer matching.

DeepSeekMath-V2 achieves gold-level scores on IMO 2025 and CMO 2024, along with a groundbreaking 118/120 on the Putnam 2024 contest. Released under the Apache 2.0 license and hosted on Hugging Face, it is fully open source for research and commercial use.