Blog
Thoughts on AI, ML, and study strategies for postgraduate life.
All Posts
0x Alpha: OpenCode's New Saviour
After DeepSeek's price hike and the collapse of Operation Cheapseek wrecked OpenCode subscriptions, an anonymous 1M-context model named 0x Alpha dropped for free. Here is the backstory, performance analysis, and technical breakdown.
Anthropic Nuked Thrice in a Day
Three Opus-shaped models dropped in a single day — DeepSeek v4 Pro going official, Grok 4.6, and open-weight Qwen 3.8 Max — and Anthropic's premium is getting harder to explain.
Mimo v2.5 Pro: DeepSeek Fan in Me is Impressed
An honest, practical review comparing Xiaomi's MiMo v2.5 Pro and DeepSeek v4 — exploring real-world coding performance, hallucination rates, pricing, and provider limits.
Kimi K3: A justified price jump?
An in-depth review of Moonshot AI's newly released Kimi K3, a 2.8T Mixture-of-Experts model, exploring its architecture, pricing, and performance.
DeepSeek v4 Pro: A 2-Month Review
An honest, hands-on review of DeepSeek v4 Pro after two months of daily use — exploring its performance, cost, strengths, and limitations in real-world workflows.