PinnedMubashir Rahim·Feb 16The RTX 5080 Is 3x Faster Than NVIDIA’s $3,999 “AI Supercomputer” — And It Doesn’t MatterI benchmarked the DGX Spark against consumer GPUs. The speed gap is brutal. But speed isn’t the whole story.A response icon34A response icon34
PinnedMubashir Rahim·Feb 9We Replaced an RTX 5080 With a $3,999 AI Supercomputer : Here’s WhyA $999 GPU vs. a $3,999 machine. On paper it looks like a bad trade. In practice, it’s not even close.A response icon1A response icon1
Mubashir Rahim·May 2920 Hours With Opus 4.8: Initial Views From Inside Real Agentic WorkI have spent roughly 20 hours with Opus 4.8 over the past few days, around five to six of which were active hands-on work rather than…
Mubashir Rahim·Mar 13GPT-5 scores 98% on Math Benchmarks. It Gets Half of Basic Arithmetic Wrong.I tested 5 frontier LLMs on raw arithmetic. Just numbers. The results should worry anyone using AI for anything involving money.
Mubashir Rahim·Mar 10Apple’s M5 Max Just Made My DGX Spark Benchmark Look OutdatedI benchmarked NVIDIA’s $3,999 AI box last month. Apple’s new laptop chip just changed the conversation. (fair warning xD No pretty graphs…A response icon10A response icon10
Mubashir Rahim·Feb 8I Tested 10 AI Models to Predict Saudi Arabia’s Potato Production.What happens when machine learning models violently disagree with each other?
Mubashir Rahim·Dec 11, 2025When AI Forgets Who It Is: My Claude Had an Identity Crisis (And Calling itself “Perplexity AI”)Welcome to 2025, where even AI models need therapy.
Mubashir Rahim·May 13, 2025The GPU Paradox: Why Your GPU Might Be Slower Than Your CPU for LLM InferenceIntroduction: A Counter-Intuitive Discovery
Mubashir Rahim·Aug 17, 2023Unveiling the Evaluation Matrics: Mastery of Machine Learning Through the Lens of GamingUnveiling the Gamer’s Guide to Mastering Machine Learning Performance. Join us on a journey where gaming prowess meets machine learning…