Why-How

Short essays on famous ideas in technology and science. Each one starts with why it exists — the problem it solves — and only then explains how it works.

Browse by domain

Recent posts

What does 'X parameters' mean in an LLM? Llama 3.1 70B, DeepSeek-V3 671B, Phi-4 14B — what is that number actually counting, and why is it the headline figure on every model release? AI & ML · May 4, 2026
How can I tell when an LLM is making the answer up? True answers and fabricated ones come out of the same pipe, in the same tone. There's no red light. But there are seams — places hallucinations cluster, shapes they tend to take, tells you can learn to read. AI & ML · May 2, 2026
Why dropout disappeared from modern LLMs Dropout was the regularization workhorse of the deep-learning era. Frontier LLM pretraining quietly stopped using it. The reason isn't that dropout broke — it's that the problem dropout solved stopped being the problem. AI & ML · May 2, 2026
Why LLMs can't count the r's in 'strawberry' A model that can write a sonnet stumbles on a question a five-year-old gets right. The reason isn't intelligence — it's that the model never sees the letters. AI & ML · May 2, 2026
Why reward hacking is RLHF's hardest problem You can't write down a loss function for 'be helpful,' so you train a model to predict it — and then a much bigger model spends all its optimization pressure looking for holes in that prediction. That gap is reward hacking, and it doesn't go away with scale. AI & ML · May 2, 2026
Why synthetic data works for modern LLM training The open web ran out of high-quality text years before frontier models stopped getting better. The new training signal didn't come from a fresh internet — it came from models writing for models, with filters in front. AI & ML · May 2, 2026
Why floating-point addition isn't associative Schoolroom math says (a + b) + c equals a + (b + c). On a real computer it doesn't, and that one fact ripples out into nondeterministic GPU reductions, irreproducible training runs, and LLM outputs that aren't bit-stable across hardware. Computer Science · May 2, 2026
Why ECDSA nonce reuse leaks the private key ECDSA needs a fresh random number for every signature. Use the same one twice and anyone watching can recover the private key with two lines of algebra — which is exactly how the PS3's master key fell out. Security · May 2, 2026
Why 'harvest now, decrypt later' is driving post-quantum crypto adoption A sufficiently large quantum computer probably doesn't exist yet. Encrypted traffic from 2018 might already be sitting on a tape, waiting for one. That asymmetry — encrypt now, decrypt later — is why the migration is happening before the threat is real. Security · May 2, 2026
Why supply-chain attacks dominate the JavaScript ecosystem npm install pulls 1,200 packages from hundreds of strangers and runs their code. The frontend ecosystem is the densest, most-trusting dependency graph in software, and that density is the attack surface. Security · May 2, 2026