you wrote an 87-word email. the point was in the last 6 words. the reader gave up before they got there.

in 1963, Barbara Minto joined McKinsey as their first female MBA. she noticed every consultant buried the answer on the last page. she flipped it.

answer first. reasons second. proof third.

McKinsey made it mandatory. Amazon adopted it. Google teaches it. 40 years later, almost nobody outside these firms has heard of it.

I turned her method into 4 Claude prompts:

→ prompt 1: paste your messy draft, get a clean pyramid back
→ prompt 2: write a 4-sentence opener that makes the reader lean in
→ prompt 3: stress-test your logic for gaps and overlaps
→ prompt 4: a simulated CEO grades if they'd act on it or ignore it

the gap isn't talent. it's a method from 1985 nobody uses.

all 4 prompts in this article. rewrite one email tonight.

CPU vs GPU vs TPU vs NPU vs LPU, explained visually:

5 hardware architectures power AI today.

Each one makes a fundamentally different tradeoff between flexibility, parallelism, and memory access.

> CPU

It is built for general-purpose computing. A few powerful cores handle complex logic, branching, and system-level tasks.

It has deep cache hierarchies and off-chip main memory (DRAM). It's great for operating systems, databases, and decision-heavy code, but not that great for repetitive math like matrix multiplications.

> GPU

Instead of a few powerful cores, GPUs spread work across thousands of smaller cores that all execute the same instruction on different data.

This is why GPUs dominate AI training. The parallelism maps directly to the kind of math neural networks need.

> TPU

They go one step further with specialization.

The core compute unit is a grid of multiply-accumulate (MAC) units where data flows through in a wave pattern.

Weights enter from one side, activations from the other, and partial results propagate without going back to memory each time.

The entire execution is compiler-controlled, not hardware-scheduled. Google designed TPUs specifically for neural network workloads.

> NPU

This is an edge-optimized variant.

The architecture is built around a Neural Compute Engine packed with MAC arrays and on-chip SRAM, but instead of high-bandwidth memory (HBM), NPUs use low-power system memory.

The design goal is to run inference at single-digit watt power budgets, like smartphones, wearables, and IoT devices.

Apple Neural Engine and Intel's NPU follow this pattern.

> LPU (Language Processing Unit)

This is the newest entrant, by Groq.

The architecture removes off-chip memory from the critical path entirely. All weight storage lives in on-chip SRAM.

Execution is fully deterministic and compiler-scheduled, which means zero cache misses and zero runtime scheduling overhead.

The tradeoff is that it provides limited memory per chip, which means you need hundreds of chips linked together to serve a single large model. But the latency advantage is real.

AI compute has evolved from general-purpose flexibility (CPU) to extreme specialization (LPU). Each step trades some level of generality for efficiency.

The visual below maps the internal architecture of all five side by side.

Notice the thread connecting all five. Every generation exists to move data less, because the math was never the hard part. Feeding the math units fast enough is.

The same battle plays out one layer up in software. During LLM inference, a single GPU produces terabytes of KV cache per day, and nearly all of it gets thrown away and recomputed, which is a big reason agent workloads cost what they do.

I wrote a full breakdown of how a disaggregated caching layer fixes this, with up to 14x faster time-to-first-token. The article is quoted below.

You should also check the @lmcache GitHub repo: https://github.com/LMCache/LMCache

(don't forget to star 🌟)

👉 Over to you: Which of these 5 have you actually worked with or deployed on?

If you want to become S-tier with software engineering, star + save these 15 GitHub repos: