Visualizers

Learn by moving things

Interactive explainers for the ideas behind retrieval, embeddings, and agents. A formula tells you what is true; dragging a vector shows you why. Each one also appears inside the article it belongs to, and runs entirely in your browser, so drag, poke, and break it.

Portrait of Sachin Gupta rendered in binary

Throughput: a live rate lane

From: Sizing a System on a Napkin

Requests per second, made physical. Beads stream past a one-second gate and the flow speed is the rate. Switch between creates (~40/s), clicks (~4,000/s), and the busy-hour peak (~8,000/s) to feel how far apart they are.

Throughput: how many per secondauto
~40writes / sec
1 secondrequests

100M a month, over ~2.6M seconds

~40/s
~4,000/s
~8,000/s

Storage: a disk filling over time

From: Sizing a System on a Napkin

Scrub the months and watch used space creep up in a straight line, links so far times 500 bytes. A 4 TB capacity frame shows the link shortener stays small for years.

Storage: it grows, but slowlyauto
0.00 TB0.0B links
0 months (0 yr)
1y
2y
3y
emptyone modest disk: 4 TB

Each month adds 100M links x 500 bytes. It grows in a straight line, and a single disk lasts years.

Bandwidth: a thin trickle in a fat pipe

From: Sizing a System on a Napkin

Bandwidth is reads times bytes. Two sliders set the flow in MB/s against a 1 Gbps pipe, showing how little the link shortener uses, and what it would take to fill it.

Bandwidth: bytes per second through a pipedrag the sliders
2.0 MB/s2% of a 1 Gbps link
pipe capacity: ~125 MB/s (1 Gbps)the actual flow

Bandwidth is just reads times bytes. At 4,000 reads a second of 500 bytes, that is about 2 MB/s, a thread through a fat pipe. Push the sliders to see what would actually fill it.

Visualizers — Sachin Gupta