30-second learning · 41 short videos

One engineering idea, played out step by step

Some things are faster to watch than to read. Each short takes one idea — why a rack losing power does not take your shop down, why b = a is not a copy, where Netflix actually streams from — and plays it out a step at a time, most of them in under a minute. Every step is written out under the player too, so you can go back over it at your own pace instead of the player's.

Filter by category

Showing all 41 shorts

MSDEVBUILD Azure Security Secure a Mobile API in 5 Steps 8 steps · 25s

Secure a Mobile API in 5 Steps

In order. The order is most of the value.

  • The five layers a mobile API needs, and what each one actually stops
  • Why authentication and authorization are two separate steps, not one
  • Why monitoring is last but is the step you will be glad you did

Azure · Security 8 steps

MSDEVBUILD Azure A 118-character URL. 15 characters back. 13 steps · 65s

A 118-character URL. 15 characters back.

Designing a link shortener on Azure — and the one digit that decides whether you ever see your click stats.

  • Why a link shortener is really two systems: one write, a thousand reads
  • How the short code is made, and what Cosmos DB does when two people roll the same one
  • Why the redirect must be 302 and not 301, and why Front Door answers most clicks

Azure · Architecture 13 steps

MSDEVBUILD Azure A million taps a second. One tiny message. 14 steps · 70s

A million taps a second. One tiny message.

25 million people tapping 🔥 at the same moment — and not one of those taps is ever delivered to anyone.

  • The live-reactions use case in four steps — what the app actually owes every viewer
  • Why sending every reaction to every viewer is impossible, in one line of maths
  • The five Azure services that count instead — Front Door, Container Apps, Event Hubs, Stream Analytics, Web PubSub

Azure · Architecture 14 steps

MSDEVBUILD Azure One line of code. Seven stops. 13 steps · 64s

One line of code. Seven stops.

What really happens between await http.GetAsync(...) and the JSON landing back in your app.

  • The seven stops every HTTP call to an Azure API makes, and what each one costs
  • Why the first call is slow: 100 of the 160 ms is DNS and the TLS handshake, before your code runs
  • Why new HttpClient() in a loop pays that cost forever, and IHttpClientFactory pays it once

Azure · Architecture 13 steps

MSDEVBUILD Azure The queue was in order. Your workers were not. 12 steps · 50s

The queue was in order. Your workers were not.

A customer paid $120 for an item they removed, and Service Bus did nothing wrong.

  • That FIFO delivery and FIFO processing are different things, and concurrency is what separates them
  • Why SessionId is a partition-key choice, not a yes/no on ordering
  • The one setting that silently gives back the race after you enable sessions

Azure · Messaging 12 steps

MSDEVBUILD Azure You tapped View Product. 12 steps · 50s

You tapped View Product.

Five Azure services answer before the page paints. Here is the order they run in.

  • Why the edge answers first and your API never hears about a cache hit
  • What cache-aside actually does on a Redis miss, including the write-back
  • Why product images take a different path and never reach your API

Azure · Architecture 12 steps

MSDEVBUILD Azure The outage lasted 18 minutes. Your retries lasted 4 seconds. 12 steps · 50s

The outage lasted 18 minutes. Your retries lasted 4 seconds.

720 good orders in the dead-letter queue, and not one of them was poison.

  • Why ten redeliveries take about four seconds, not four minutes
  • That MaxDeliveryCount caps attempts and says nothing at all about time
  • How to build the backoff Service Bus does not give you, with ScheduledEnqueueTimeUtc

Azure · Messaging 12 steps

MSDEVBUILD Azure What is MCP? 8 steps · 27s

What is MCP?

Model Context Protocol — how an AI assistant reaches GitHub, your database, your files and Slack.

  • Why a model on its own cannot reach GitHub, SQL or Slack
  • How an MCP client and server carry one request down and back
  • Why M + N connectors beats M × N bespoke integrations

Azure · AI 8 steps

MSDEVBUILD copilot LLMs have no memory 9 steps · 28s

LLMs have no memory

So how does Copilot know your code? It never remembers. It re-reads, re-ranks and re-sends your context on every single request.

  • Why the model holds nothing between two requests
  • How your editor rebuilds the whole prompt every time
  • What competes for the same fixed token budget

GitHub Copilot 9 steps

Free app · no app store

These are built for a phone

Every short is drawn at full portrait height, the shape a phone already is. Installed, it opens full-bleed with no address bar across the top — and the whole library reads offline.

How it works

Get new posts by email

New technical articles, Azure AI and GitHub Copilot updates, and upcoming events. No spam, unsubscribe anytime.

navigate open