Recommending for Long-Term Member Satisfaction at Netflix (8 minute read)
Netflix optimizes recommendations for long-term member satisfaction by going beyond short-term engagement metrics like clicks. It uses a contextual bandit framework where user actions, like completing a season or giving a thumbs-down, are used to define a proxy reward function. This function aims to capture long-term satisfaction and is iteratively refined through a process called reward engineering. To address the challenge of delayed feedback, Netflix predicts missing feedback using observed data and integrates it into the proxy reward function, allowing for better bandit policy updates.
|
blocking=render: Why would you do that?! (9 minute read)
The "blocking=render" attribute allows developers to explicitly mark resources as render-blocking, meaning the browser will not display the page until the resource is loaded. However, this is generally not recommended, as it can negatively impact performance. The most common use case for "blocking=render" is in client-side A/B testing, where it can prevent a jarring experience for users by making sure that the page is not rendered until all variations are ready.
|
How to make your web page faster before it even loads (13 minute read)
Events that occur before the first byte of data is received by the browser are called pre-TTFB (Time to First Byte) events. The PerformanceNavigationTiming API captures these pre-TTFB events and provides insights into processes like DNS lookup, connection establishment, and TLS negotiation. This article explores optimization strategies for these events, such as caching, DNS configuration, and minimizing request waterfalls.
|
|
The Fragile Future of Open Source (4 minute read)
Free and open source software is facing declining industry support and being negatively affected by developer burnout. FOSS projects may experience decline or collapse as investment shifts away from traditional software and fewer people have the time or resources to contribute.
|
|
RDF Processing with N3.js (GitHub Repo)
N3.js is a powerful JavaScript library for working with RDF data. It offers features for parsing, writing, storing, and reasoning with RDF, making it a valuable tool for developers working with semantic data.
|
|
Building Postcards for “Airbnb” Scale (8 minute read)
Airbnb's Media team developed a system to generate unique, personalized postcards for group trips. The project was complex and involved matching relevant postcards to destinations, having localized text formatting, and optimizing performance for various platforms. The team built a flexible template system, used a destination matching algorithm, and implemented an asynchronous postcard generation workflow for fast delivery.
|
The journey of an internet packet: Exploring networks with traceroute (10 minute read)
This article explains how traceroute works and how it can be used to diagnose network problems. It first explains the journey of an internet packet and how it traverses multiple routers before reaching its destination. It then introduces traceroute, which uses the Time To Live (TTL) field in IP headers to trace the path a packet takes. Traceroute identifies each router along the path by sending ICMP packets with increasing TTL values and listening for Time Exceeded Messages.
|
Scaling Rails & Postgres to Millions of Users at Microsoft: Lessons & Takeaways (44 minute read)
This post contains an interview with Andrew Atkinson, a PostgreSQL expert at StepChange, which discusses his experience scaling Rails applications for millions of users. Atkinson shares valuable lessons learned from his work at Microsoft, where he optimized the infrastructure for Flip, a video discussion app. He highlights the challenges of scaling a Rails application with Postgres, particularly when dealing with heavy read and write operations, and provides practical advice for developers on optimizing queries, identifying bottlenecks, and utilizing tools like PG stats and explain plans. The post encourages developers to leverage ActiveRecord's features like counter caches and sharding to improve performance and scalability.
|
|
GPT Engineer (Website)
GPT Engineer allows you to build web apps, sync with GitHub, and deploy them in one click all with natural language.
|
|
|
Love TLDR? Tell your friends and get rewards!
|
|
Share your referral link below with friends to get free TLDR swag!
|
|
|
|
|
Track your referrals here.
|
|
Want to advertise in TLDR? 📰
|
If your company is interested in reaching an audience of web developers and engineering decision makers, you may want to advertise with us.
If you have any comments or feedback, just respond to this email!
Thanks for reading,
Priyam Mohanty, Jenny Xu & Ceora Ford
|
|
|
|
