- Hungry Minds
- Posts
- 🍔🧠 Containers: Netflix's Predictive CPU Isolation
🍔🧠 Containers: Netflix's Predictive CPU Isolation
PLUS: Meta Llama 3 state-of-the-art 🦙, Tinder's API gateway architecture 🛰️, Next-gen Atlas unveiled for commercialization 🤖
Alex Zajac
April 22nd, 2024
In partnership with
Happy Monday! ☀️
Welcome to the 301 new hungry minds who have joined us since last Monday!
If you aren’t subscribed yet, join smart, curious, and hungry folks by subscribing here.
2 new things this week:
I launched Hungry Minds on Substack! A lot of friends in Software Engineering are hanging out there so I want to give it a try. 🤘
This is the first-ever sponsored issue of Hungry Minds! 🥳
Let’s dive right in:
🍔 THIS WEEK’S MENU 🥗
📚 Netflix's predictive CPU isolation for containers. Understand Tinder's API Gateway architecture. Explore advanced retrieval augmented generation.
🗞️ Boston Dynamics commercializes next-gen electric Atlas robot. Meta unveils state-of-the-art Llama 3 models. Adobe develops AI-powered generative video tools.
👨🏻💻 Quick byte: Memory-efficient large model training in PyTorch using gradient checkpointing.
Reading time: 5 minutes
A word from our sponsor:
Free SOC 2 Compliance Checklist from Vanta
Are you building a business? Achieving SOC 2 compliance can help you win bigger deals, enter new markets, and deepen trust with your customers — but it can also cost you real time and money.
Vanta automates up to 90% of the work for SOC 2 (along with other in-demand frameworks like ISO 27001, HIPAA, and GDPR), getting you audit-ready in weeks instead of months and saving you up to 85% of associated costs.
Food for Thought
A mindset, an example, and an action item to start the week
‘Everything you've ever wanted is on the other side of fear.’
Mindset: Embracing fear is the key to unlocking your fullest potential.
Example: SpaceX regularly tackles the fear of failure in aerospace innovation, achieving historic milestones.
Action Item: Identify one fear today and take a small step to confront it directly.
The Rabbit Hole
Deep dives, trends, and resources curated to stay ahead
💾 SIDE DISHES 💾
ARTICLE (rag to riches)
A Comprehensive Tutorial on Advanced Retrieval Augmented Generation
GITHUB REPO (coding interviews)
A complete study plan to prepare for coding interviews
ESSENTIAL (linux)
Build your own Linux system from the source
ARTICLE (sheets in the backend)
How Row Zero stream sheets content from the server at scale
ESSENTIAL (git)
The guide to git I never had
ARTICLE (SAAS architecture)
Why self-hosting is still a great option
TOOL (ohmygit)
An open-source game to learn about git
PAPER (infinite attention)
The paper from folks at Google on infinite attention
The Weekly Digest
Software, AI, and startup news worth your time
Brief: Boston Dynamics unveils a new, all-electric "Atlas" robot, retiring the iconic hydraulic version, showcasing advanced capabilities in a symbolic video farewell.
Brief: Meta introduces the next generation of Llama models, Llama 3, showcasing significant advancements in performance, reasoning, coding, and optimization for real-world scenarios.
Brief: Adobe introduces a cutting-edge AI model for video generation, offering features like object addition and generative extend, while emphasizing metadata tools like Content Credentials for media authenticity.
Brief: Elon Musk's x.AI introduces Grok-1.5 Vision, a cutting-edge multimodal model excelling in real-world understanding and interaction, showcasing progress and potential in AI innovation.
Brief: TikTok unveils TikTok Notes, a new photo-sharing app that mirrors Instagram but with added features like headlines for images above captions and a Pinterest-style homepage layout.
Brief: Microsoft's $1.5B investment in UAE's G42 signifies a strategic move amid escalating U.S.-China tech tensions, positioning the country as a key player in the AI race.
The Quick Byte
One coding tip because you’re technical after all
This week’s coding challenge:
This week’s tip:
When training large neural networks in PyTorch, memory efficiency becomes crucial. Using techniques like gradient checkpointing can significantly reduce memory usage at the cost of additional computation, allowing for the training of larger models or larger batches.
Wen?
Large Model Training: Essential for training models that would otherwise not fit into GPU memory.
Resource-Constrained Environments: Useful when training on hardware with limited memory resources.
Why?
Resource Efficiency: Allows for more efficient use of computational resources by trading off memory usage for compute.
Scalability: Enables scaling up of model complexity and training dataset size without requiring proportional increases in hardware capacity.
Flexibility: Offers a strategic option to manage memory usage during model development and training phases.
The Job Feast
A fresh cheese job board if you’re looking for a change
Burp-A-Laugh
The most important meal of your day
That’s it for today! ☀️
Enjoyed this issue? Send it to your friends here to sign up, or share it on Twitter!
If you want to submit a section to the newsletter or tell us what you think about today’s issue, reply to this email or DM me on Twitter! 🐦
Thanks for spending part of your Monday morning with Hungry Minds.
See you in a week — Alex.
Icons by Icons8.
*I may earn a commission if you get a subscription through the links marked with “aff.” (at no extra cost to you).