- TechTok Newsletter
- Posts
- Cheaper GPT-4o 🧠, Learn From Catastrophes 💡, Worldwide IT Crash Explained 🏛️
Cheaper GPT-4o 🧠, Learn From Catastrophes 💡, Worldwide IT Crash Explained 🏛️
Today's TechToks: Corporate Windows machines outage worldwide, OpenAI releases cheaper & safer GPT, The complexity of problems , Useful Products, Trending GitHub Repos & more!
TechTok by João Martins
July 23, 2024
Today’s Summaries
See all as stories in techtok.today
Today's Picks🍒
💥 Global system outage on 8.5 million Windows devices interrupts banks, hospitals and airlines following CrowdStrike’s Falcon software update
💰 OpenAI releases cost-effective GPT-4o mini
🧠 From MIT Technology Review: The complexity of technical problems and learning from catastrophes
📦 From AWS Engineering Blog: A library of well-architected AWS solutions, ready to deploy
GitHub's Trending Repositories
💭 Memory layer for LLMs
⏱️ Quicken compilation time by caching build artifacts
🦙 Distribute and run LLMs with a single file
🏛️ Open source library of Computer Science books
Products of the Week
📃 Summarize books into 10 minute audios
🧠 Organize and search your bookmarks with LLM
💼 Find jobs or conduct fair interviews with AI
And 6 more products to scale automized messaging, make gif animations from statistics and much more.
🧐 Daily Picks
Most impactful articles and news of this week on techtok.today
A faulty cybersecurity update caused global outage to around 8.5 million devices, less than one percent of all Windows machines. But it affected the corporate clients of CrowdStrike, including airlines, broadcasters, and banks. The issue was in the Falcon software update — not a cyberattack — that made Windows systems inoperable by getting stuck in a boot loop, also causing issues in Red Hat Enterprise Linux last month. Here’s an explanation.
OpenAI released GPT-4o mini with pricing over 60% cheaper than GPT-3.5 Turbo, and instruction hierarchy, which increases safety by prioritizing the instructions in trusted system prompts.
This MIT story navigates three perspectives on the complexity of technological problems, ranging from the concept of cloud problems — complex, unpredictable and often with a social dimension — to describing systems as both structure and behavior. It navigates the books: Wicked Problems, Rational Accidents and Systems Ultra.
MIT Technology Review
✨ "Too long don't README"
GitHub trending repos of today
Mem0 is a self-improving memory layer for LLMs, enabling personalized experiences across applications. It provides multi-level memory retention, adaptive personalization, and a developer-friendly API, allowing you to enhance your applications and streamline development.
What can you use this for? Easily use the same memory across different LLMs, keep details for relationship management, and track your personal health data.
NativeLink is an efficient build cache and remote execution system. It accelerates software compilation and testing by caching build artifacts, and distributing tasks across multiple machines.
What can you use this for? Accelerate build and test processes, parallelize workloads, and ensure consistent build environments.
llamafile lets you distribute and run LLMs with a single file. It turns LLMs into a standalone executable that can be run locally, called “llamafile“.
What can you use this for? Make your fine tuned LLM easily deployable by turning it into a llamafile, prompt private data to local LLMs by using llamafiles.
Repos to learn something new
A list of free programming books and learning resources in multiple languages, featuring programming playgrounds to run code in-browser. You may explore the library here.
What can you use this for? Study programming computer science topics like embedded systems, blockchain, cloud computing, and many other subjects.
Quick Note
Help us reaching more people and get your shout-out in our newsletter 🤩!
Copy and share your unique referral link:
{{rp_refer_url}}
📱 Product Picks
Curated products from Product Hunt and MicroLaunch
Scalable and personalized automation for WhatsApp, Email, and SMS.
📃 SoBrief
Read books in 10 minutes, summarized into 12 listenable key takeaways.
Organize and search your bookmarks with an AI assistant and interactive tools.
AI job search copilot with filtered matches, deep insights, and real-time company trends.
⚖️ Dialog AI
Conduct fair, interactive interviews with real-time AI and custom insights for hiring.
Transform your statistics into captivating animated GIFs.
🧲 Blaze
Finds high intent leads by scanning social posts, engagements and connections with AI.
AI chat service victor over GPT4o in MMLU and HumanEval, eco-friendly (only 0.5W per response).
A structured way to track your journey, reflect on your progress, and stay motivated.
Free Instagram Widget: Embed your Instagram feed on any website.
🤗 Hugging Face
Check this trendy AI model online for free
Provides a transcription of videos with clickable timestamps for each word, with a downloadable json and an in-browser interactive transcription.
What can you use this for? Jump to when a certain word is said in a video, count how many times a word was mentioned, calculate time taken by pauses in dialogue.
👩🏽💻 Engineering Blogs
Articles from engineering blogs of big tech companies
A blog post on AWS Solutions Library: a curation of ready-to-deploy architectures and guides built by AWS and AWS Partners, with over 1,300 solutions separated by industry, technology or organization type.
Popular solutions include Live Streaming on AWS and Guidance for Game Analytics Pipeline on AWS.
Guidance for Game Analytics Pipeline on AWS
