- TechTok Newsletter
- Posts
- Google's Biggest Acquisition 🏛️, OpenAI Co-Founder's New Company 🧠, New Code and Math LLMs 🐍
Google's Biggest Acquisition 🏛️, OpenAI Co-Founder's New Company 🧠, New Code and Math LLMs 🐍
Today's TechToks: Google's $23 bi acquirement, IOS 18 open beta, Andrej Karpathy's Eureka Labs, Codestral & Mathstral, useful AI products & repos, and more!
TechTok by João Martins
July 17, 2024
Today’s Summaries
See all as stories in techtok.today
Today's Picks🍒
🐍 Mistral releases language models specialized in code generation, math and scientific discovery.
🧠 OpenAI Co-Founder Andrej Karpathy announces new company for AI native education.
🍎 iOS 18 is in public beta featuring Apple Intelligence’s AI tools.
🏛️ Google is close to its biggest acquisition ever: Wiz cybersecurity for $23 billion.
🌀 The first generative feature film: a different movie at each screening.
✨ From Engineering blogs: Building and scaling Notion’s data lake.
▶️ From HuggingFace spaces: Math-focused language model.
GitHub's Trending Repositories
🕸️ Web scraper with persistent storage.
👀 A library for computer vision.
💠 The open source to scaling an Ethereum ecosystem.
👋 A course on algorithms and data structures.
Trending Products of the Week
🐝 Ditch multiple AI tool subscriptions for a single one.
📄 Pre-built multi-step Webflow forms within a click.
🚀 Predict customer behavior with ML.
And 7 more products to track website loading speed, extract data, automate custom support, and much more.
🧐 Daily Picks
Most impactful articles and news of this week on techtok.today
Mistral has launched two new language models - Codestral Mamba 7B, a fast and efficient code generation model, and Mathstral 7B, a specialized model for math-related reasoning and scientific discovery. These are State Space Models, alternate to Transformer models (like GPT, Claude and Gemini). Here’s a comparison to other code-specific models:
Former director at Tesla and co-founder of OpenAI Andrej Karpathy announced a company for AI-integrated education, Eureka Labs. The first course will be an undergraduate-level class on training AI, LLM101n.
Google is in advanced talks to acquire cloud cybersecurity startup Wiz, currently offering $23 billion. The acquisition is likely to bring further antitrust scrutiny from US and EU regulators.
The new iOS 18 with Apple Intelligence is now in public beta, with new features including AI-powered photo editing and writing assistance.
The first generative feature film is a documentary that selects different footage to show at each screening, picking from 530 hours and making possibly 52 quintillion (1018) different versions of ‘Eno’.
✨ "Too long don't README"
GitHub trending repos of today
Crawlee is a web scraping and browser automation library for Python and Node.js. It scrapes data and persistently stores it in machine-readable formats, providing an unified interface for HTTP and headless browser crawling and automatic parallel crawling.
What can you use this for? Crawl for links, extract data from websites, and automate browser interactions with JavaScript-heavy websites. Here are quick examples from the docs.
Supervision is a Python library for reusable computer vision tools, including functionality to load datasets, run models, and annotate images or videos. Model-agnostic and integrated with popular libraries (Ultralytics, Transformers, MMDetectionIt) with customizable annotators.
What can you use this for? Estimate how long an athlete keeps the ball, how fast they kick it, and wait time analysis of people in line. Here’s how to get started at the docs.
Optimism is an open-source project that scales Ethereum using an L2 blockchain, with the aim of building effective decentralized economies. It provides the OP Stack, OP Mainnet, and Base.
What can you use this for? Build decentralized Apps on Ethereum, explore and modify the codebase, and participate in Optimism's governance. Tutorials here at the docs.
Repos to learn something new
Hello Algo is an open-source data structures and algorithms course. It has animated illustrations and runnable code examples on the fundamentals of data structures and algorithms.
What can you use this for: Learn and review data structures with examples in Python, C++, Java, C#, Go, Swift, JS, TS, Dart, Rust, C, Kotlin, Zig. Here’s the english website.
Quick Note
Help us reaching more people and get your shout-out in our newsletter 🤩!
Copy and share your unique referral link:
https://newsletter.techtok.today/subscribe?ref=PLACEHOLDER
📱 Product Picks
Curated products from Product Hunt and MicroLaunch
Website monitoring to continuously test and track loading speed.
Predict customer behavior on your site with ML, save on ads.
Pre-built multi-step Webflow forms within a click.
Ditch multiple subscriptions with 50+ versatile AI solutions in one tool.
Think Notion, but for analytics: pull data from 150+ sources with or without code.
Redesign websites in seconds with AI.
Instant, automated support by AI Agents trained on internal knowledge.
🎮 Shram
Gamified work management.
🌯 Lekko
FAANG-class dynamic config for all: wraps functions in code to make them runtime configurable by users across the company.
🏦 Ceedar
AI accountant for your startup.
🤗 Hugging Face - Space of the Week
Check this trendy AI model online for free
🏆 Math Olympiad Solver (NuminaMath 7B TIR)
The demo for NuminaMath: language models trained to solve math problems using tool-integrated reasoning (TIR). The 6.96 billion parameter model won the first progress prize of the AI Math Olympiad (AIMO), with a score of 29/50 on public and private tests sets.
What can you use this for: Calculate finance management outcomes, choose a workout plan by calculating gains, get a general algorithm from asking a specific question.
All of these use cases can be best if you inform details, but the language model does have knowledge like ‘how much protein for muscle growth by kilogram’ (it considered 0.56 - 0.84 grams per kilogram).
What would you like to see more often on TechTok? |
👩🏽💻 Engineering Blogs
Articles from engineering blogs of big tech companies
In the past three years, Notion has managed a tenfold increase in data volume, doubling every 6-12 months, inciting them to build and scale a data lake. In this blog post, they detailed the tech stack and 5 design decisions taken for it, resulting in: less money, more time, and stronger infrastructure for AI.
🧠 Learn to Use AI Tools (Sponsored)
AI & ChatGPT Mini Crash Course - Eliminate workplace burnout & save 16+ hours every week. Learn 20+ AI tools, prompting techniques & hacks for free.