TLDR 2022-04-04

Tesla factory footage 🚗, Apple + LG foldable display 💻, limitations of neural networks 🧠

📱
Big Tech & Startups

Apple Collaborating With LG to Develop iPads and MacBooks With Foldable OLED Displays and Ultra-Thin Cover Glass (2 minute read)

Apple and LG are collaborating to develop a foldable OLED display panel with ultra-thin cover glass for future iPad and MacBook models. LG has growing expertise in foldable OLED displays and sells them to other manufacturers such as HP and Lenovo. Apple's display could be around 20-inches in size. It may be creating a new product category for the device. The device won't be launched until after 2025.

Drone footage shows everything going on inside Tesla's Berlin Gigafactory (2 minute read)

Tesla released a video taken by a drone inside the Tesla Gigafactory in Berlin. The video shows the inside of the new factory and was shot by a company called Skynamic. It shows exactly how a Tesla assembly line works. The video is available in the article.
🚀
Science & Futuristic Technology

Israeli bee tech startup Beewise pulls in $80m investment for robotic beehives (4 minute read)

Beewise is an Israeli startup that makes a robotic beehive that can monitor and care for honeybees around the clock. The Beehome uses robotics, artificial intelligence, imaging, a software platform, and a mobile app to look after up to 24 bee colonies at a time. It automatically controls climate and humidity and sends alerts when human intervention is required. The Beehome can even harvest the honey that its bees produce. Beewise recently announced a new version of its Beehome after raising $80 million in funding.

Some AI Systems May Be Impossible to Compute (4 minute read)

There are fundamental theoretical limits to how stable and accurate AI systems can be. While there is mathematical proof that stable and accurate neural networks exist for a wide variety of problems, there may be no algorithms that can successfully compute them. There are limitations to what computers can achieve that can not be overcome with more data, computing power, or time. Knowledge about these limits could inspire more research. The limits imply the existence of a classification theory for describing neural networks that can be computed by algorithms.
💻
Programming, Design & Data Science

Transformers for software engineers (21 minute read)

The Transformer model architecture, originally introduced for machine translation, is now used for a wide range of domains, including text, audio, video, and more. This article reverse engineers the Transformer architecture and uses the language of software engineering to explain how these models work. It is an introduction to the details of the Transformer model for those with a software engineering background, particularly for those looking to do interpretability work.

Yaade (GitHub Repo)

Yaade is a collaborative API development environment. It is self-hosted, secure, and persistent and it features a dark mode default. Requests can be called on localhost as well as on remote servers. Screenshots are available.
🎁
Miscellaneous

Ask HN: What bits of fundamental knowledge are productivity multipliers? (Hacker News Thread)

Last Friday I posted this story on "productivity multipliers" with a broken link, here is the correct link.

Life Advice from NYC Chess Hustlers (13 minute read)

New York's Washington Square Park has hosted many chess hustlers for more than 80 years. Some have made up to $1,700 of tax-free money in a day. The hustlers make money by playing against challengers or by giving lessons. This article features interviews with three chess players from the park where they talk about chess, how they make money playing the game, and life.

How to test a city-wide wireless network from inside a chest drawer (5 minute read)

Thingsquare is a company that builds large Internet of Things solutions. These networks can cover department stores and even entire cities. The company uses sub-GHz radio frequencies with IPv6 mesh networking to create its network. This article discusses how Thingsquare tests its large-scale mesh networks by creating small-scale setups with MAC-level packet filtering and some software.
Get the most important tech, science, & coding news in a free daily email. Read by +1,250,000 software engineers and tech workers.
Join 1,250,000 readers for