This post from Bill Gates discusses AI and how it will change how people interact with technology. Devices will soon be able to understand natural language and perform tasks based on spoken instructions. This represents a huge shift in how people use computers. AI will upend the software industry and bring about the biggest revolution in computing since the GUI was introduced.
Google allows workers to choose 60% and 80% full-time schedules as well as other variants of part-time work, but as with any other company, there are times when employees work more than 40 hours a week to meet deadlines, cover teammates, or deliver products and services.
While it is possible to find extra work to earn a side income as a software engineer, it may be better to just focus on finding a better job instead.
This article looks at predictions by experts on how many years it will be before transformative AI is built - the estimates vary between 4 to 40 years.
Web Application Firewalls may seem like a good idea at first, but they have horrible performance, are easily bypassed, and have a high false positive rate - this article presents some alternatives.
This Hacker News thread discusses the bear case against OpenAI, comparing the excitement around AI tech to how people jumped on the crypto bandwagon years ago.
CAR T cell therapy is a powerful type of immunotherapy that has begun to revolutionize cancer treatment. It involves engineering a patient's T cells so they recognize and attack cancer cells. New research has shown that disrupting a single gene in CAR T cells can make them more potent and able to fight tumors longer. The approach improved the therapy's effectiveness against multiple cancers in mice.
This repository is an effort to create a similar experience to OpenAI's GPTs. It allows developers to configure over 60 LLMs, prompts, over 100 tools, vector databases, retrieval algorithms, and chat history databases. The project uses LangChain, LangServe, and LangSmith.
This article presents a mental model of LLMs and how their limitations can be overcome. A mental model of LLMs is needed to know where they work well, where they don't, and how best to use them. Most people already have a similar model for Google without realizing it - it's assumed everyone knows how to talk to Google, but that wasn't always the case. LLMs are basically like booksmart Harvard graduates who can Google anything to answer any question you ask them.
OpenAI is working on GPT-5 and it plans to raise more capital from Microsoft to support its development efforts.
Coding as a valuable skill might soon fade from the world.
SpaceX may launch the second Starship orbital test as soon as this Friday. Elon Musk said on X that SpaceX will receive a launch license from the Federal Aviation Administration in time for the launch. The last orbital attempt was almost seven months ago - it ended in a mid-air explosion that led to a months-long federal investigation. The FAA and the US Fish and Wildlife Service have yet to announce the completion of an environmental review for the test. The most critical part of the test flight will be hot staging, where Starship's upper stage separates from its boost and ignites its own engines.
An AI meteorology model from Google DeepMind called GraphCast significantly outperforms conventional weather forecasting methods in predicting global weather conditions up to 10 days in advance. GraphCast outperformed the world's leading conventional system in 90% of 1,380 metrics. It can predict hundreds of weather variables over 10 days at 0.25° resolution globally in under a minute. GraphCast is estimated to be about 1,000 times cheaper in terms of energy consumption compared to traditional methods.
Android developers with personal accounts will have to test their apps with a minimum of 20 people for at least 2 weeks prior to publication.
Microsoft built its own custom AI chip for training large language models. It has also developed Arm-based CPU for cloud workloads. Both chips will arrive in 2024. They could potentially lessen the industry’s reliance on Nvidia. Microsoft is already designing the second-generation versions of these chips. More details about the new chips are available in the article.