Search the archive

Schedule a demo to learn more

FALSEHOODS PROGRAMMERS BELIEVE ABOUT NULL POINTERS

Null pointer behavior is usually more complex than expected. Optimizations, platform-specific quirks, and the evolution of standards have created numerous edge cases and unexpected behaviors, particularly in C and C++. These misconceptions range from simple assumptions about program crashes to the intricacies of null pointer addresses and pointer provenance. Developers should avoid relying on intuition when it comes to null pointers and check docs diligently instead.

HOW ARE IMAGES REALLY STORED?

This article explains how images are stored, starting with the fundamental concept that pixels are the smallest units of color. Grayscale images store one byte per pixel, while color images use RGB (or RGBA) triplets, which require three or four bytes per pixel. Image formats like GIF use palettes to reduce storage by indexing colors, while PNG employs Deflate compression after applying filters to minimize pixel differences before compression. JPEG achieves high compression ratios through chroma subsampling (reducing color information) and the Discrete Cosine Transform (DCT), which exploit the human eye's lower sensitivity to color detail.

THE MODERN WAY TO WRITE JAVASCRIPT SERVERS

This article advocates for a modern JavaScript server API based on the Request/Response model, contrasting it with Node.js's traditional approach. The new API avoids socket binding, which simplifies testing by allowing server functions to be tested directly with simple object inputs and outputs. Benchmark results show the Request/Response API is substantially faster.

UNEXPECTED BENEFITS OF BUILDING YOUR OWN TOOLS

Building your own custom tools has its benefits. For example, this developer created a simple exploit editor for his game which drastically sped up his workflow and enabled higher-quality content creation through increased iteration. A utility he created for managing embedded devices at his day job eliminated tedious tasks for him.

HELL IS OVERCONFIDENT DEVELOPERS WRITING ENCRYPTION CODE

When devs write their own encryption code, it often leads to security issues. A recent example of a startup's flawed encryption implementation shows this, despite its claim of not "rolling its own crypto." People should use better, readily available cryptographic tools to prevent these recurring mistakes.

ADD F*CKING TO YOUR GOOGLE SEARCHES TO NEUTRALIZE AI SUMMARIES

Including expletives in Google search queries prevents Google's AI-generated summaries from appearing. This method is simpler than previously discovered techniques for disabling AI summaries.

LUME

Lume is a CLI and local API server for creating, managing, and running macOS and Linux virtual machines (VMs) on Apple Silicon using Apple's Virtualization.Framework. It offers a range of commands for VM creation, management (including starting, stopping, and deleting), image pulling, and configuration modification.

EARTHSTAR

Earthstar is a database for private, distributed, offline-first applications. It offers a new specification for building user-owned, connected applications. Earthstar uses a JavaScript library and prioritizes offline functionality, data deletion, and self-hosting.

PAGE ASSIST

Page Assist is an open-source browser extension that enables interaction with local AI models from any webpage. With a sidebar and a ChatGPT-like web UI, it supports various Chromium-based and Firefox browsers and allows users to chat with webpages, ask questions about content, and use features such as default keyboard shortcuts.

EVERYONE KNOWS YOUR LOCATION: TRACKING MYSELF DOWN THROUGH IN-APP ADS

This author investigated the geolocation data leak from Gravy Analytics and found several apps on his phone secretly transmitting his location data without consent. He then attempted to purchase his own location data from data brokers but found the cost too high. The author discovered that even with location services disabled, his location and IP address were still being shared with numerous ad platforms and companies. His investigation revealed that this data is collected and resold through a complex system that involves supply-side platforms like Unity Ads and demand-side platforms like Moloco Ads.

CHATBOT SOFTWARE BEGINS TO FACE FUNDAMENTAL LIMITATIONS

LLMs are showing fundamental limitations in solving complex, multi-step reasoning problems, particularly those requiring compositional tasks. Research shows that LLMs struggle with these tasks because their underlying transformer architecture has inherent mathematical limits on their ability to handle increasingly complex computations. While techniques like chain-of-thought prompting and embedding tricks can improve performance on some tasks, these are essentially workarounds, not solutions to the core limitations.

LIFE IS MORE THAN AN ENGINEERING PROBLEM

Science fiction author Ted Chiang explores his work's influence on scientists and engineers, especially in the context of AI. Chiang's focus stems from philosophical questions, often using speculative scenarios to dramatize them, and he views science fiction as the ideal medium for this exploration. He's skeptical about current AI implementations, arguing that while impressive, they lack genuine understanding and could be misused.

RLHF BOOK

This book provides a gentle introduction to reinforcement learning from human feedback (RLHF), covering its origins, core methods, popular algorithms, and future directions for language model applications.

SNIFFNET

Sniffnet is a free, open-source, cross-platform app for monitoring internet traffic.

THE NEXT CHAPTER IN SWIFT BUILD TECHNOLOGIES

Apple is open-sourcing Swift Build, an extensible build engine used by Xcode, to create a unified, cross-platform build experience for Swift projects.

REACT'S DECLARATIVE MODEL ISN'T PERFECT

Despite React's dominance in UI development, its declarative nature creates challenges for unit testing.

GOOGLE QUIETLY ANNOUNCES ITS NEXT FLAGSHIP AI MODEL

Google revealed its next-gen flagship AI model, Gemini 2.0 Pro Experimental, in a changelog for its Gemini chatbot app. The model was available to Gemini Advanced users beginning Thursday, but Google has since removed the mention of the model from its changelog. The new model provides better factuality and stronger performance for coding and mathematics-related tasks. It is still in early preview and can display unexpected behaviors. The model doesn't have access to real-time information and isn't compatible with some of the app's features.

OPENAI IN TALKS FOR HUGE INVESTMENT ROUND VALUING IT UP TO $300 BILLION

OpenAI is in early talks to raise up to $40 billion in a funding round led by SoftBank. SoftBank, which will invest between $15 billion to $25 billion, is helping assemble investors for the rest of the round. The round would value OpenAI as high as $300 billion - the AI startup was last valued at $157 billion in October. The funding will be used to help OpenAI fulfill its roughly $18 billion commitment to the Stargate project, which will construct new data centers in the US to power OpenAI's technology.

FDA APPROVES NEW CLASS OF PAINKILLER – THE FIRST IN MORE THAN 20 YEARS

The US Food and Drug Administration has approved the use of a novel painkiller called suzetrigine for the treatment of short-term moderate-to-severe pain in adults. The new drug is the first of a new class of analgesics that aren't addictive. It selectively targets a sodium channel critical to pain signaling in the peripheral nervous system, inhibiting pain signals from being sent to the brain. The drug is not an opioid, which means that medical professionals now have an alternative to treating pain that doesn't require using a potentially addictive analgesic.

OPENAI PARTNERS WITH U.S. NATIONAL LABORATORIES ON SCIENTIFIC RESEARCH, NUCLEAR WEAPONS SECURITY

The US National Laboratories will be using OpenAI's AI models for scientific research and nuclear weapons security. Up to 15,000 scientists working at the National Laboratories will be able to access OpenAI's reasoning-focused o1 series. The new partnership will involve scientists using OpenAI's technology to enhance cybersecurity to protect the US power grid, identify new approaches to treating and preventing disease, deepen the understanding of fundamental mathematics and physics, and work on nuclear weapons. OpenAI will also work with Microsoft to deploy one of its models on a supercomputer at Los Alamos National Laboratory.

JAVASCRIPT TEMPORAL IS COMING

JavaScript Temporal is a replacement for the Date object that makes date and time management reliable and predictable. It adds support for time zone and calendar representations and many methods for conversion, comparisons, computations, formatting, and more. JavaScript Temporal is starting to be shipped in experimental releases of browsers, making working with dates and times in JavaScript more simple. This post provides a quick overview of the new JavaScript object along with examples of how to implement it.

TYPEZERO

TypeZero is a zero-configuration TypeScript boilerplate for modern backend development. It features a blazing-fast development workflow with hot reload, enterprise-grade code quality tools, future-proof ESM architecture, and more. TypeZero can help developers build production-ready Node applications in seconds.

GOOGLE OFFERING ‘VOLUNTARY EXIT' FOR EMPLOYEES WORKING ON PIXEL, ANDROID

Google has announced a voluntary exit program for employees working in its Platforms & Devices group, which is responsible for Pixel hardware and Android software. The program applies to US employees in the division - Google has many people around the world working on its products, but the program is only for those stateside. Employees will receive a severance package. The program does not coincide with any product roadmap changes.

APPLE TOPS 1 BILLION SUBSCRIPTIONS, NEARLY $100B IN SERVICES REVENUE IN 2024

Apple's Services division, which includes the App Store, iCloud, Music, TV+, and other subscriptions, had an all-time revenue high of $26.3 billion last quarter. It generated nearly $100 billion in revenue over the past year. The company has over 1 billion subscriptions across its services. Expansions of offerings like Apple Arcade and Fitness+ contributed to the growth.

IS ENGINEERING STRATEGY USEFUL?

There's always an engineering strategy, even if there's nothing written down - strategy can be valuable for organizations that take it seriously.

CS16.CSS

cs16.css is a CSS library based on the Counter-Strike 1.6 UI.

MICROSOFT MAKES OPENAI'S O1 REASONING MODEL FREE FOR ALL COPILOT USERS

Microsoft will now offer Think Deeper, which was previously part of the $20 per month Copilot Pro subscription, for no extra cost to all Copilot users.

MARK ZUCKERBERG TEASES A 2025 RETURN TO ‘OG FACEBOOK'

It is clear that Meta is in desperate need of younger Facebook users to sustain itself for the next generation.

ADVICE FOR A FRIEND WHO WANTS TO START A BLOG

A collection of advice given from one friend to another about blogging.

BOREDOM, CHAOS, & THE EXPERIMENTAL CURVE

The barriers between rationality and experimentation are eroding as bandwidth constraints continue to collapse.

AI2'S NEW MODEL SURPASSES DEEPSEEK V3

AI2's Tulu-3, a 405B parameter open-weight language model, surpasses DeepSeek V3 and even OpenAI's GPT-4o on key benchmarks.

MISTRAL SMALL 3

Mistral has released a very powerful 24B model that achieves strong performance, especially in multilingual data. It is the perfect size for deployment and strength.

FIGURE AI DETAILS PLAN TO IMPROVE HUMANOID ROBOT SAFETY IN THE WORKPLACE

Figure AI is establishing the Center for the Advancement of Humanoid Safety to address gaps in safety for robots in workplaces. Led by former Amazon Robotics safety engineer Rob Gruendel, the initiative will focus on testing and certifying robots to industrial safety standards. The company aims to provide transparency with quarterly updates on testing processes and improvements.

AI-POWERED BIOACOUSTIC MONITORING

Acoupi is an open-source Python framework that simplifies the deployment of AI-based bioacoustic monitoring on low-cost devices. It integrates recording, processing, and real-time messaging.

3D OCCUPANCY PREDICTION

SliceOcc introduces a novel vertical slice representation for 3D semantic occupancy prediction in dense indoor environments. It achieves state-of-the-art performance using an RGB camera-based model.

EXPLAINABLE QUERY OPTIMIZATION

Reqo is a new query optimization model that leverages Bi-GNN and probabilistic ML to improve cost estimation accuracy. It introduces an explainability technique that highlights the contribution of query subgraphs.

products get noticed. Learn more about advertising with TLDR.

BYPASSING LLM GUARDRAILS WITH VIRUS

VIRUS is a method designed for generating adversarial data that can bypass moderation systems and disrupt the safety alignment of large language models.

RIGGING CHATBOT ARENA RANKINGS

Researchers demonstrate that crowdsourced voting on Chatbot Arena can be manipulated to boost or lower model rankings using strategic rigging techniques, impacting the leaderboard's reliability.

QWEN2.5-VL COOKBOOKS

Qwen2.5-VL, an amazing new vision language model, has a companion set of cookbooks that show how to use the model for various different tasks.

ARTIFICIAL INTELLIGENCE IS TRANSFORMING MIDDLE-CLASS JOBS. CAN IT ALSO HELP THE POOR?

Generative AI adoption is surging globally, with 66% of leaders prioritizing AI skills over traditional experience. However, access disparities hinder adoption in developing regions, where only a small percentage can leverage GenAI due to limited digital infrastructure. Addressing infrastructure and education gaps is crucial to prevent AI from widening global inequalities.

A NEW WAY TO TEST AI FOR SENTIENCE: MAKE IT CONFRONT PAIN

Researchers from Google DeepMind and LSE conducted a study using a text-based game to explore AI "sentience," testing LLMs by having them choose between options with varying pain and pleasure associations. Findings revealed some models prioritized avoiding pain over scoring points, suggesting a potential framework for assessing AI consciousness.

AI'S CODING PROMISES, AND OPENAI'S LONGEVITY PUSH

The second wave of AI coding is advancing, allowing models to prototype, test, and debug code, potentially moving developers into more supervisory roles. OpenAI has entered longevity science with a model that designs proteins to transform cells into stem cells, claiming results surpassing human efforts. Cleaner jet fuels from alternative sources are gaining momentum, promising significant emission reductions and prompting industry shifts.

OMI RAISES $2M TO BUILD THE FUTURE OF AI WEARABLES

Omi has raised $2M to develop an AI wearable that enhances mind and productivity.

AI ISN'T VERY GOOD AT HISTORY

A new benchmark, Hist-LLM, revealed that leading LLMs like GPT-4, Llama, and Gemini struggle with high-level historical questions.