Full-text search across article titles and summaries. Refine with sector, year, content type, and reading-time filters.
Team trust is vital for the success of any organization because it fosters a safe work environment, promotes open communication, and increases overall productivity. However, trust at work is earned, not given. Leading by example, showing consistency in actions and words, and involving your team in decision-making are some of the strategies to build trust.
This is the ultimate guide on how to think about and define who your competitors are, understanding how much to care about each one, and what to do with them.
This article outlines 8 key metrics that PMs should consider when measuring product success.
Authenticity is about staying true to one’s core values. Even though many companies list authenticity as a key cultural value, it is often interpreted in the corporate context as adhering to the values of the company. This article suggests a three-step approach for navigating the delicate balance of “work-appropriate authenticity”. First, observe behaviors around you. Second, figure out how you feel about these behaviors. Finally, make a decision to either accept it or leave the environment.
Effective writing is crucial for workplace collaboration. It is important to make sure your writing will get your recipient to take action and achieve the outcome you want. “The Robot Voice Method” is recommended for ensuring that your message conveys your tone and intent accurately. This method involves reading your own writing in a bland and machine-like voice to mimic the limitations of written communication so that you can perceive the message with greater clarity and identify ways to improve it.
This post maps out the career journeys of product managers, covering both entry points into the PM role as well as exit points from the PM role.
Data, findings, and insights are three related but distinct concepts. Data refers to unanalyzed user observations. Findings capture patterns among data points. Insights are actionable opportunities based on user research and business context.
How to find satisfaction as a PM.
Lyft transitioned from using Apache Druid to ClickHouse for its sub-second analytics system. The move was driven by several factors, including a steep learning curve and maintenance complexity associated with Druid, cost considerations, and ClickHouse's better alignment with Lyft's infrastructure and engineering expertise. ClickHouse's benefits included simplified infrastructure management, a reduced learning curve for engineers, native data deduplication support, lower costs compared to Druid, and specialized functionalities that made various tasks easier.
Standard Ebooks served over a million ebook downloads with a VPS, all without relying on JavaScript or a database. Instead, it used classic web technologies and PHP in an efficient manner. The ebooks are actually stored on Git repositories and an ebook build process is run on the VPS whenever an ebook is released or updated.
ReactDOM.flushSync is a function that forces React to flush any pending work and update the DOM synchronously. This post is an in-depth look at what React does on updates and how flushSync can solve problems when React updates the UI later than you need it to. The article provides a good distillation of how React updates the DOM even if flushSync isn’t used immediately in a use case.
Cold-blooded software uses boring technology and can usually run without changes for years. Warm-blooded software has more flexibility but requires constant energy to maintain. It’s important to choose the right context for a project.
Managers who continue thinking about work after hours may actually harm their leadership performance. A study found that not taking a break from work in the evening, especially for new managers, drained their mental resources and reduced leadership effectiveness. Leaders who mentally switched off from work the night before felt more recharged and identified more strongly with their role as a leader.
Four WebAssembly (Wasm) developments are anticipated in 2024: Wasm becoming an ideal match for AI workloads, the completion of critical Wasm standards, a shift towards server-side use cases, and the emergence of web applications that can run Wasm on both client and server sides. Wasm isn’t quite production-ready yet, but 2024 may be the year that it makes the transition.
Sponsored
QA Wolf [11] gets web apps to 80% automated end-to-end test coverage in just 4 months. And they include unlimited parallel runs on their infrastructure + 24-hour maintenance and triage at no additional cost.
A collection of libraries and tools aimed at helping developers create adaptive, accessible, and robust user experiences. These libraries include a React implementation of Adobe’s design system (React Spectrum), unstyled React components and hooks (React Aria), and a set of React Hooks for cross-platform state management (React Stately).
A comprehensive list of New Graduate software engineer positions across the industry.
Jan is an open-source ChatGPT alternative that runs offline on your computer. It runs on any hardware.
US software companies, particularly smaller ones, are facing unexpected high tax bills due to changes in Section 174 of the US tax law. This change requires R&D costs, including software engineer labor, to be capitalized and amortized over 5 years (or 15 years if labor is done outside the US). As a result, companies have to pay taxes on profits they technically haven't realized yet, leading to layoffs and reduced hiring of software engineers. Large tech companies are able to manage the tax increase better than startups.
Design to code is the process of automating the conversion of visual designs into clean, functional code that matches the design's layout, styling, and interactions. It involves collaboration between designers who create visual designs and developers who translate those designs into code using languages like HTML, CSS, and JavaScript. New AI tools aim to automate the manual collaboration process.
Implicit returns and if-expressions can lead to more concise and readable code. This post provides examples in Rust, Ruby, and other languages. On top of implicit returns, block expressions and single expression functions are also useful to further simplify code.
On December 29th, a package named "everything" was released to the NPM registry that installed all other public packages. This resulted in an extensive network of dependencies across the entire registry that made it impossible to unpublish any package on the site.
Reading books on soft skills and leadership is useful in the workplace as both a dev and a manager. To read books with limited time, first start with book summaries to evaluate relevance, then read selectively from the actual book while taking notes. Finally, connect and apply the lessons to your own personal experiences.
The One Billion Row Challenge (1BRC) explores how far modern Java can be pushed for aggregating one billion rows from a text file.
Go is an "80%-language," meaning it can achieve 80% of the performance and functionality of other languages with only 20% of the effort.
Sponsored
Chasing down assets and collecting screenshots for a SOC 2 audit gets tedious. Grow faster and build trust with your customers with the leader in compliance automation [5]:
The latest rumors suggest that the Apple Vision Pro is launching by the end of February. Recently leaked videos show how much multitasking users can expect to accomplish with Apple's Safari web browser inside of visionOS. Safari on the Apple Vision Pro lets users place multiple windows and tabs in 3D space inside of the headset, creating a fully customizable work experience. The leaked videos are available in the article.
Meta plans to possibly unveil an AR glasses prototype in 2024 that it claims is the most advanced piece of technology on the planet in its domain. The device is still prohibitively expensive to manufacture and there is still a lot of work to be done to reduce the price and form factor. Meta may be working on a version of the Ray-Ban Meta smartglasses with a small viewfinder display. AI breakthroughs and the integration of AI into the current Ray-Ban Meta smartglasses have significantly affected Meta's product roadmap.
Chinese labs working on replicating LK-99 findings have discovered a superconductor that works at -23 C. While it is not exactly a room-temperature conductor, it is still a significant discovery. Similar materials have been discovered previously, but they require high pressure to exhibit their superconducting properties. The labs' results have already been replicated.
Mobile ALOHA is a system for imitating mobile manipulation tasks that are bimanual and require whole-body control. It is able to autonomously complete complex mobile manipulation tasks such as cooking, opening doors, and much more. This thread contains a video showing a robot with Mobile ALOHA cooking a Cantonese recipe. Mobile ALOHA is low-cost and open-source.
The AHA stack combines Astro, htmx, and Alpine.js to create modern web applications with a simple set of mental models and workflows. It features a thin client, a simple yet awesome backend, a lightweight frontend, and a declarative approach. The AHA stack is minimalist and uses as few dependencies as possible.
Caching involves storing temporary data for quick access. Eviction algorithms determine what data to keep and what to discard. SIEVE is a simple and effective cache eviction algorithm with just one queue. It can also serve as a cache primitive for designing more advanced eviction policies. SIEVE is not scan-resistant, so it can sometimes show a higher miss ratio than other tools.
The protocol force encourages the layer between consumers of value and producers of value to be as thin as possible. IP, HTTP, and HTML are great examples of this effect. The protocol force minimizes the cost of reaching as many consumers or producers as possible. This force is present in any layered stack that has producers and consumers as outer layers. One of the key challenges of designing versatile tools is that this middle layer is constantly tested and often broken.
A lot of innovation will be happening in 2024. AI and web3 are creating a more intelligent, resilient, and decentralized internet and new energy technologies will power our lives in new and sustainable ways. This article looks at the technologies that will be developed this year and opportunities for the teams building these new technologies. 2024 will be a terrific year for tech.
Most business leaders are confident with a hybrid work strategy.
The Samsung Galaxy S24 will have a new design with flat metal sides and some models will feature a titanium body.
The tech could be used to advance quantum computing and is a potential replacement for silicon.
App stores opening up to competing browsers will enable competitors to replace app stores with directories of Progressive Web Apps.
The endurance test determined that the batteries will be able to power electric vehicles that can drive 500,000 kilometers with virtually no loss of range.
Blue Skuti is the first human to ever force Tetris into a kill screen on the NES classic.
Nikon, Sony Group, and Canon are developing camera technology to embed digital signatures in images, allowing differentiation from AI-generated fakes. Nikon's mirrorless cameras will include tamper-resistant digital signatures for professionals, providing details like date, time, and location. Sony plans to add an online tool called Verify to its mirrorless SLR cameras in spring 2024 to check images for digital signatures, highlighting any AI or tampered content. Canon will introduce similar technology for images and videos by 2024. These advancements address the growing challenge of distinguishing real images from sophisticated AI-generated fakes.
Intel is forming a new independent company called Articul8 AI focused on artificial intelligence software, with investment from DigitalBridge Group and other investors. Articul8, stemming from Intel's work with Boston Consulting Group, will have an independent board and will not be publicly traded, will focus on creating cost-effective, secure AI solutions for business use.
One of the big drawbacks of current language models is that you need extremely expensive human preference data to make them work well. A big research direction recently has been figuring out if we can have the language models perform self-play to improve without collecting this data. SPIN is a new method that takes a step in that direction with just SFT data - it can dramatically improve a base model’s performance across various tasks.
Finding contours and boundaries in images is a classic computer vision problem. However, when introducing noise, quality shifts, or out-of-distribution examples, many existing methods fail to perform well. This new learned method performs extremely well on sensor readings and is only 207k parameters. It uses a two-stage training process and greatly improves state of the art.
This research leverages bracketing photography and a novel temporally modulated recurrent network (TMRNet) to significantly improve photo quality in low-light conditions. By training with synthetic data and adapting to real-world images, this approach outperforms existing multi-image processing methods.
GPT-fast is a great push from the Pytorch team to speed up vanilla models. It can achieve an impressive 218 tokens per second across 8 GPUs.