Search the archive

16 TIPS TO LANDING CLIENTS ON LINKEDIN

A cheat sheet for turning LinkedIn into a client engine with offers and messaging that convert.

you want

HOW WE DECREASED GITLAB REPO BACKUP TIMES FROM 48 HOURS TO 41 MINUTES

GitLab reduced its GitLab repository backup times from 48 hours to 41 minutes by identifying and fixing a performance bottleneck in a 15-year-old Git function. The issue was from an inefficient O(N²) algorithm used to handle duplicate references within the `git bundle create` command. GitLab replaced the nested loops with a map data structure, leading to a 6x performance improvement.

A JAVASCRIPT DEVELOPER'S GUIDE TO GO

This is a guide for JavaScript devs transitioning to Go. It covers fundamental concepts like compilation, packages, variables, structs, error handling, and concurrency, comparing them to their JavaScript equivalents. It also goes over important Go-specific concepts, like its explicit error handling and its powerful concurrency model using goroutines.

SEARCH PARAMS ARE STATE

This article discusses the common issue of treating URL search parameters as an afterthought in web development, leading to inconsistent data handling and potential bugs. TanStack Router integrates search parameter schemas directly into the routing layer. This allows URL search parameters to be treated as state, making them easily shareable and resistant to reloads. The approach also allows for type-safe validation, inference, and coordination of search parameters across different routes and components.

FIELD NOTES FROM SHIPPING REAL CODE WITH CLAUDE

There are three effective modes of AI collaboration: AI as First-Drafter, Pair-Programmer, and Validator, highlighting the shift from writing code to editing and directing AI. To use AI effectively as a collaborator, it's useful to have infrastructure components like CLAUDE.md files and anchor comments, along with git workflows, to manage AI-generated code.

WHAT "WORKING" MEANS IN THE ERA OF AI APPS

AI startups are achieving faster revenue and quicker funding rounds than pre-AI startups. Both enterprise and consumer AI companies are showing impressive growth, with consumer companies now generating significant revenue compared to pre-AI.

SELF-HOST & TECH INDEPENDENCE: THE JOY OF BUILDING YOUR OWN

Devs should own their domain and host their own content to maintain control and build long-term value. This dev self-hosts various services like their blog, second brain, and home server.

ODYC.JS

Odyc.js is a small JavaScript library designed to make video game creation accessible even without prior programming experience.

BEACHPATROL

Beachpatrol is a CLI tool that automates and replaces your everyday web browser using Playwright scripts. It lets users control a Chromium or Firefox browser with the command line, integrating browser tasks into custom workflows and OS scripts.

WHAT WAS RADIANT AI, ANYWAY?

This a deep dive into Bethesda's Radiant AI system, exploring its origins, promises, and realities in the Elder Scrolls IV: Oblivion. The article traces the evolution of Radiant AI from Morrowind through subsequent Bethesda titles like Skyrim and Starfield, detailing how it was intended to create a dynamic and living world through non-scripted NPC behavior. While the system never fully achieved its initial ambition, it laid the groundwork for later AI advancements in Bethesda games and is still used in current versions of the engine.

THE HIDDEN TIME BOMB IN THE TAX CODE THAT'S FUELING MASS TECH LAYOFFS

A change to Section 174 of the US tax code, implemented as part of the 2017 Tax Cuts and Jobs Act, disallowed companies from immediately deducting research and development expenses, requiring them to amortize them over several years. This shift made R&D more expensive, impacting cash-strapped tech companies and contributing to mass layoffs that were often blamed on other factors like over-hiring or AI.

I READ ALL OF CLOUDFLARE'S CLAUDE-GENERATED COMMITS

Cloudflare's open-sourced OAuth 2.1 library, which was primarily written by Claude, has a commit history with prompts. Some patterns in the creation of it include prompt-by-example and iterative refinement with specific feedback. Prompts could be treated as part of the source code in the future, allowing for easy regeneration of codebases with improved models.

AI CODING AGENTS

A performance comparison of AI coding agents like GitHub Copilot, OpenAI Codex, Cursor Agents, Devin, and Codegen that shows their success rates in generating merged pull requests - OpenAI Codex currently leads at 83.3%.

FOCUS AND CONTEXT AND LLMS

While LLMs can complete complex coding tasks with sufficient context and algorithmic supervision, the current hype around "agentic coding" as a hands-off solution is misplaced, as these models require intense management and context curation from skilled engineers for valuable results.

FALSEHOODS PROGRAMMERS BELIEVE ABOUT AVIATION

Programmers make numerous incorrect assumptions about aviation data.

REACT NATIVE TEAM OFFICIALLY FREEZES THE LEGACY ARCHITECTURE

The new React Native architecture was promoted to stable in October 2024 for all projects.

Watch the recording

AMAZON ‘TESTING HUMANOID ROBOTS TO DELIVER PACKAGES'

Amazon is building a 'humanoid park' to test robots that could eventually take the jobs of delivery workers. The indoor obstacle course is about the size of a coffee shop. The company is developing artificial intelligence software that can power the robots. It will use hardware developed by other companies. Amazon is aiming for the robots to be able to travel in its Rivian vans and make deliveries from them.

GOOGLE STARTS TESTING ‘SEARCH LIVE' IN AI MODE

Google has started tests of its AI Mode's new 'Search Live' experience, which lets users have real-time conversations with Google. The feature can answer questions while surfacing the sites used to inform the answer. It may ask clarifying questions to refine queries, and users can ask follow-up questions. The feature is currently rolling out to the Google app in Android and iOS.

LONGEVITY IS NOW A FACTOR WHEN PICKING AN EMBRYO FOR IVF

Nucleus Genomics is a DNA testing and analysis company that plans to provide DNA testing for embryos to predict and reduce chronic diseases. It will analyze up to 20 embryos for $5,999, providing an analysis of up to 900 conditions. The embryos will be given probabilities for the likelihood they will get these chronic conditions and then it will be up to the parents to decide which embryos to use. Nucleus also provides possibility assessments for attributes such as IQ, height, and eye color.

WALMART AND WING EXPAND DRONE DELIVERY TO FIVE MORE US CITIES

Wing and Walmart have announced plans to roll out drone delivery to more than 100 Walmart stores in Atlanta, Charlotte, Houston, Orlando, Tampa, and the Dallas-Fort Worth area. Walmart appears to be becoming confident in drone delivery - the expansion is a nearly five-fold increase in Wing's operations with the retailer. Wing is now decidedly out of the pilot and trial phase and scaling up its business.

Read the Delve case study.

SHOW HN: IOS SCREEN TIME FROM A REST API (HACKER NEWS THREAD)

The founders of Clearspace, a startup that builds tools to help people reduce their screen time, have released an API that makes iOS Screen Time accessible on the web. The API allows developers to build tools that leverage screen time data. Example use cases include apps that auto-report large deviations in screen time, show group visibility, and automatically complete tasks every time a limit is passed.

CLAUDE COMPOSER CLI

Claude Composer CLI is a wrapper for Claude Code that adds enhancements around automation, UX, and configuration. It automatically handles permission dialogues based on configurable rules and uses rulesets to define exactly which actions to allow automatically. Users can configure which tools Claude can use and stay informed without switching contexts through system notifications.

GOOGLE ROLLING OUT UPGRADED GEMINI 2.5 PRO PREVIEW

Google has released an upgraded preview of Gemini 2.5 Pro. It will be generally available in the coming weeks. The model leads in coding across benchmarks like AIDER Polyglot and has top-tier performance on GPQA and Humanity's Last Exam. It has an Elo score of 1460 on LMArena and 1443 on WebDevArena.

AMC NETWORKS INKS DEAL WITH AI COMPANY RUNWAY

AMC plans to use AI company Runway's technology to both generate marketing images and help pre-visualize shows that have yet to be produced. The AI production images could help showrunners figure out what would or wouldn't work on the air without the high cost of shooting the material or designing from scratch. They are not intended to be used in shows at this point. Other studios are also tinkering with AI models, but they are staying silent about it due to the early nature of the tech and the sensitivity of the issue with Hollywood guilds.

GOOGLE'S CHROME BROWSER GETS 'HIGHEST SCORE EVER' ON SPEEDOMETER PERFORMANCE TEST

Speedometer 3, created collaboratively by Google, Apple, Intel, Microsoft, and Mozilla, was designed to measure web responsiveness through several workloads like HTML parsing, JavaScript and JSON processing, pixel rendering, CSS application, and more.

APPLE SAYS ITS APP STORE HELPED GENERATE $1.3T IN BILLINGS AND SALES, MOST WITHOUT A COMMISSION

Apple says that 90% of the billings and sales on its global App Store business did not involve paying Apple a commission.

PERPLEXITY RECEIVED 780 MILLION QUERIES LAST MONTH, CEO SAYS

The AI search engine is seeing more than 20% growth month-over-month.

WHAT A DEVELOPER NEEDS TO KNOW ABOUT SCIM

SCIM (the System for Cross-domain Identity Management) is a standard used by identity providers for communicating three kinds of changes to other software.

X IS TESTING A WAY TO TAKE COMMUNITY NOTES TO THE NEXT LEVEL

The new feature will display a callout box that shows when a post has been widely liked by users with differing opinions.

THE FOUNDER OF DEVIANTART IS MAKING A $22,000 DISPLAY FOR DIGITAL ART

Layer is a screen designed to showcase digital art in the best quality possible.

>> Read the research

>> Join the webinar [6] for expert analysis

GEMINI 2.5 PRO GETS AN UPGRADE

The updated preview leads all models on several major benchmarks (GPQA, Aider, and LMArena) and addresses regressions in formatting and creativity from the previous 2.5 Pro update.

ELEVEN V3: MOST EXPRESSIVE AI TEXT-TO-SPEECH MODEL LAUNCHED

Eleven Labs has launched Eleven v3, an AI text-to-speech model recognized for its expressiveness. It supports a wide range of languages, including Afrikaans, Arabic, French, and Mandarin. This advancement enhances multilingual voice applications.

CURSOR'S ANYSPHERE $9.9B VALUATION

Anysphere, maker of AI coding assistant Cursor, raised $900M at a $9.9B valuation as ARR surged past $500M.

OPENAI THREAT INTELLIGENCE REPORT: JUNE 2025

LLMs aren't giving bad actors fundamentally new capabilities. OpenAI has published 10 examples of models accelerating existing hacking, fraud, and misinformation operations. This includes North Korean actors using GPT to scale fraudulent IT worker schemes, Russian groups developing sophisticated malware, and Cambodian scammers generating multilingual “task scams” promising victims $500/day for liking TikTok posts.

LATEST ADVANCEMENTS IN SEARCH AND RECOMMENDATION SYSTEMS (4 HOUR VIDEO)

This 4-hour session, presented during the AI Engineer World's Fair 2025, covers recent innovations in search and recommendation systems.

APPLE RESEARCH FINDS CRITICAL LIMITATIONS IN REASONING MODELS

When OpenAI's o3, Claude, and DeepSeek-R1 were tested in puzzle environments, their performance collapsed beyond certain complexity thresholds, despite generating detailed "thinking" processes. The models exhibit a counterintuitive scaling limit where their reasoning effort actually decreases as problems become more complex, and they fail to improve even when given explicit solution algorithms.

LLM-DRIVEN DATA ANNOTATION

To address label uncertainty in LLM-based annotation, this paper introduces a method that captures multiple possible labels and uses a teacher-student framework called CanDist to distill them into a single output.

CLAUDE COMPOSER CLI

Claude Composer CLI is a tool that enhances Claude Code with automation, UX, and configuration. It reduces interruptions while giving users flexible control and tools to configure Claude. The tool provides users with system notifications to keep them informed. Users can control which permission dialogs are automatically accepted.

PORTRAITS: PERSONALIZED AI COACHING BUILT ALONGSIDE REAL EXPERTS

Google Labs launched Portraits, an AI coaching tool featuring experts like Kim Scott, to provide AI-driven guidance. The tool uses Gemini's capabilities to simulate expert advice through interactive avatars.

INTRODUCING MODIFY VIDEO

Modify Video allows professionals to reimagine environments, lighting, and textures in videos without altering motion or performance. It offers tools for restyling, retexturing, and editing specific elements like wardrobe and props. Outperforming competitors, Modify Video maintains motion consistency, offering multiple output variants and using advanced performance signals for high-fidelity creative control.

CLAUDE GOV MODELS FOR U.S. NATIONAL SECURITY CUSTOMERS

Anthropic trained custom models for the US government optimized for intelligence and defense use cases that have already been deployed in classified environments.

TOKASAURUS

Tokasaurus is a large language model inference engine optimized for throughput-intensive workloads.