I have tried journaling many times but nothing stuck. So I decided to make my own app for myself with these features: * Voice first * Private first * AI chat * Automatic tagging, meaning extraction, and semantic retrieval using embeddings stored locally * Export to LLM the AI angle is especially interesting for me. It allows me to ask questions like: * "remind me highlights and crazy nights in the last 3 months" * "How was I feeling during my trip in Spain and how big of a problem was my breakup

Your monitoring catches problems on everything except itself. Here is how an always-firing Watchdog alert plus an external heartbeat check turns silence into a signal, so you find out when your own alerting dies. TL;DR A monitoring system can't reliably monitor its own failure, so use a dead man's switch. Create an always-firing Prometheus Watchdog alert and route it to an independent external heartbeat service. As long as the monitoring pipeline is working, the Watchdog continuously refreshes the heartbeat. If Prometheus, Alertmanager, or the delivery path fails, the heartbeat stops and the external service alerts you through a separate channel. The key is independence: the system responsible for detecting that your monitoring is down must not depend on the monitoring stack itself. One of the traps of creating alerts on a monitoring stack is the hidden assumption that the mechanism evaluating the alert is running properly and has the ability to evaluate it. Prometheus watches your hosts and Alertmanager delivers the warnings. But what watches Prometheus? If something goes wrong and the monitoring stack fails in the middle of the night, no alerts are going out but there is definitely a problem. That is the failure mode that you should be most concerned about, because it is the one your monitoring cannot report on. The fix is an old idea with a grim name: a dead man's switch. A train's dead man's switch stops the train when the operator stops holding it down. The safe state requires continuous positive action, while the absence of that action is what triggers the response. Applied to monitoring, it means building one alert whose silence is itself the alarm. Step one: an alert that always fires This feels backwards the first time you see it, and it took me a little time to get it right. Basically, you create an alert with a condition that is always true, so it fires constantly, forever, on purpose. In the Prometheus world this is conventionally called Watchdog. - aler

Our field-tested gear includes power banks, electrolytes, outdoor blankets, earplugs, and more.

Why identity-local signals and topology signals are two layers of the same blast radius The credential with the widest blast radius sometimes has no secret to flag. See how GitGuardian and Anyshift rank risk by what actually breaks. By Louis Fradin • 23 Jul 2026 • 7 min read 👉 TL;DR: Identity-local signals show whether a credential or machine identity is risky. Topology signals show what breaks if that identity is abused. GitGuardian identifies and ranks exposed credentials and risky machine identities; Anyshift's graph adds downstream context by showing which services depend on the resources those identities reach. Together, they help teams prioritize by both credential severity and operational blast radius. A leaked credential is also a topology problem A leaked credential creates risk beyond the identity itself. Its real impact depends on the services and resources connected to what that credential can access. Identity-local signals answer the first question: how risky is this credential or machine identity on its own? Is it plaintext? Guessable? Stale? Overprivileged? Production-exposed? Tied to an admin identity? Those signals matter because they identify the secrets and machine identities most likely to be abused. But they do not answer the next question: what breaks if that credential is used? That answer lives in the topology around the credential. A database credential may sit on one pod and unlock one datastore, but the operational blast radius extends to every service that depends on that datastore. Some of those services never hold the credential at all. Some may not even have a secret signal to score. Want to run the same analysis on your own stack? Explore the Anyshift Graph API to query dependencies, blast radius, and production impact directly. Learn more That is where identity-local signals and topology signals become two layers of the same blast radius: one tells you why the credential is dangerous, and the other tells you how far the damage can tr

Nothing has had a strong visual identity since the Ear 1 were released in 2021 for $99 and challenged the notion that a highly featured headset requires a high price. The Ear 3A have the same transparent look as Nothing's previous earbuds, with a clear case and stem design accented by a solid-colored body - […]

Indian government exam portals have a rule that has quietly shaped a lot of my code: your photo must be under 50KB . Not "small". Not "optimised". Under 50KB, or the upload is rejected. Every free tool I found for this wanted me to upload the photo to a server, wait in a queue, and create an account. For a file I just wanted to shrink. So I wrote it myself, in the browser. This post is about the actual technique — hitting an exact byte target with canvas — and, honestly, about where my implementation still falls short. The naive version, and why it fails The obvious approach: canvas . toBlob ( blob => download ( blob ), ' image/jpeg ' , 0.7 ); Pick a quality, hope for the best. The problem is that JPEG quality has no predictable relationship to output size. Quality 0.7 on a flat, low-detail portrait might land at 18KB. The same 0.7 on a noisy, high-detail photo lands at 210KB. You cannot compute the quality you need — the encoder decides, and it depends entirely on image content. So you can't calculate it. You have to search for it. Binary search on quality toBlob is cheap enough to call repeatedly, and quality is monotonic — higher quality never produces a smaller file. That's exactly the setup binary search wants. function encode ( canvas , quality ) { return new Promise ( resolve => canvas . toBlob ( resolve , ' image/jpeg ' , quality ) ); } async function compressToTarget ( canvas , targetBytes ) { let lo = 0.05 , hi = 0.95 , best = null ; for ( let i = 0 ; i < 8 ; i ++ ) { const mid = ( lo + hi ) / 2 ; const blob = await encode ( canvas , mid ); if ( blob . size <= targetBytes ) { best = blob ; // fits — remember it, try for better quality lo = mid ; } else { hi = mid ; // too big — back off } } return best ; } Eight iterations over the range 0.05–0.95 narrows quality to about ±0.002, far finer than anyone can see. Each iteration is one encode; on a typical phone photo the whole loop runs in well under a second. Two details that matter more than they look: Keep

Sharing a holiday home with family or friends is great until the admin starts. Who’s in next weekend? Did someone already claim Easter? Who was meant to book the cleaner? Where’s the WiFi password / insurance cert / “how to winterize the outdoor taps” note? For most groups this lives in five group chats, a half-maintained Google Calendar, and a Drive folder nobody trusts. I kept running into that pattern — so I built Shared Holiday Homes : software for families, friends, and co-owners who already share a place and need less chaos, not another generic calendar. The problem isn’t “finding a free date” Generic calendars are fine at showing blocks of time. Shared holiday homes need more than that: Double-booking protection that isn’t “hope nobody overwrites the event” Rules for peak weeks, min/max stays, booking windows, and optional approval Fairness visibility — who actually used the place this year Named jobs with owners and due dates (cleaning, maintenance, “fix the pump”) A home for house knowledge — docs, arrival notes, appliance quirks, emergency info If your group is small and high-trust, Google Calendar can work. Once you’re coordinating multiple households, peak seasons, and maintenance, the “calendar + WhatsApp” stack starts creating the arguments it’s supposed to prevent. I wrote a longer comparison here if you want the practical breakdown: Shared Holiday Homes vs Google Calendar What I built (and what I didn’t) The product is intentionally narrow. Private co-owner groups don’t need a full property-management system or a fractional-ownership marketplace. They need an operating layer for one shared house. In scope: One shared booking calendar Booking rules / seasonal rotations Shared task list Document library House guides (the handbook people can actually find) Out of scope on purpose: Selling property shares Matching investors Full bookkeeping / STR channel management That boundary mattered. Every time I was tempted to add “just one more admin feature,” I a

I've been following WWDC since 2011* and over the years I've developed my own process for watching sessions. Over the past 15 years, I've gone through all the stages from denial (I missed all the videos, for example, in 2014, when I couldn't accept Swift's appearance), to bargaining ("I still CAN watch all of them!") and finally accepting and creating my own approach to watching WWDC. So here's my current approach: I try to watch Keynote and Platforms State of the Union at the time of their live broadcast or in the early days. When all the sessions become available, I sit down and go through all the titles and descriptions and choose what I will watch. All sessions fall into 4 categories: Essential is a must watch and should never be missed Nice To Watch is something interesting to me personally and may be applicable to what I work with Only If I Have Time is for optional sessions that interested me, but if I skip them, then nothing terrible will happen Everything else Thus, I clearly identify a fairly small set of sessions that I definitely need to watch, usually it's about 10 sessions and I don't feel any FOMO or pressure that there is so much new and how and when to watch it all. Most of the time, it takes me all summer to slowly watch everything from Essential, a few (or all, it depends) Nice To Watch, and sometimes a couple, and sometimes nothing at all from Only If I Have Time. An important rule for me is to watch Essential first. Then I can do whatever I want. Another personal kink of mine is to watch the sessions in the order of their numbering. For example, in my Essential category, the very first video is 227. Create UI prototypes using agents in Xcode , and then 258. What’s new in Xcode 27 and so on. So, here is my personal list of WWDC’26 sessions, divided into these categories: Essential Create UI prototypes using agents in Xcode What’s new in Xcode 27 Xcode, agents and you Get the most out of Device Hub What’s new in Swift What’s new in SwiftUI Moderni

Is this the future of foldable phones? Samsung clearly thinks so. The company has given its new wide foldable the "Z Fold 8" name, positioning the phone not as a widescreen oddity, but as the new normal, the default design for foldables going forward. Rumor has it that Apple feels the same way, with a […]

The build method in Flutter widgets is synchronous. That means it doesn’t like to wait for anything. But sometimes, we need to wait for a value to arrive in order to display it. Let’s think of a simple weather app that displays only the temperature of a city. The app needs to make a request to the backend, get the temperature value, and finally display it. It will have to wait for a response from the backend, but as we discussed, the build method does not like to wait for anything. So how do we solve this issue? Enter: FutureBuilder . FutureBuilder takes a value of type Future and displays widgets until it is resolved. In fact, we can specify which widgets to display not only while loading but also when an error occurs. Let’s see how we can use FutureBuilder in a simple app. First, create an app in a directory of your choice: flutter create future_builder --platforms = macos You can choose whichever platform you want. Open the project in your preferred IDE, and navigate to lib/main.dart . Replace the entire content of the file with the following: import 'package:flutter/material.dart' ; void main () { runApp ( const MyApp ()); } class MyApp extends StatelessWidget { const MyApp ({ super . key }); @override Widget build ( BuildContext context ) { return MaterialApp ( home: const MyHomePage ()); } } class MyHomePage extends StatelessWidget { const MyHomePage ({ super . key }); Future < int > _getTemperature () async { await Future . delayed ( Duration ( seconds: 3 )); // Dummy delay of three seconds. return 25 ; } Future < int > _getTemperatureError () async { await Future . delayed ( Duration ( seconds: 3 )); throw Exception ( 'An error occurred while retrieving the temperature value.' ); } Future < int ? > _getTemperatureEmpty () async { await Future . delayed ( Duration ( seconds: 3 )); return null ; } @override Widget build ( BuildContext context ) { return Scaffold ( body: Center ( child: FutureBuilder ( future: _getTemperature (), builder: ( context , snapshot )

A pattern I’ve seen many times in software projects is that documentation starts too late and documents the wrong thing. A team ships a feature, the code works, the tests pass, and everyone moves on. Maybe someone adds a README section, maybe not. If they do, it usually explains how to run something, how to call an endpoint, or what a component does. That kind of documentation is useful, but it often misses the part future developers need most. It misses the decision. Six months later, someone opens the same part of the codebase and asks the usual questions. Why is this data model shaped like this? Why is this rule handled in the backend instead of the frontend? Why is this integration synchronous? Why does this permission check live here? Why did the team choose this simple approach instead of something more flexible? The code can show what exists, but it rarely explains why it exists. That is where a lot of engineering context disappears. The Problem Is Not Always Missing Documentation When people complain about documentation, the usual diagnosis is that there is not enough of it. The README is outdated. The setup instructions are incomplete. The API docs are missing examples. The architecture diagram no longer matches reality. All of those problems are real. But I think there is another documentation problem that is easier to miss: the docs describe the system without preserving the reasoning behind it. This matters because software is full of trade-offs. A piece of code may look strange because it was written badly, but it may also look strange because it was solving a constraint that is no longer visible. Maybe the team chose a simpler data model because they were still validating the product. Maybe they avoided a generic abstraction because they had only one real use case. Maybe they accepted duplication because the two workflows looked similar but were expected to diverge. Without the reasoning, future developers have to guess. That guessing creates waste. So

OpenWorker shipped in late July 2026. It is MIT-licensed, runs on your own machine, and takes your API key instead of selling you inference. The pitch is narrow and worth repeating exactly: it is an agent that hands you finished work , not a chat transcript. A drafted document on disk. A Slack reply with the real numbers in it. A calendar that has actually been rearranged. There are a lot of desktop agents right now. This post is about what makes this one structurally different, what state it is actually in, and how to get it running. What it is in one paragraph OpenWorker is a desktop app: a Tauri shell around a React UI, sitting on top of a local Python agent server. You give it an outcome ("prepare a customer brief from these three files and the Jira tickets"). It decomposes that into steps, reaches into your files, terminal, and connected SaaS apps, and produces an artifact. Before anything consequential happens - sending a message, running a shell command, writing to your calendar - it stops and asks. The engine is built on aisuite , Ng's provider-agnostic LLM library. That matters more than it sounds like: OpenWorker is explicitly positioned as a reference implementation of what you can build on aisuite, so the codebase doubles as a worked example if you are building your own harness. The four things that actually distinguish it 1. There is no OpenWorker inference service You paste a key, or you point it at Ollama and use none at all. The curated list covers OpenAI, Anthropic, Google, plus OpenAI-compatible vendors like DeepSeek, GLM, Kimi, Qwen, MiniMax, Mistral, and Grok, plus open-weight models through Together and Fireworks. Roughly thirty models are marked as verified for tool-calling work; you can point it at any other model string and accept the risk yourself. The practical consequence: your cost is your provider bill, and swapping models is a dropdown, not a migration. 2. The permission model is typed, not a confirmation dialog This is the part I would

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Samsung’s chip workers are jumping ship to rival SK Hynix Lee, an engineer at Samsung’s semiconductor division, used to work late. But lately, he’s been clocking out on time and heading…

Anthropic published Verbalizable Representations Form a Global Workspace in Language Models on July 6, and the vocabulary it introduced is suddenly everywhere: J-space, the Jacobian lens, a global workspace inside Claude. Most of the discussion so far is about interpretability and alignment auditing, which is fair, since that is what the paper is about. I had a narrower and more mercenary question: can the workspace tell an inference runtime which parts of the KV cache it is safe to throw away? Three days after the paper landed, the first pre-registered gate on that question passed. As of this week the signal has replicated on three models and ships inside EVOKE , my KV cache memory manager built on a forked llama.cpp. This post covers what J-space is, why it makes a good KV cache eviction signal, the numbers across Qwen2.5-7B, Qwen3-8B, and Qwen3-4B, and the caveat that comes with them. What J-space is, in one paragraph The Jacobian lens is the instrument and J-space is the phenomenon. The lens isolates directions in a model's residual stream that encode a token the model could verbalize next, and those directions form a low-dimensional workspace: roughly 10% of activation variance, concentrated in the middle layers, carrying whatever the model is "holding in mind" at each position. Anthropic's headline application is alignment auditing, reading reasoning the model never voices. What makes independent work possible is that they released companion code under Apache-2.0 along with fitted lens matrices for open Qwen models on Hugging Face , so anyone can apply the lens to an open-weights model on a single GPU. The systems problem: KV cache eviction Every long-running LLM session eventually outgrows its KV cache budget. An agent session in a coding harness crosses tens of thousands of cached tokens within a few turns, and something has to decide which entries stay in GPU memory. The standard answers, H2O and SnapKV, rank cache blocks by accumulated attention history: k