标签：#t

My puzzle app, Keyhole, carries 296 dark stories, each with an illustrated card. A dark story is a situation that looks impossible until you drop one false assumption you did not know you were making, and the illustration must show the situation and never the reveal. Draw the aeroplane over the desert and story one is over before the player has read it. In August I ruled that the app does not ship while any card is still flagged by the judge. "End of story," I wrote in the decision, and then spent two days learning what that sentence cost. Two things get judged, the text and the art, and one design is shared by both. The judge is a model, run blind: it sees the finished card and the story the player sees, and neither the finding that triggered the redraw nor the old card. That is the whole trick. A judge that knows what was wrong last time grades the fix. A judge that knows nothing grades the card. Blindness is what makes a pass mean something, and it is why the judge is a separate call from the writer and from the illustrator, never the same conversation. The text pass first. A rubric written for the genre, with one test at its centre, "name the one assumption the solver will make that is false", and four semantic questions after it: does the reveal explain everything the situation promised, does the situation give the reveal away, is there a contradiction, can the answer be reached by yes/no questions without knowledge nobody has. Over all 296 stories it flagged 27: five unanswered, nine spoilers, ten sense breaks, three unsolvable. The fix lane rewrites only what a finding names, the deterministic gate must still pass, and the blind judge reads the result cold before it is written back. A fact-check over the rewrites then cleared them, or left a truth note where no honest fix existed. The art pass is where the numbers live. Each open card was redrawn from a scene brief and judged blind, in waves. The judge wrote a note on every failure, and the lever changed from

Hello, fellow version-bumping enthusiasts, sleep-deprived Rustaceans, and accidental software...

When iOS 27 lands later this month, it will have a feature called iPhone Handoff that lets you switch between two phones using the same number. It was briefly mentioned during the WWDC keynote back in June, but there were no details at the time. Now there's a demo clip showing how to set up […]

There are plenty of durable, versatile rugged laptops on the market, but you'll rarely see one get pulled out at the coffee shop. Why?

Agentic Methods: Coding With AI Agents, Designing For Agents TL;DR "Agentic methods" covers two distinct things colliding right now: AI agents that code alongside the team (read, write, run, verify, in a loop), and agentic architectures we design into our own systems (orchestrating autonomous agents on the product side). In both cases, the same principle applies: an agent is only useful if the contract around it is explicit — scope, errors, permissions, stopping points. The Tech Lead role doesn't disappear, it shifts: fewer lines typed, more specification, review, and governance. The underlying topic isn't tooling, it's clarity — exactly like a well-modelled business workflow. Table of Contents Introduction — one word, two meanings Coding with AI agents: what actually changes From autocomplete to the agentic loop The developer's role shifts toward review Explicit guardrails Designing agentic architectures An agent is a box with a contract Orchestration or autonomy: a choice, not a default Observability: if you can't replay it, you can't debug it Where humans remain irreplaceable A Tech Lead checklist for adopting these methods Conclusion — agents reveal a team's maturity Introduction — one word, two meanings "Agentic" has been everywhere for a few months, but it means two different things depending on who's talking: Coding with AI agents : a tool that reads code, writes diffs, runs commands, launches tests, and iterates until it reaches a correct result — instead of suggesting one line at a time. Designing agentic systems : a software architecture where autonomous agents (often themselves LLM-based) make decisions, call tools, and cooperate to accomplish a business task — a support chatbot that triggers refunds, a document pipeline that routes complex cases to a human on its own. These are two separate topics, but the same underlying principle runs through both: an agent — human, AI, or service — is only reliable when it operates inside an explicit frame. It's the s

Building StudySift Without Third-Party Dependencies Introduction What if a useful study tool could be built without installing a single third-party package? For the Zero Dependency Hackathon, I built StudySift , a command-line tool that converts lecture transcripts into structured, revision-friendly study notes. The idea is simple: give StudySift a transcript and automatically extract useful information such as keywords, definitions, examples, and important points. The interesting part was the constraint. The project had to run using Python's standard library only , with no third-party runtime dependencies. The Problem Lecture transcripts can be long and difficult to revise. Important definitions, examples, keywords, and important statements can be spread throughout the transcript. Students often have to manually read the entire transcript, identify important sentences, and create their own notes. I wanted to reduce this manual work. StudySift takes a text transcript as input and processes it into organized notes. The basic workflow is: Lecture Transcript ↓ StudySift ↓ ┌─────────────────┐ │ Definitions │ │ Important Points│ │ Examples │ │ Keywords │ └─────────────────┘ **What I Built** StudySift is a Python command-line tool. The user provides a transcript file: python src/main.py examples/lecture.txt StudySift processes the transcript through several stages: 1. Read the input file 2. Split the text into sentences 3. Extract words 4. Remove common words 5. Count word frequencies 6. Detect definitions 7. Detect examples 8. Identify important sentences 9. Score sentences 10. Sort sentences by importance 11. Generate structured notes The goal is not to pretend that a collection of simple rules is a complete natural-language understanding system. Instead, StudySift is a lightweight and transparent approach to turning transcripts into useful revision material. **The Zero-Dependency Challenge** The biggest constraint was that StudySift could not depend on third-party runt

OpenAI has officially introduced GPT-6 Astra , a new model it describes as its most capable and aligned to date. The launch centers on advanced computer use, software engineering, browsing, cybersecurity tasks and professional knowledge work. Astra is initially rolling out to a limited group of organizations, followed by availability for paid ChatGPT users and developers across the OpenAI API, Microsoft Azure and AWS Bedrock . The formal release supersedes earlier speculation around a potentially "special" model rollout. OpenAI’s official GPT-6 Astra announcement establishes the substantive news: a staged, multi-platform deployment with defined API pricing, large context capacity and capabilities aimed at completing more complex digital tasks. For businesses, the important question is less whether Astra is unusual and more whether its computer-use functions can reliably reduce manual work in existing processes. OpenAI positions the model for tasks such as filling forms, updating CRM records, managing calendars, researching the web, installing and troubleshooting software, and producing documents, spreadsheets and presentations that follow a user’s templates and style. What GPT-6 Astra adds Astra is designed to work across tasks that ordinarily require moving between software interfaces, web pages and business documents. That is a significant expansion from using a language model solely to draft text or answer questions. In the right workflow, a model that can navigate authorized tools and complete multi-step tasks could help teams reduce repetitive administrative work. OpenAI also highlights Astra’s performance in code generation and professional knowledge work. Its stated ability to install, test and troubleshoot software points toward more autonomous technical workflows, while its document-generation capabilities could be relevant for recurring reports, proposals, analysis packs and operational templates. The model’s published limits and access paths are also nota

A cryptography library had a bottleneck no test could see : encrypting half a megabyte took fifty seconds. Every test passed. They had been passing for months. The cause is a trap that keeps recurring: a correct, well-documented optimisation that fixes the constant and not the order — and whose comment, precisely because it is well written, convinces the reader the problem is already solved. What the code did Quipu renders encrypted data as a sequence of symbols. To do that it converts the whole message into a single huge integer and repeatedly divides it to extract digits, the same way you would convert a base-10 number to base 2 by hand. The code did not divide one digit at a time. It carried a sensible optimisation: divide by the largest power of the base that fits in a machine word, extracting nine digits per pass instead of one. The comment explaining it opened by saying that doing it one at a time would be quadratic , and then described the improvement. All true. And the result was still quadratic: extracting nine digits per pass divides the work by nine; it does not change how the work grows. That sentence — "doing it this way would be quadratic" — reads in the past tense, as if it described the previous state. It described the current one. The measurement, which is the only thing that says so Size Time Factor per doubling 64 KiB 0.79 s — 128 KiB 3.16 s ×4.0 256 KiB 12.6 s ×4.0 512 KiB 50.7 s ×4.0 Exactly four, three times running. That is textbook quadratic: every time the input doubles, the time quadruples. Extrapolating, ten megabytes would have cost about five and a half hours . And here is the point: a correctness test sees none of this . A slow algorithm produces exactly the same bytes as a fast one. The suite stayed green, and would have stayed green forever. The fix is two hundred years old Nothing had to be invented. Divide-and-conquer radix conversion is a classical algorithm: instead of peeling digits off one end, you split the number in half — div

Most local business lead scrapers on Apify are Google Maps scrapers with a website-crawling step bolted on. lukaskrivka/google-maps-with-contact-details is the most used (87,957 users, 4.63 stars). flash_scraper/local-business-leads is the outlier: it discovers businesses on OpenStreetMap instead of Google Maps, and includes MX email verification in its $3 per 1,000. Every figure below was read from Apify's public Store API ( GET /v2/store ) on 2026-09-05 — including every user count, so they are all on the same footing. The per-actor endpoint ( GET /v2/acts/<id> ) can read one higher: it gives flash_scraper/local-business-leads 33 rather than 32, and code-node-tools 33 as well. Prices, users and ratings change; the Pricing tab on each actor page is authoritative. Disclosure: I publish flash_scraper/local-business-leads , one of the actors compared here. Its limits are listed in the same detail as everyone else's, including the one that will disqualify it for many buyers. How prices are normalised These actors bill per event, and the events differ in kind, which makes headline prices misleading. Some charge per place found. Some charge separately for the website crawl that actually produces the email. Some charge again to verify that the email is deliverable. The table lists the primary per-result event multiplied by 1,000 at the free-plan rate , then names the add-on events, because a $5 per 1,000 place price with a $100 per 1,000 email-verification add-on is not a $5 tool. Paid Apify plans get tiered discounts on several of these actors, ours included — and on the add-on events the discount can be enormous. lukaskrivka's three $100-per-1,000 add-ons fall to $4.00 (email verification), $7.50 (lead enrichment) and $10.00 (social-profile enrichment) per 1,000 on Bronze, and lower again above it (Store pricing record read 2026-09-05). Our own free-plan-to-Diamond spread is about 30 percent. So if you are on a paid plan, re-read every figure below off the Pricing tab:

Every Azure project eventually has the same conversation. Where do we store the connection string? Someone suggests an environment variable, and someone else points out that environment variables end up in deployment pipelines, in Docker compose files, in Terraform state, and occasionally in accidental commits. A secret manager gets proposed, and the secret manager needs its own credentials to access the secrets. The problem recurses. Managed Identity doesn't solve the secret manager problem by adding another layer. It removes the credential from the equation entirely for workloads running inside Azure, so the application doesn't authenticate with a stored credential but as itself, using an identity that Azure manages automatically. What Managed Identity actually does When you enable a Managed Identity on an Azure resource, Azure creates an identity in Microsoft Entra ID tied to that resource's lifecycle. The resource can then request short-lived tokens from the Azure Instance Metadata Service endpoint at 169.254.169.254 , which is only reachable from within Azure infrastructure, and those tokens are what the resource uses to authenticate against other Azure services. There's nothing to store, nothing to rotate manually, and nothing that can be leaked in a repository because the credential never exists as a static string anywhere in your codebase or configuration. The key property from Microsoft's documentation is precise: managed identities give code running on an Azure resource access to other resources without developers needing to handle or put credentials directly into code. The emphasis on "code running on an Azure resource" matters because Managed Identity only works from within Azure. A local development machine can't reach the Instance Metadata Service endpoint, which means the local development flow still needs an alternative authentication mechanism, typically az login or a service principal configured for development only. System-assigned vs user-assigne

ChatGPT.com reached about 1.09 billion monthly US visits in July 2026, a 48.38% year-over-year increase, according to Semrush Traffic Analytics data. In the same comparison, Bing.com traffic fell about 50.43%. The contrast does not show AI replacing conventional search overnight. Google and YouTube still led the US dataset by a wide margin. It does show that the places where people first discover information are changing, while many website analytics setups remain poorly equipped to show the full effect. The July 2026 snapshot puts ChatGPT ninth among the leading US sites measured by Semrush. Businesses that still view organic discovery mainly through Google rankings and familiar referral reports risk missing a growing part of the customer journey: a user may ask an AI assistant for options, follow a recommendation, and arrive on a website without a cleanly identifiable source in Google Analytics 4. The underlying Semrush US Trending Websites data compares July 2026 traffic with July 2025. It is a view of US web traffic in Semrush's ranked-site dataset, not a count of every search or AI interaction. Still, the scale of the movement makes AI-assisted discovery a practical measurement issue, not simply a trend to monitor. What the traffic shift means for SEO measurement The key implication is not that businesses should abandon established search channels. Google recorded roughly 25.31 billion monthly US visits in the July snapshot, while YouTube recorded about 10.27 billion. Those figures underline how large the established platforms remain. What has changed is the need to distinguish where discovery happens from the source that ultimately appears in analytics. ChatGPT's growth can create new paths to content, products and services. But referral details may be unavailable when users move from an AI interface to a website, particularly when the originating referrer is stripped. In GA4, those visits can be grouped as Direct or remain otherwise difficult to classify. Web

Originally published at nlocoding.com 41% of enterprise database engineers already use AI tools daily to generate, optimize, or review SQL—up from just 14% in 2023 (Gartner, 2026). The new database arms race is invisible. Enterprises process 7.4x more data per person than they did five years ago. That’s not a typo. AI-assisted database development isn’t just about speed; it’s about not drowning in schema drift and query chaos. If you’re not automating, you’re lagging by $8,200 per developer per year (Forrester, 2026). 73%of data teams say AI reduced query errors (Redgate, 2026) AI-assisted database development is rewriting the rules in 2026 AI-assisted database development is now the backbone for 52% of Fortune 500 engineering departments, slashing schema build time by 48% on average (Stack Overflow Developer Survey, 2026). Developers no longer waste days hand-writing migration scripts or debugging malformed indexes. Instead, GPT-5-powered copilots like Tabnine and DataPilot draft DDL, suggest denormalization strategies, and catch performance anti-patterns before they hit production. The result: projects ship 23% faster, according to Fivetran’s 2026 benchmark. If you’re still relying on manual SQL, you’re not just slower—you’re more expensive. Find one workflow, automate it, and measure the delta. That’s how the best teams start. ⚠️ Common Mistake: Treating AI-generated schema suggestions as gospel. Blind trust leads to silent data loss or bloated tables. Always review before merging. Schema design is now a conversation, not a bottleneck Most people get this wrong: schema design is not just a technical hurdle—it’s a communication bottleneck. In 2026, 64% of product teams report that AI-driven schema prototyping (using tools like dbdiagram.io+AI Assist, $7/month) reduced handoff time between engineering and product by 58% (LinearB, 2026). Instead of four revision meetings, you get a Slack thread with three alternative schemas, clear tradeoffs, and a side-by-side diff

OpenAI acknowledged its role in a recently reported incident where AI agents took over a German wiki forum.

There are almost certainly keys on your keyboard that you never touch. Assign new functions to give them a job.

In the case of IMAX, bigger really is better.

According to the Recording Industry Association of America (RIAA), CD sales exploded in the first half of 2026. A new report from the organization says 17.5 million CDs were sold in the first six months of the year, up from just 12 million during the same period in 2025. That's a jump of 45.7 percent, […]

You might not think to restart your router as often as you reboot your laptop, but doing so is an easy step to help it perform at its best.

“Let’s split it into agents” has become the AI equivalent of “let’s make it a microservice.” Sometimes the boundary is useful. Sometimes it only creates more state, more coordination, and a harder failure to explain. The most dangerous assumption is that separate agents should run in parallel. Parallelism is safe only when the branches are genuinely independent. If one branch changes the world while another is evaluating it, both agents can make locally reasonable decisions that are unsafe together. Google ADK 2.0 makes workflow topology explicit through graph-based Workflow objects. That is valuable because sequences, branches, and joins become part of the program instead of an agreement hidden in a supervisor prompt. Series note: This is Part 5 of Reliable Google AI Agents in TypeScript . The examples were checked against @google/adk 2.0.0 in September 2026. Start with the dependency, not the agent count Imagine a system preparing a hotel recommendation. It needs live inventory, company travel policy, and a final recommendation. Inventory lookup and policy evaluation can run concurrently because both observe the same request and neither changes shared state. The final decision must wait for both. Now consider a different pair of operations: one agent changes the reservation; another calculates an upgrade using the current reservation. Those branches are not independent. Running them concurrently can make the upgrade decision depend on state that no longer exists. Before drawing a parallel branch, ask: Do both operations only read the same starting state? Can either operation change data the other consumes? Can either produce an irreversible side effect? Is there a deterministic way to combine their results? What happens when one succeeds and the other times out? If those answers are unclear, parallel is an optimization you have not earned yet. Encode safe parallelism as fan-out and join ADK’s TypeScript Workflow graph can express two independent branches and a joi

Much of the traditional software operations we deal with are online, we click a button, wait for a moment, and the transaction or operation is completed. But there is a big area that deals with software operations that require offline processing. For example, background processing of jobs, e.g., OpenAI training/improving its existing GPT models behind the scenes using the data it gathers from its users. Batch Processing Whenever such an offline system runs a job that typically generates output from a batch of inputs, we call that batch processing. Inputs here are immutable, which avoids side effects. Benefits of batch processing: You can time travel. In case of any failure or unintentional outputs, you can jump to the last input checkpoint before a batch processing job. This handling is often referred to as human fault tolerance. Using batch processing and offline systems, compute usage efficiency can be improved. For example, whenever a heavy computation needs to be done, it's better to do it in bulk on maybe a GPU compute rather than crashing the CPU host where the server is online. Though the boundary between online and batch processing is not always clear. For example, a long-running database query could also be categorised as batch processing. Another alternative to batch processing is stream processing, which we will understand in the next article. MapReduce MapReduce is a batch processing algorithm that is utilized by Hadoop, CouchDB, and MongoDB as well. It is a balanced approach that is less extreme than completely parallelizing the jobs. There are several other frameworks like this that are now replacing MapReduce. For example, DataFrames APIs, query languages, etc. We will see MapReduce in detail sometime later. Simulating Batch Processing with Unix Tools (Single Host) If you are a Linux user, this simulation could be very easy for you to grasp. If not, just put it in ChatGPT or any AI tool to understand the command in detail if interested. A typical Ngin

I maintain a small collection of single-purpose web tools. Last month I built a favicon checker: you type a URL, it reads the icon declarations in the HTML head, probes every referenced file, and also hits /favicon.ico directly, because plenty of software still requests that path without ever reading your HTML. The first thing you should do with any auditing tool is point it at your own stuff. So I did. Ten sites, all built by me, all shipped and verified in browsers I actually use. All ten failed. Not seven out of ten. Ten. Failure one: SVG-only icon sets Every site had a nice crisp favicon.svg and nothing else. Modern browsers request it, render it at any size, everything looks great in Chrome and Firefox. Then something older comes along: a bookmark sidebar, an RSS reader, a corporate proxy portal that lists your link, that one intern running Opera 12. These clients do not parse your <link> tags. They request /favicon.ico and hope. All ten sites returned a 404 for that path. The fix is not glamorous. You need an actual .ico file, ideally with 16, 32, and 48 pixel frames packed inside, plus a PNG for iOS. More on that below. Failure two: no apple-touch-icon Nine of the ten sites had no apple-touch-icon.png . When someone saves such a site to an iOS home screen, Safari does not use your favicon. It takes a screenshot of the page, letterboxes it, and calls that your app icon. If you have ever seen a bookmark that looked like a cropped text paragraph, that is why. The fix is one file and one tag: a 180 by 180 PNG, referenced with <link rel="apple-touch-icon" href="/apple-touch-icon.png"> . Done. No JavaScript, no media queries, no dark mode variants needed. iOS rounds the corners itself. Failure three: the 404 that would not leave This is the one that cost me an evening, so pay attention if any of your sites sit behind Cloudflare. I generated the missing icons, deployed them, and re-ran the checker. Still 404. I deployed again. Still 404. I started doubting my build,