标签：#AR

Not everyone wants a screen on their wrist.

Let's poke at the differences between scroll- driven and scroll- triggered animations. A First Look at Scroll-Triggered Animations originally handwritten and published with love on CSS-Tricks . You should really get the newsletter as well.

Internmaxxing Somebody on your timeline this week called intern code "API slop."...

Pre: What is BlockSocial? BlockSocial is the ultimate social network for developers, bringing the energy of short-form video to the world of open source. Think of it as Facebook meets Instagram—a place to showcase your code, find inspiration, and build your developer brand through "Reels" and interactive dashboards. Github link: https://github.com/Hfs2024/BlockSocial 1. User Scenario & Workflow (The Fork System) The Setup User A : Publishes a post saying: "I love drinking Pepsi every day." User B : Is shy, but wants to tell their friend this is an unhealthy habit. User C : Is a malicious user who gossips. The Fork Mechanism User B creates a fork to discuss this post with User C via the POST /api/share endpoint. Data Copying : It copies the entire post contents except comments, likes, reports, and downloads. Chain Prevention : You can fork a forked post, but the system will fork the original source root, not the fork itself. Scope : It shares with only one user at a time to prevent unexpected group creation. Database Payload for Forks The following fields are appended to the document structure: { "share" : true , "shareId" : "post._id" , // The original post ID "sharedBy" : "req.currentUser.username" , // The user who shared or forked "shareTo" : "shareTo" , // The friend receiving the share "shareComment" : "comment || ''" // A quick comment on the post } Moderation & Enforcement Workflow If User C breaks trust and leaks the conversation, User B can report them via the POST /report/user endpoint. Verification : Administrators review interaction history to verify the violation. Account Termination : Bad users receive a permanent lifetime account ban. Data Scrubbing : All associated messages from the malicious user are removed. Blacklisting : The account is fully banned. The Blindspot : Face-to-face interactions remain outside system moderation boundaries 😅 2. Technical Implementation Details Dynamic Comment Identity Logic When a user submits a comment via POST /api/c

For years I treated burnout as a personal failing. If I was tired, I needed more sleep. If I was anxious on Sunday night, I needed to meditate. If I dreaded standup, I needed a better attitude. None of it worked, because I was treating an organizational problem as a character problem. Senior engineer burnout rarely looks like simple exhaustion. It looks like your pull request reviews getting slower. It looks like tech debt you keep meaning to document and never do. It looks like every "quick question" landing in your DMs, because you are the person who knows where everything is. The load is structural. You cannot meditate your way out of an org chart. Here is the framework that finally helped me, and that I now keep as a runbook. First, diagnose: acute or systemic A rough sprint is not burnout. A hard quarter is not burnout. Those are acute, and they resolve when the spike passes. Systemic burnout is different. The recovery never comes, because the structure that caused it never changes. You finish the death-march launch and the next one is already scheduled. You clear the queue and it refills by lunch. The mistake is applying acute fixes (a long weekend, a vacation) to a systemic problem. You come back rested, the structure grinds you down again in two weeks, and now you also feel like the rest "did not work," which makes it worse. A quick self-check. In the last month: Do you feel recovered after a weekend, or does Sunday-evening dread start by Saturday night? Is your reduced capacity tied to one specific deadline, or is it just how things are now? If your single worst recurring task vanished tomorrow, would you feel fine, or would something else immediately take its place? If your answers point to "it is just how things are now," you are dealing with systemic burnout, and the fixes are structural, not personal. Reclaim deep work with routing, not willpower Deep work does not survive on discipline. It survives on routing. The senior engineer's calendar is a public

Your retriever returned the right documents. The similarity scores look fine. The answer is still wrong. If you've shipped RAG, you've seen this — and it's the failure that survives every retrieval upgrade. What everyone tries Reranker. Higher top-k. Hybrid search. A better embedding model. All of these chase the same goal: documents more similar to the query. They help when the right document wasn't being retrieved. They do nothing when the right document was retrieved and the answer is still wrong. Why it doesn't work Similarity answers "is this chunk about the same topic?" It does not answer "does this chunk contain the facts needed to support the answer?" Those come apart constantly. A chunk can be highly similar — same vocabulary, same subject — and contain nothing that actually grounds the answer. Hand the model a pile of on-topic text and it will produce a fluent, plausible, even cited-looking answer. The grounding is cosmetic: the text was nearby, not load-bearing. High similarity with a wrong answer isn't a contradiction. You asked retrieval to find related text. It did. Nobody asked whether the text was enough. The one shift Stop treating retrieval output as evidence. Treat it as candidate material that has to pass an explicit evidence check before it can support an answer. Put a step between retrieval and generation: does the retrieved set actually contain the facts this answer requires? If not, abstain. When the documents don't contain the facts, the system should return nothing rather than a confident guess. Relevant context in, only sufficient evidence allowed through. That's the line between a RAG demo and a RAG system you can trust in production. I write about the three boundaries where production RAG dies — query, evidence, output — from the angle of shipping under security and model constraints. Read the full version on my blog , where this connects to the practical RAG Failure Diagnosis Kit for teams debugging production RAG.

A few months ago, I was overwhelmed by everything happening in AI. Every week there was a new coding assistant, a new workflow, or someone claiming they built an app in just a few hours. It felt like if you weren't keeping up, you'd be left behind. I tried almost everything. Cursor. ChatGPT. Claude Code. Lovable. At first, I kept switching between tools, hoping one of them would magically make me a better developer. It didn't. The biggest lesson I learned wasn't about choosing the best AI tool. It was learning how to work with AI. These days, I don't start by asking AI to write code. I start by explaining the problem. I describe the feature, the business requirements, the edge cases, and what I want the final result to look like. Sometimes I ask ChatGPT to help me plan the implementation first. Once everything is clear, I pass that plan to an agentic coding assistant and start building. That one change made a huge difference. I spend less time writing boilerplate and more time thinking about architecture, user experience, and solving the actual problem. AI still gets things wrong, so I review everything before it goes into production. But instead of writing every single line myself, I'm guiding the process. Looking back, the first few months were the hardest. Now it just feels normal. The tools will keep changing, but I think the real skill is learning how to communicate with AI and use it as part of your development process. That's something worth investing in.

Every Solana program eventually hits the same question: where do I put my data, and how do I find it again later? Programs are stateless, so a program's data lives in separate accounts, each at an address. The moment you store something, you owe an answer to a problem databases tend to hide from you: what address does this live at, and how does the program find it again tomorrow? Program Derived Addresses are Solana's answer. The name scares people off, but the idea is mostly "an address you compute instead of remember, that only your program can control." The problem, in code Say each user gets a counter account. The normal way to make an account is to generate a fresh keypair and store data at its public key: import { Keypair } from " @solana/web3.js " ; const counter = Keypair . generate (); // counter.publicKey is something random, e.g. 7Hx4...9fT // create the account at that address, write count = 0 It works. But the address is random, so nothing connects this user to that address . Tomorrow, when the user comes back to increment, how does your program find their counter? You're forced to keep a lookup table somewhere: // the mapping you now have to store and never lose const counters = { " 9fYL...user1 " : " 7Hx4...9fT " , " B2k9...user2 " : " Qz1p...4dR " , // ...times ten thousand users }; Lose that table, lose the data, even though the accounts are right there on chain. You're storing files in a warehouse and writing the shelf number on a sticky note. The fix: compute the address from what you already know What if the address were a function of the user instead of random? Give a function the word "counter" and the user's public key, and it hands back a fixed address. Same inputs, same address, every time. No table. That's a PDA. PDAs are 32-byte addresses derived deterministically from a program ID and a set of seeds. The seeds are the meaningful inputs you pick (here, "counter" + the user's key). With @solana/web3.js , the library Anchor's client uses: im

OpenAI’s Bonnie Xu discusses Kepler, an internal AI data analyst agent built to query 600+ petabytes of data. She explains how they overcome context window limits using MCP, automated code crawling, and RAG. Xu also shares how their team leverages scoped semantic memory for self-learning and utilizes AST-based LLM grading to build a robust, regression-free evaluation pipeline. By Bonnie Xu

We break down the current iPad lineup to help you figure out which of Apple’s tablets is best for you.

CircleCI has launched Chunk Sidecars, a new capability designed to bring CI-style validation directly into an AI coding agent's inner development loop By Craig Risi

Forget about patchy internet connections and dead spots in the house. These WIRED-tested multiroom mesh systems will get you online in no time.

Researchers say these coral strongholds may help repopulate more degraded reefs.

Even after your movies end, these art televisions look stunning on any wall.

Miami-based AI startup Subquadratic came out of stealth mode last month with a huge claim. It announced that it had solved a mathematical bottleneck that had been holding back large language models for almost a decade. The details were thin, and many people were unconvinced. But Subquadratic has started to bring the receipts, sharing the…

Apple’s fall macOS release will let you build Shortcuts by typing what you want to happen. But Claude Code and Codex users don’t have to wait.

Find the right wearable for your lifestyle, workouts, and goals.

Single-model systems are simple. Multi-model systems are powerful. The challenge isn't choosing models — it's designing the architecture that orchestrates them. A multi-model system isn't about having more models. It's about having the right model for the right task at the right time. Architecture patterns Five patterns cover most use cases: Pattern Complexity When to use Tradeoff Single Model Lowest Prototyping, simple tasks Limited capability Sequential Low Multi-step workflows Higher latency Parallel Medium Independent tasks Higher cost Hierarchical High Complex reasoning Complex orchestration Ensemble Highest Critical decisions Highest cost Pick the simplest one that works. Complexity is real, and it compounds. Sequential architecture Process tasks through a chain of models, each specializing in a step. Pattern 1: Pipeline Pipeline pattern — each model's output feeds the next: class ModelPipeline : def __init__ ( self ): self . models = [ { " model " : " qwen2.5-1.5b " , " task " : " classify " }, { " model " : " qwen2.5-7b " , " task " : " extract " }, { " model " : " qwen2.5-32b " , " task " : " reason " }, ] def process ( self , input : str ) -> str : current = input for model_config in self . models : current = self . call_model ( model_config [ " model " ], self . create_prompt ( model_config [ " task " ], current ) ) return current Latency adds up. Three models in sequence means three times the latency. Only use this when each step actually needs a different model. Pattern 2: Router Router pattern — classify the task, route to the specialist: class ModelRouter : def __init__ ( self ): self . classifier = " qwen2.5-1.5b " self . specialists = { " code " : " qwen2.5-coder-7b " , " math " : " qwen2.5-32b " , " creative " : " claude-sonnet-4 " , " general " : " qwen2.5-7b " , } def route ( self , prompt : str ) -> str : task_type = self . classify ( prompt ) model = self . specialists . get ( task_type , self . specialists [ " general " ]) return self . call_m

There is a comfortable lie that has taken root in information security domain. It goes like this: "We have invested in the best tools. We have CSPM. We have CNAPP. We have ASM, ASPM, EDR, SIEM. We are covered." Boards believe it. CISOs present it. Security budgets are built around it. And while everyone is looking at dashboards full of green, two things are quietly true: One — attackers are not trying to beat your tools. They are looking for what your tools were never designed to see. Two — the most dangerous gaps in your security posture today are not gaps in your tooling budget. They are gaps in visibility, knowledge, and the fundamental understanding of what your own systems are supposed to do — and whether they actually do it. No tool solves that. Only people do. The Tool Trap The security industry has industrialized the idea that protection is a purchasing decision. Spend enough. Deploy enough. Integrate enough. And you will be secure. This thinking has produced something genuinely useful — a generation of powerful tools that automate detection, surface known misconfigurations, and reduce the manual burden on security teams. CSPM catches publicly exposed storage buckets. EDR detects malware. SIEM correlates suspicious behavior across logs. These tools matter and they work — within the boundaries of what they were designed to see. But there is a boundary. And most organizations have no idea where it is. The boundary is this: every security tool operates on known patterns, defined rules, and observable configuration state. None of them operate on context. None of them understand intent. And intent — what a system was supposed to do, how it was supposed to be accessed, what data it was never supposed to expose — is exactly where the most dangerous vulnerabilities live today. The Dangerous Gap Nobody Has Named Yet Security practitioners are familiar with the concept of IoM — Indicator of Misconfiguration. A wrong setting. An overpermissioned role. A flag that flipp

PSYONIC's prosthetic touch data is now training ABB robots. Gatik signed the first Fortune 50 commercial autonomous freight contract with PepsiCo. Burro drove Physical AI onto the construction site. Experts set $20k as the humanoid price target. And someone just called Edge AI the Windows of robotics. This week, Physical AI crossed three invisible lines at once. A company that makes prosthetic hands figured out that the touch data from amputees is exactly what industrial robots need to learn how to grip. A Fortune 50 company signed not a pilot but a commercial contract for autonomous freight. A 44-horsepower robot drove off the warehouse floor and onto the construction site. And two separate conversations about software and pricing suggest that the next wave of robotics adoption will be driven by access, not capability. Here is what happened, and why it matters beyond the headlines. Value Description Fortune 50 PepsiCo becomes first to sign a commercial contract for autonomous freight with Gatik $20k Target price point for humanoid robots, Robotics Summit consensus: achievable by 2028–2030 1M hours Burro's field experience backing the Grande 44 autonomous outdoor platform 100+ Pressure sensors per fingertip in PSYONIC's Ability Hand, now training ABB GoFa A Prosthetic Hand Is Now Teaching an Industrial Robot How to Grip The standard approach to teaching a robot how to handle objects has been simulation, teleoperation, or labor-intensive physical demonstrations. PSYONIC and ABB just introduced a different source of data : the hands of people who have already learned to feel again. PSYONIC's Ability Hand is a prosthetic with more than 100 pressure sensors per fingertip . The company has been collecting kinesthetic data from users with upper-limb amputations. That data, which captures how a human hand adjusts grip pressure, contact area, and force across thousands of everyday tasks, is now being fed as training data into ABB GoFa robot arm models. The implication is no