标签：#AR

Mach's approach differs sharply from some of its peers.

AI makes us faster, but does it make us better engineers, or just more dependent? As a follow-up...

Let me confess something a little creepy. I have a habit of peeking at other people's dev posts. Not stealing the writing — relax. I run a tiny read-only job that fetches the public pages on dev.to, Zenn, and Qiita and counts only the boring parts: titles, post times, like counts. Who published what, at what hour, and how far it traveled. Then it tallies the lot. The reason is petty: my own posts weren't landing. The content is already in my hands — so I wanted to know how much the rest, the when and how you publish , actually moves the needle. By the numbers, not by gut. So I counted across three platforms. And the conditions that make a post fly turned out to be roughly mirror images between Japan (Zenn / Qiita) and the English-speaking world (dev.to). Here's the story. First, my most important disclaimer This post is full of numbers, so let me put up a guardrail before any of them. This is correlation, not causation . A result like "weekend posts don't do well" could mean the weekend itself is bad — or it could mean people who post on weekends are just dashing something off on the side. The data can't separate those. Please read it that way. Also, I only keep aggregate numbers I computed myself . I don't store or reuse anyone's article body (read-only GET, count the features, throw the page away). I peek, but only at the overall shape . Nobody gets singled out here. With that out of the way — four findings I enjoyed. 1. The best hour to publish is just your readers' time zone This one came out cleanest. On Qiita , posts published in the morning win (+32pt in the GOOD group). Midday is +14pt. Evening is -32pt, late night -14pt. Zenn likes midday too (+27pt). Late night is -15pt. dev.to is the exact opposite. Late night Japan time scores +7pt — Japanese evening is actually weak. The trick is obvious once you see it. dev.to's readers are English-speaking, mostly US. Late night in Japan is the US working day. Zenn and Qiita readers are in Japan, so the Japanese morni

13.5.0 Before We Begin During its design, Rust drew inspiration from many languages, and functional programming had a particularly strong influence on Rust. Functional programming often includes passing functions as values to parameters, returning them from other functions, assigning them to variables for later execution, and so on. In this chapter, we will discuss some Rust features that are similar to what many languages call functional features: Closures Iterators (this article) Improving the I/O Project with Closures and Iterators Performance of Closures and Iterators If you find this helpful, please like, bookmark, and follow. To keep learning along, follow this series. 13.5.1 What Is an Iterator To talk about iterators, we first need to talk about the iterator pattern. The iterator pattern allows you to perform a task on each element in a sequence, one by one. In that process, the iterator is responsible for: Traversing each item Determining when the sequence has finished iterating Rust iterators are lazy: unless you call a method that consumes the iterator, the iterator itself does nothing. In other words, if you write an iterator in your code but never use it, it is as if it did nothing at all. Take a look at an example: fn main () { let v1 = vec! [ 1 , 2 , 3 ]; let v1_iter = v1 .iter (); } v1 is a Vector , and v1.iter() creates an iterator for v1 and assigns it to v1_iter . But v1_iter is not used yet, so the iterator can be considered to have no effect. Now let’s use the iterator to traverse the values: fn main () { let v1 = vec! [ 1 , 2 , 3 ]; let v1_iter = v1 .iter (); for val in v1_iter { println! ( "Got: {}" , val ); } } This is equivalent to using each element in the iterator once in a loop. 13.5.2 The Iterator Trait All iterators implement the Iterator trait. This trait is defined in the standard library and looks roughly like this: pub trait Iterator { type Item ; fn next ( & mut self ) -> Option < Self :: Item > ; // methods with default implementa

13.4.0 Before We Begin During its design, Rust drew inspiration from many languages, and functional programming had a particularly strong influence on Rust. Functional programming often includes passing functions as values to parameters, returning them from other functions, assigning them to variables for later execution, and so on. In this chapter, we will discuss some Rust features that are similar to what many languages call functional features: Closures (this article) Iterators Improving the I/O Project with Closures and Iterators Performance of Closures and Iterators If you find this helpful, please like, bookmark, and follow. To keep learning along, follow this series. 13.4.1 Closures Can Capture Their Environment Closures have a capability that functions do not: a closure can access variables in the scope where it is defined. Take a look at an example: fn main () { let x = 4 ; let equal_to_x = | z | z == x ; let y = 4 ; assert! ( equal_to_x ( y )); } The closure part is: let equal_to_x = | z | z == x ; Some people may find it hard to distinguish the roles of = and == here, so let’s rewrite it another way: let equal_to_x = | z | { z == x ; } In other words, the closure takes z as its parameter, compares it with x (which is 4, because x = 4 was defined above), and returns a boolean. If they are equal, the result is true ; otherwise it is false . Here the closure directly accesses the variable x in the same scope, which functions cannot do. But this feature has a cost: it introduces memory overhead . In most cases we do not need a closure to capture its environment, and we do not want the extra overhead either. That is why functions are not allowed to capture variables from the environment, and defining and using a function never introduces this kind of overhead. 13.4.2 How Closures Capture Values From Their Environment Closures capture values from the environment in three ways, just like functions receive parameters in three ways: Taking ownership, whose trait

Pagination is a key component on web-applications that let users navigate through pages making easy to read/find records. Also, pagination is a great strategy to improve performance by avoiding to load entire dataset at once. However, while working with Kaminari, a popular pagination gem in Rails, I encountered an unexpected issue that revealed an interesting edge case. Identify the issue Basically, pagination in frontend was not working properly. Datatable should load 316 total rows, although when the user started to load 15 records per page, frontend is showing inaccurate total rows. Multiple of 15 should ends at 0 or 5. There were pages with 64 records, crazy world. Some pages were loading 12 or 11 rows. There is no issues or error messages in the frontend or backend. Lost in debugging-land After discarding Angular frontend errors, I started to dig into backend controller and Kaminari configuration. Nothing seems wrong. Everything looked good: test suite, smoke tests, desktop debugging. Despite of test results, I started to wonder: what if returned-data is wrong after all? and... Bingo! Finally, after checking every response I noticed that there were duped records in two different pages(pagination requests). Those duped records were skipped from Angular data-table and that's why loaded/total rows did not match. Bingo: a sort of mistake This tricky bug has a simple explanation: bad sorting. Kaminari uses a SQL query using LIMIT/OFFSET strategy: SELECT * FROM posts ORDER BY id LIMIT 25 OFFSET 0 ORDER BY : sort the collection. LIMIT : number of records per page. OFFSET : is used to skip a specified number of rows before starting to return rows from a query. This works perfectly using ORDER BY id because primary key is unique. Check table A. id title body created_at lock 101 Welcome to the Platform First post introducing the new platform features. 2026-06-16 08:15:22 false 102 Summer Update Announcing the latest improvements and updates. 2026-06-16 08:15:22 true 103

업무 자동화 시스템을 만들 때 가장 먼저 드는 질문이 있다. "에이전트 하나면 안 되나?" 단일 에이전트로 시작하면 구조가 단순하고 디버깅도 쉽다. 그런데 실제 업무 맥락에서는 단일 에이전트가 빠르게 벽에 부딪힌다. 이 글은 왜 여러 에이전트가 협업하는 구조가 필요한지, 그리고 그 구조를 실제로 어떻게 설계하는지를 기술적으로 짚는다. 단일 에이전트로 충분하지 않은 이유 단일 에이전트가 실패하는 지점은 복잡한 기능 탓이 아니라 컨텍스트 길이와 직렬 실행의 구조적 한계 때문이다. LLM 기반 에이전트에 하나의 긴 작업을 맡기면 세 가지 문제가 동시에 발생한다. 첫째, 컨텍스트 창이 소진된다. 데이터 수집, 변환, 검증, 발행을 하나의 루프에서 처리하면 중간 상태가 프롬프트에 누적되고, 모델은 앞부분 지시를 잊는다. 둘째, 직렬 실행은 병목을 만든다. API 호출이 5개 있고 각각 2초라면, 단일 에이전트는 10초를 기다린다. 셋째, 에러 격리가 불가능하다. 한 단계가 실패하면 전체 루프를 재시작해야 한다. 반면 여러 에이전트가 협업하는 구조에서는 각 에이전트가 명확한 책임 경계를 갖는다. 한 에이전트가 데이터를 수집하고, 다른 에이전트가 변환하고, 또 다른 에이전트가 검증한다. 에러는 해당 에이전트 범위 안에서 처리되고, 독립된 작업은 병렬로 돌린다. 에이전트 협업 구조를 어떻게 설계할까? 나무숲이 실제 자동화 프로젝트에서 가져가는 구조는 오케스트레이터-워커(Orchestrator-Worker) 패턴이다. 이 패턴은 Anthropic이 공개한 에이전트 설계 가이드라인 에서도 다루는 구조로, 책임 분리가 명확하다는 점이 핵심이다. 역할 책임 범위 주요 판단 오케스트레이터 전체 작업 계획, 워커 배정 어떤 워커를 호출할지, 순서와 병렬 여부 워커 에이전트 단일 도메인 작업 실행 도구 호출, 결과 반환 검증 에이전트 출력 품질 검사 재시도 요청 또는 다음 단계 진행 상태 관리 에이전트 간 공유 컨텍스트 보존 어떤 정보가 다음 에이전트에 전달되는지 예를 들어 콘텐츠 자동 발행 파이프라인이라면, 오케스트레이터가 "오늘 발행할 항목 목록"을 받아 수집 워커, 요약 워커, 포맷 워커를 순서대로 호출한다. 검증 에이전트는 포맷 워커 출력을 보고 발행 가능 여부를 판단한다. 에이전트 간 데이터 흐름과 오케스트레이션 설계 오케스트레이터는 각 워커를 직접 호출하고, 그 결과를 다음 워커의 입력으로 넘긴다. 이때 중요한 설계 결정이 두 가지다. 메시지 구조를 명시적으로 정의한다. 에이전트 간 데이터는 자유형 텍스트가 아니라 스키마가 있는 구조로 전달해야 한다. JSON 스키마나 Pydantic 모델을 쓰면 에이전트 출력이 다음 에이전트의 입력 형식을 충족하는지 런타임 전에 검사할 수 있다. 오케스트레이터는 작업의 의미를 이해하지 않는다. 좋은 오케스트레이터는 라우터에 가깝다. "이 입력은 A 워커에게, 그 결과는 B 워커에게"를 결정할 뿐, 각 작업의 도메인 로직에 관여하지 않는다. 이 원칙을 지키면 워커를 교체하거나 추가할 때 오케스트레이터 코드를 손댈 필요가 없다. 간단한 파이썬 예시로 구조를 보면 이렇다: from anthropic import Anthropic from pydantic import BaseModel client = Anthropic () class WorkerOutput ( BaseModel ): status : str # "success" | "retry" | "failed" payload : dict error_message : str | None = None def call_worker ( system_prompt : str , user_input : str ) -> WorkerOutput : response = client . messages . create ( model = " claude-opus-4-5 " , max_tokens = 1024 , system = system_prompt , messages = [{ " role " : " user " , " content " : user_input }],

Why Your Agent's Search Results Look Right and Are Wrong: The Index Distribution Problem You've built an agent. It has a search tool. You query it with something reasonable — a factual question, a comparison, a technical lookup — and it returns results. The results look right. The sources are real. The snippets are plausible. The agent synthesizes them into a confident answer. And the answer is wrong. Not obviously wrong. Not hallucinated-in-a-hallucinatory-way wrong. Structurally wrong — wrong in a way that passes every surface-level check because the error is baked into the retrieval layer before the model ever sees the context. This isn't a prompt engineering problem. It isn't a context window problem. It's a distribution problem , and it has a structural ceiling that no amount of better prompting will fix. The Index Is a Frozen Decision Here's the thing most agent builders don't internalize: a search index is not a neutral representation of knowledge. It's a frozen set of decisions about what matters and what doesn't. Every index — whether it's a BM25 inverted index, a dense vector store, or a commercial web search API — encodes a distribution shaped by past relevance judgments. Someone, at some point, decided which documents were "relevant" to which queries. That could be explicit (human raters labeling search results) or implicit (click logs, dwell time, link graphs). Either way, the index now encodes a probability distribution over what the system considers a good answer to a given query. That distribution is not semantic truth. It's past relevance consensus . Consider what happens when you embed a corpus and build a vector index. Your embedding model was trained on data that reflects certain assumptions about what concepts are close to each other. Your chunking strategy encodes assumptions about what granularity of information is useful. Your ranking model — whether it's cross-encoder reranking or a learned relevance model — was trained on labeled data that

We’ve all been there: waking up feeling like a zombie despite getting eight hours of sleep. While wearables give us data, they often fail to give us foresight . What if you could predict your stress levels 24 hours in advance? 🚀 In this tutorial, we are going to tackle HRV prediction (Heart Rate Variability) using a state-of-the-art Temporal Convolutional Network (TCN) . By leveraging the Oura Ring API and deep learning, we’ll transform non-stationary biometric time series into actionable insights. Whether you're into time series forecasting or building the next big health-tech app, mastering Temporal Convolutional Networks (TCN) is a game-changer for handling long-term dependencies without the vanishing gradient headaches of traditional RNNs. For those looking for more production-ready examples and advanced biometric signal processing patterns, I highly recommend checking out the deep-dives at WellAlly Blog , which served as a major inspiration for this architecture. The Architecture: Why TCN? Traditional LSTMs are great, but they process data sequentially, making them slow and prone to memory loss over long sequences. TCNs, however, use Dilated Causal Convolutions , allowing the model to look back exponentially further into the past with fewer layers. Data Flow Overview graph TD A[Oura Cloud API] -->|Raw JSON| B(Pandas Preprocessing) B -->|Cleaned HRV/Activity| C{Feature Engineering} C -->|Sliding Windows| D[TCN Model Training] D -->|Dilated Convolutions| E[Stress Trend Prediction] E -->|24h Forecast| F[Dashboard/Alerts] style D fill:#f9f,stroke:#333,stroke-width:2px Prerequisites To follow along, you'll need: Tech Stack : Python, TensorFlow/Keras, Pandas, Scikit-learn. Data : An Oura Cloud Personal Access Token (or use the mock data generator provided). Difficulty : Advanced (Buckle up! 🏎️). Step 1: Fetching Biometric Data First, we need to pull our "Readiness" and "Sleep" data. Oura provides high-resolution HRV samples (usually 5-minute intervals during sleep).

TL;DR — Harness engineering is one thing: getting reliable judgment out of a reasoner you didn't train — given an instruction, bounded by a spec that can override that instruction, and verified before the result ships. It is not "the wrapper around the model." It's a property of code, not a place in your stack, and it lives on both sides of every tool call. The generic scaffolding — loops, dispatch, memory — melts into the model a little more every quarter. What survives is the part that stays external to the model: the specification of what the agent may do, and the verification that it did. That's the discipline. The rest is plumbing. I've written before that Agent = Model × Harness , and that the harness is the half you actually engineer. I still believe the formula. But stop there and it oversells two things — so this is me correcting my own framing, precisely, with code. The two things the formula gets wrong The × oversells separability. Multiplication implies the factors are independent. They aren't. A better model doesn't leave your harness untouched — it dissolves parts of it. Chain-of-thought prompting became reasoning models. ReAct-style tool loops became native tool use. Half your RAG plumbing is being quietly absorbed by longer context and better-trained retrieval. Every model generation collapses a layer of harness the previous one needed. "Harness absorbs software engineering" oversells the annexation. It doesn't absorb it. It partly consumes it — pulls in the slice that encodes the agent's behavior — while the deterministic spine (payments, ledgers, auth) and the dumb tools stay exactly what they were. Relocation with an annexation at the seam. Not a takeover. Both corrections point at the same uncomfortable question: if the model eats the scaffold every quarter, isn't the harness the melting part — the thing you'd be foolish to bet a career on? What melts, and what doesn't Here's the resolution, and it's the whole point. The harness mechanism melts.

AI-generated answers now influence B2B buying decisions at the top of the funnel. Buyers use ChatGPT, Perplexity, Claude and Gemini to shortlist vendors before they ever visit a website. Which means the question "where does my brand actually appear in AI-generated answers?" is now a critical GTM intelligence question. Most brands don't know the answer. This post gives you the methodology to find out. What an AI visibility audit is — and isn't An AI visibility audit is not a technical SEO audit. You're not looking for broken links, page speed issues, or crawl errors (though those matter for a related reason). You're auditing citation density — how frequently and authoritatively your brand appears across the specific sources that LLMs draw from when generating recommendations. Those sources break down into 12 distinct surfaces. A brand can score 9/10 on one surface and 1/10 on eleven others — and the aggregate result is a brand that appears for some queries and not others in ways that seem inexplicable but are structurally predictable. The goal of this audit is to make that structure visible. The 12 surfaces — quick reference Before diving into the methodology, here's the full surface map: # Surface Primary platforms 1 AI Interfaces ChatGPT, Perplexity, Gemini, Claude, Copilot 2 Search + AI Search Google AI Overviews, Bing Copilot, Perplexity Search 3 Reviews + Reputation G2, Clutch, Capterra, industry-specific platforms 4 Earned Media + Publishers Trade press, business press, analyst reports 5 Owned Content + Website Brand site, blog, schema markup, answer-object pages 6 Technical + Developer GitHub, dev.to, Stack Overflow, Hacker News 7 Social + Authority LinkedIn, executive publishing, award citations 8 Data + Knowledge Graphs Wikipedia, Wikidata, Crunchbase, ZoomInfo 9 Marketplaces + Ecosystems Clutch category pages, Agency Spotter, partner directories 10 Case Studies + Proof Published case studies, award wins, client outcomes 11 Community + Q&A Reddit, Quora, Lin

If you've shipped software in the last three years, you've probably watched your job description...

Scheduling features look simple until you build them. Google Calendar speaks its own REST API with events.insert ; Microsoft 365 wants Graph and POST /me/calendar/events ; Apple and a long tail of providers expect CalDAV. The moment your app needs to read a user's events, drop a meeting on their calendar, or check whether three people are free at 2pm, you're staring down three integrations that disagree on field names, time formats, and recurrence rules. The Nylas Calendar API gives you one interface over all of them. Connect a user's account once, get a grant_id , and read calendars, manage events, send RSVPs, and compute free/busy with the same request shape whether the backing provider is Google or Microsoft. This post walks the calendar surface from both sides: the HTTP API your backend calls, and the Nylas CLI for testing the same operations in a terminal. I work on the CLI, so the terminal snippets below are the commands I actually run when I'm poking at a calendar. Calendars, events, and the calendar_id A connected account has one or more calendars , and every event belongs to exactly one of them. Most operations take a calendar_id , and the special value primary resolves to the account's default calendar — so you don't need to look up an ID to act on the main calendar. One exception: iCloud doesn't support primary , so for iCloud accounts you pass a real calendar ID from nylas calendar list . An event carries a title , a when object holding its start and end times, a list of participants , an optional location , and flags like busy . That schema is identical across providers, which is the whole point: you read a Google event and a Microsoft event into the same struct. See the Calendar API overview for how calendars, events, and availability fit together. Before you begin You need a Nylas API key and a connected account with calendar scopes. The CLI gets you there in two commands: nylas init # create an account, generate an API key nylas auth login # connect

Half a year ago I asked a simple question: during an online call, could a short, to-the-point hint appear on my screen in a second or two — while the other person is still talking? Not an after-the-fact transcript, but help in the moment. The result is a desktop assistant (macOS + Windows). Below is an honest breakdown of what turned out to be hard, and which solutions worked. Engineering only, no marketing. Architecture in one paragraph On the device there are only two things: audio capture and a thin UI overlay. All the "brains" (provider keys, prompts, model selection) live on the server. The client gets a short-lived per-session token and streams audio; the server returns the transcript and the generated answer. I picked this split not for "security theater" but because otherwise keys and prompts would have to be baked into the binary — and both leak instantly. Hard part #1: system audio, not the microphone The mic only captures you. You need the other party's audio — i.e. the system output. And that's where the platform pain starts: macOS. For a long time there was no native "give me system audio" API; the classic path was a virtual audio device (BlackHole/Soundflower-style) or, in recent versions, ScreenCaptureKit, which can hand you a process's audio. ScreenCaptureKit turned out to be the best option: no kernel extensions for the user to install. Windows. WASAPI loopback saves you — you can grab whatever is going to the output device, without virtual cables. Takeaway: "system audio capture" is not one feature but two different subsystems for two OSes, and most of the early bugs were about permissions and device selection, not about audio itself. Hard part #2: latency is everything A hint that arrives 6 seconds late is useless — the conversation has already moved on. The latency budget has three parts: STT (speech → text). Streaming only. Batch "recognize after the phrase ends" immediately adds 1–2 seconds. The key metrics weren't "overall accuracy on a benchm

Thinking about trying FtrIO? The new CLI makes it easy to start with just one feature flag If you've been curious about FtrIO (the .NET feature toggle library that replaces if (featureFlags.IsEnabled(...)) with a [Toggle] attribute woven directly into your compiled IL) but weren't sure where to start, the experimental release of FtrIO.onetwo just made that first step a lot smaller. You don't need to commit to a full migration. You don't need to rip out your existing flag library. You just need one method and 20 minutes. What FtrIO.onetwo does FtrIO.onetwo is a .NET CLI audit tool. Its default mode scans your source tree, finds every FtrIO toggle reference, and tells you exactly what's live right now: dotnet tool install --global FtrIO.onetwo --version 1.1.1-experimental ftrio.onetwo --source C: \P rojects \M yApp But in this experimental release it does two new things: ftrio.onetwo import : pulls your current flag state from LaunchDarkly, Flagsmith, flagd, environment variables, or an HTTP endpoint directly into appsettings.json . Your existing flag library keeps working unchanged. ftrio.onetwo migrate : scans your .cs files for LaunchDarkly or Flagsmith SDK call patterns using Roslyn, cross-references them against your live flag state, and generates a report showing exactly what each flag would look like in FtrIO and how to migrate it. The "try it on one flag" workflow The migrate report categorises every flag it finds: ✅ Ready to migrate : boolean flag, no targeting rules, straightforward [Toggle] replacement ⚠️ Needs review : targeting rules, number flags, needs a decision ❌ Cannot migrate : JSON flags, recommend moving to IConfiguration For every ready flag it shows the suggested refactor. Something like: new - checkout - flow → NewCheckoutFlow File : Services \ OrderService . cs : 42 Current code : if ( client . BoolVariation ( "new-checkout-flow" , user , false )) { ValidateCart (); ApplyDiscounts (); ProcessPayment (); } Suggested action : Extract the if bloc

A smart contract does not need an overflow, reentrancy bug, or broken access-control check to lose money. Sometimes, the exploit is hidden inside an ordinary division: uint256 result = amount * rate / SCALE; The expression looks harmless. It may even produce the expected answer in every unit test. But financial smart contracts operate with integer arithmetic. Fractions are discarded, rounding direction changes who receives value, and an error of one unit can be repeated across thousands of transactions. In a financial protocol, rounding is not merely a mathematical implementation detail. Rounding is a value-transfer policy. Every division should therefore answer three questions: Which direction does the calculation round? Which party benefits from that direction? Can the rounding advantage be repeated or amplified? This article examines the most dangerous precision problems in Solidity and the engineering patterns used to prevent them. Solidity Does Not Have Native Fixed-Point Arithmetic Most financial formulas use fractions: interest = principal × rate × time fee = amount × fee percentage shares = assets × total shares ÷ total assets collateral value = token amount × oracle price Solidity primarily performs these calculations with integers. For unsigned integers: uint256 result = 5 / 2; The result is: 2 The fractional component is discarded. For positive values, this behaves like rounding down: 2.5 → 2 This appears insignificant until the result represents: vault shares; debt; collateral; protocol fees; interest; rewards; liquidation bonuses; exchange rates; token prices. The lost fraction does not disappear economically. One party receives less value, while another party retains the remainder. Precision Loss Is Not Always Small Consider a protocol calculating a percentage: function calculateFee( uint256 amount, uint256 feeBps ) public pure returns (uint256) { return amount * feeBps / 10_000; } For a 0.3% fee: amount = 100 feeBps = 30 fee = 100 × 30 ÷ 10,000 fee =

In this article, we will explore how to implement an LSTM using PyTorch and Lightning . For more details about LSTMs, there is a separate series of articles available here . Imports To begin, we first import the required modules. import torch import torch.nn as nn import torch.nn.functional as F Introducing a New Optimizer We also introduce a new optimizer: from torch.optim import Adam Adam is used to fit the neural network to the data. It works similarly to SGD, but in practice, Adam often converges faster and adapts the learning rate more effectively. Lightning and Data Utilities Next, we continue with the remaining imports: import lightning as L from torch.utils.data import TensorDataset , DataLoader Defining the LSTM Model We define the neural network by creating a Lightning module. class LSTMByHand ( L . LightningModule ): def __init__ ( self ): # Create and initialize weight and bias tensors def lstm_unit ( self , input_value , long_memory , short_memory ): # LSTM computations def forward ( self , input ): # Forward pass through the unrolled LSTM def configure_optimizers ( self ): # Configure Adam optimizer def training_step ( self , batch , batch_idx ): # Compute loss and log training progress Initializing the Model Now let’s implement the __init__ method. This is where we initialize all weights and biases. class LSTMByHand ( L . LightningModule ): def __init__ ( self ): super (). __init__ () mean = torch . tensor ( 0.0 ) # Mean of the normal distribution std = torch . tensor ( 1.0 ) # Standard deviation # ------------------------- # Forget Gate (l = "lr") # ------------------------- self . wlr1 = nn . Parameter ( torch . normal ( mean = mean , std = std ), requires_grad = True ) self . wlr2 = nn . Parameter ( torch . normal ( mean = mean , std = std ), requires_grad = True ) self . blr1 = nn . Parameter ( torch . tensor ( 0.0 ), requires_grad = True ) # ------------------------- # Input Gate (p = "pr") # ------------------------- self . wpr1 = nn . Parameter

I expected bcrypt to silently drop characters past 72. I did not expect it to bake in half an emoji. That's what happens with a specific password combination I tested. The original password still works. But strip the emoji (a password manager, a different keyboard, a Unicode normalizer) and you're locked out. Your Laravel validator passed it as valid the whole time. The 72-Byte Rule bcrypt has a hard input limit of 72 bytes. Not characters - bytes. When you call password_hash($password, PASSWORD_BCRYPT) , PHP silently truncates anything past byte 72. Most developers know this in theory. But for ASCII-only apps, it never bites. 72 ASCII characters is already a very long password, and the silent clip is harmless in practice. The trouble starts with multi-byte scripts. How Many Characters Fit? Character set Bytes per char Effective bcrypt limit ASCII 1 72 chars Cyrillic 2 36 chars CJK (Chinese, Japanese, Korean, common block) 3 24 chars Emoji 4 18 chars Past the byte limit, a longer password adds no security at all. A 200-character Cyrillic password hashes identically to its own first 36 characters. Byte 73 and beyond simply do not exist from bcrypt's point of view. So "longer always produces a stronger bcrypt hash" is not true. A Cyrillic user with a 37-character password gets silently truncated at char 36. The hash is still consistent. The user logs in fine, but any variation past character 36 doesn't matter to bcrypt. Annoying from a security standpoint, but it does not break login. The Split-Byte Trap The 72-byte limit cuts at a byte boundary, not a character boundary. If a multi-byte character falls on that cut, bcrypt bakes in an incomplete UTF-8 sequence. // 35 Cyrillic chars = 70 bytes, emoji = 4 bytes, total = 74 bytes $password = str_repeat ( 'А' , 35 ) . '🔑' ; $hash = password_hash ( $password , PASSWORD_BCRYPT ); password_verify ( $password , $hash ); // ? password_verify ( str_repeat ( 'А' , 35 ), $hash ); // ? password_verify ( str_repeat ( 'А' , 36 ), $h

Many of those videos were reportedly filmed on “near-perfect copies” of the Polymarket website, while featuring trades and winnings that were not real.

Originally published on Mohamad Alsabbagh's Blog . AI coding assistants are excellent at compressing known work into fast drafts. That speed is the preface boost : routine implementation arrives almost immediately. The hidden risk is that teams begin treating the model's first plausible answer as architecture. Because LLMs are trained and aligned around historically common patterns, they can pull engineering teams toward Generative Monoculture : less diverse solutions, narrower exploration, and fewer designs shaped by the exceptional constraints of the system in front of them. Give the same prompt to three engineers using the same assistant and you often get the same shape back: a tidy service layer, a familiar API boundary, a conventional retry wrapper, and code that looks clean enough to merge. That answer is useful. It may even be the right answer for ordinary work. The danger is what happens when ordinary work becomes the default posture for extraordinary constraints. Large Language Models are not neutral architecture engines. They are probabilistic systems trained over historical work and tuned toward answers people tend to reward. Used well, that makes them extraordinary accelerators. Used passively, it creates an optimization paradox: teams gain immediate implementation velocity while becoming anchored to a consensus baseline that may be too average for the actual system. 1. The Default Is a Local Optimum Wu, Black, and Chandrasekaran define Generative Monoculture as a narrowing of model output diversity relative to the diversity available in the training data. That matters because software architecture is rarely a search for the most common answer. It is a search for the answer that fits the exact failure modes, latency envelope, team topology, regulatory constraints, and operational reality of a system. The model's default is often a local optimum: a solution that is statistically likely, syntactically polished, and broadly acceptable. That can be excellent