标签：#X

Cuando el objetivo es algo tan delicado como un diagnóstico asistido, confiar en un único modelo es arriesgado. FeliniAI usa tres pipelines complementarios que se refuerzan entre sí, igual que un veterinario combina lo que ve, lo que mide y lo que sabe. Pipeline 1 — Visión: MobileNetV2 Una CNN MobileNetV2 (PyTorch, transfer learning) clasifica imágenes de la piel/pelaje del gato en categorías visuales. Elegí MobileNetV2 por su equilibrio entre precisión y ligereza: corre rápido en CPU, lo que mantiene la inferencia por debajo de 1 segundo. Alcanza un 93,4% de accuracy visual . Pipeline 2 — Clínico: XGBoost + ICADA El núcleo del sistema es un clasificador XGBoost que trabaja sobre 33 features clínicas derivadas de los criterios ICADA (los criterios estandarizados de dermatitis atópica felina): estacionalidad, distribución de las lesiones, prurito, respuesta a tratamientos previos. Sobre un dataset de 8.000 casos , este módulo logra un F1 macro de 0.9675 en validación cruzada 5-fold. La búsqueda de hiperparámetros se hizo con Optuna y la explicabilidad con SHAP. Pipeline 3 — LLM: la síntesis Finalmente, un LLM ( Llama 3.3 70B vía Groq ) integra las salidas de los dos modelos anteriores y las traduce en una recomendación legible: qué tipo de alergia es más probable, con qué confianza y qué pasos sugerir. El LLM no diagnostica solo: orquesta y comunica lo que han calculado los modelos especializados. Por qué tres pipelines y no uno Porque cada uno cubre el punto ciego del otro. La visión capta lo que una foto muestra pero un cuestionario no; el modelo clínico capta el historial que una foto no puede mostrar; el LLM convierte ambos en algo accionable. Es un patrón de ensemble heterogéneo aplicado a datos de naturaleza distinta. Resultados F1 macro (clínico): 0.9675 , accuracy 0.9909. Accuracy visual: 93,4%. 4 tipos de alergia, 33 features clínicas, <1s de inferencia. Qué aprendí Que en dominios sensibles, la arquitectura correcta no es "el modelo más grande", sino varios

Most mobile apps ship a privacy policy as a link that opens Safari. A smaller number ship an in-app rendered policy. A tiny minority ship what I think is the single highest-leverage privacy screen: a summary that mirrors your Nutrition Label in plain language, designed to be read. Why it's high-leverage The privacy summary sits at the intersection of three concerns: Users read it before granting sensitive permissions. Especially for camera, contacts, location, and health data. App Store reviewers check it exists and matches your store listing. Regulators (GDPR, CCPA) reward it. Clarity is a compliance signal, not a legal defense — but it's what an investigator asks for first. Sixty minutes of design work. Meaningful trust dividends. Often the difference between first-submission approval and a rejection loop. What to include The privacy summary should mirror your Privacy Nutrition Label but with real language humans can parse: Each data category you collect — displayed as a card or a row, with a clear icon. Why you collect it — one line, plain language. "So you can log in on another device," not "for authentication purposes." Where it's used — is it stored on our servers, shared with third parties, only on your device? Be specific. How to opt out or delete it — a link or a button to the settings page where the user can act on that category. Six to eight cards, one per data category. No more. Visual patterns that read as trustworthy Design choices that consistently score high in user-testing for trust signal: Real language, not legalese. "We store your email address so you can log in" beats "Personal identifiers are retained for authentication purposes." Muted, confident colors. No warning reds, no compliance yellows. A neutral surface with soft accent for the data-category icons. Readable typography. 16pt body, generous line-height (1.5x), enough paragraph spacing that scanning is easy. Icons per category (not just text). A camera icon for camera data, a location pin

I maintain Fulmine , a drop-in replacement for Express 5 that runs on µWebSockets.js instead of node:http . One line changes: const express = require ( " fulmine.js " ); // instead of require("express") Your middleware keeps working: helmet , cors , passport , morgan , multer , express-session and the rest. The numbers are not mine Benchmarks published by a project about itself deserve suspicion, so let me use somebody else's. HttpArena runs every framework on the same 64-core machine, in containers, under the same rules, and publishes the results. Express and Fastify are on that board too. Requests per second, from their published runs: Profile Fulmine Express Fastify Baseline (query parsing) 1,220,308 607,777 711,263 JSON (dataset + serialization) 1,111,187 395,361 522,201 Short-lived connections 1,026,789 278,163 298,779 Pipelined 7,259,814 1,009,543 1,671,338 Mixed API workload, 16 CPUs 126,282 67,724 75,633 Async Postgres 222,701 169,687 179,169 Upload (20 MB body) 2,154 2,104 1,902 That is 2.0x Express on the baseline, 2.8x on JSON, 3.7x on short-lived connections, 7.2x pipelined , and 1.9x on the mixed API profile. Against Fastify, on the same board, it is 1.7x on the baseline and 2.1x on JSON. Now the honest parts, which matter as much as the table. Look at the upload row: 1.02x. A 20 MB body is memory bandwidth and syscalls, not framework code. Everywhere the cost belongs to a library both servers call, the difference disappears: JSON.parse , zlib, OpenSSL. Speed comes from the framework only where the framework is doing the work. My entry runs in the arena's "tuned" mode, Express's and Fastify's run in "standard". On two profiles I left out of the table, static files and compressed JSON, that difference is decisive, because tuned mode allows hand-written compression and negotiation. Those rows would show 23x and 8x, and they would be measuring my entry's tuning, not the framework. I would rather not quote them. Where the speed comes from Not from one trick

Intro Day 20! I lined up 10 AIs that turn a single photo into a few seconds of video. Half ran locally on my DGX Spark, half in the cloud 🐱 What I used: DGX Spark (LTX-2.3 / Wan 2.2) / 8 cloud models via fal.ai / ComfyUI / ffmpeg The setup Item Value Input One identical photo (my cat on a desk) Length 6 seconds Settings Identical The only variable The prompt Easy prompt The cat looks at the camera and meows once. It opens its mouth, meows, then closes it. Its tail flicks and its ears twitch. Hard prompt The cat stands upright on its hind legs in a kitchen, wearing a small apron, holding a knife in its front paws and chopping vegetables on a cutting board. Steam rises from a pot behind it. Please, just watch it Some of the cats came out with very long legs. Anyway. First half is the easy prompt, second half the hard one. On the easy prompt, local and cloud were a fair match . On the hard one... cloud, I think...! Three rankings below. Ranking 1: Time Time per 6-second clip on the hard prompt. Rank Model Where Time 🥇 LTX-2.3 Cloud 41s 🥈 Wan 2.7 Cloud 92s 🥉 Happy Horse 1.1 Cloud 97s 4 Veo 3.1 Cloud 128s 5 Kling 3 Pro Cloud 205s 6 Seedance 2.0 Cloud 210s 7 LTX-2.3 Local 315s 8 Wan 2.2 Local 651s 9 daVinci-MagiHuman Cloud 710s 10 HunyuanVideo 1.5 Cloud 796s A 19x spread. Look at 1st and 7th. Same model, LTX-2.3 , nearly the same resolution. The only difference is where it ran — 7.6x . Local setup DGX Spark (GB10, 128GB unified memory, ~273GB/s). ComfyUI headless, workflows over its API. LTX-2.3 is distilled fp8 at 8 steps. At 1088×1920 peak memory hit 77.8GB, about 60% of 128GB. That was the ceiling. Dropping to 512×768 finishes in 70s, but with one-fifth the pixels. Wan 2.2 is I2V-A14B fp8, 20 steps, 480×640. Higher resolution does not finish in reasonable time. Ranking 2: Cost Rank Model Per 6 seconds 🥇 Local Electricity only 🥈 LTX-2.3 (cloud) $0.36 🥉 Wan 2.7 $0.90 4 Kling 3 Pro $1.01 5 Happy Horse 1.1 $1.08 6 Veo 3.1 $2.40 7 Seedance 2.0 $4.09 — HunyuanVideo / MagiHum

I asked an agent to swap one colour value across a file. It did. It also rewrote the line that defined the value in the first place, so the definition now pointed at itself. Nothing errored. Nothing warned. The instruction ran perfectly, which is the whole problem. Every one of these has the same shape A single condition matched more than I meant, and everything that matched got changed. The second one I still think about: hiding a set of shadow rectangles also hid a keyboard, because the keyboard's parts satisfied exactly the same single condition. Again no error, again a clean report of success. Once you see the pattern it's everywhere. It isn't a model being careless. It's an instruction that was less precise than it felt while writing it, executed with total literalness by something that has no idea what any of these objects are for. The rule: scope, and two conditions, never one Name the region it may touch. Not "the file" — this section, these frames, this layer group. Then give it two properties that must both be true. Not "everything with this colour" but "everything with this colour, inside this region, that is a fill rather than a definition". The second condition is doing the real work: it's what stops the match spreading into things that happen to share one attribute. It's a small amount of extra writing. It's the difference between a change and an incident. It cannot see the result — that's the fixed constraint An agent writes the change, the change renders somewhere it has no eyes on, and it reports success based on the instruction completing rather than the outcome being right. People treat that missing feedback loop as a tooling problem, something that will be solved in a future version. I don't think it is one. It's a sequencing problem, and sequencing is available today. The loop can't be closed by the agent. Fine. It can still be closed by a person — just not a hundred times. One, then all Run the operation on a single representative case. Render

On one build I found 127 places bound to a raw colour instead of a named role. Every single one had passed visual review. They all surfaced the moment someone asked for dark mode. That number is the whole argument. Not because 127 is large, but because none of them looked wrong. A value that was typed in and a value that came from the system are visually identical. The difference only exists in what happens next. The failure isn't that the agent breaks the rules It's that it extends them. Give an agent a design system and ask it to build. When it reaches something the system covers, it uses the system — genuinely, reliably. When it reaches something the system doesn't cover, it does not stop and ask. It invents. And what it invents is a name that sounds exactly like one of yours, sitting right next to the real ones, reading as though someone chose it on purpose. That's why this is so hard to catch by eye. A fabricated token isn't a glaring error. It's a plausible one. Six months later nobody can tell you whether it was a deliberate exception or a hallucination, and by then five components depend on it. Readable is not the same as closed Making a library available to an agent gets you components it will reuse. It does not get you a closed set. A closed set means: these values exist, everything else does not, and anything outside them fails loudly rather than passing quietly. The distinction sounds pedantic and it decides everything. A readable system produces output that mostly matches. A closed system produces output you can audit. Which is the real test I'd apply to any AI design setup: not how much of your system it covers, but what happens to the things it doesn't cover. If those slip through silently, coverage is irrelevant — you've just made the drift harder to spot. Layers, and not reaching past them Tokens have layers for a reason. Base values underneath — the raw material. Named roles on top — what a value is for. And the product interface binds to the role,

This is the stretch nobody films. The demo ends at the screenshot; the job ends about a week later, in a Figma file that someone else has to be able to open without you in the room. It's also where roughly 40% of the work lives, and where most AI-assisted design quietly falls over — not because the screens are bad, but because nothing in them is addressable. Import destroys the names Bring generated markup into Figma and everything arrives as a frame inside a frame inside a frame, with names that mean nothing. The structure survives. The meaning doesn't. The instinct at this point is to start componentising from what's on the canvas — find a button in a screen, make it a component, move on. Don't. That tree is a rendering artefact. Build your library from it and you inherit every accident in it: wrapper divs promoted to components, layout containers baked into masters, the same element modelled three different ways because it appeared in three different screens. The source markup is the specification. It knows what each thing is. So the first move is reading it and producing a record of what should exist and what it should be called — then renaming against that record, then componentising. Rename first, componentise second. Reversing those two costs more than any other ordering mistake in this stage. Library first, screens second Masters get built in a clean library section, not harvested from inside screens. The difference shows up in what ends up inside the component. Harvested masters carry their surroundings — a padding wrapper that belonged to the screen, a demo label, a background that existed to make it visible on a dark canvas. Those things then travel into every instance, and six months later somebody is asking why every card has eight pixels of phantom padding. Same-structure things get grouped into a variant set rather than left as separate components. A button that arrives as five unrelated components instead of one set is the single most common breakage

Hard debugging is rarely defeated by a lack of tools. It is defeated by three quiet habits: assuming the fault is where the symptom appears, clinging to the first explanation, and hoping a tool will do the thinking. A difficult fault is usually lost to those habits before you read a line of code. The engineers who resolve hard faults are the ones who notice these defaults and replace them with a patient, evidence-first mindset. Most hard bugs are lost before we touch them, in the attitude we bring to the session. When something breaks, the average person rushes in with three quiet habits: they assume the fault lives exactly where it shows up, they cling to the first explanation their mind offers, and they hope a tool or a smarter person will tell them what to do next. Those habits feel natural, but on hard faults they are exactly what keep us stuck. Put two engineers on the same failing board. One finds a way through in an afternoon; the other is still going three days later. The difference is rarely raw intelligence or how many commands they know. It is the mental posture each brings to the work before the first step. Handling a hard debug session is less about knowing every tool and more about managing your own assumptions, reactions, and impatience. A tough problem is usually lost in your mindset before it is lost in your methods. Habit one: starting too narrow The first habit is to fix on the most visible symptom and refuse to look anywhere else. Something breaks, so we stare at the last thing we changed, and we return to it because it is familiar and close at hand. When the answer is not there, we look harder in the same place instead of stepping back. Here is what that looks like on real hardware. A device keeps dropping off the bus. You are a kernel person, so you open the driver and read it, carefully, for three days: the probe path, the error handling, the power-management callbacks. Every line is correct, and the device still fails. The fault was a layer b

xAI's Grokipedia, an online encyclopedia with AI-generated articles that Elon Musk once promised would be a "massive improvement" over Wikipedia, apparently hasn't been updated since April 24th, according to a report from Lawfare. "As far as we can tell, no entry has changed in more than three months," Lawfare said. Grokipedia launched in v0.1 in […]

LinkBreeze is a self-hosted alternative to Linktree. I built it because Linktree's $15/mo Pro plan didn't justify the feature set, email capture is another $9/mo, embed widgets are paywalled, link scheduling is paywalled. I wanted something I actually own: my data on my server, no subscription, no tracking pixels. The interesting technical bit: the public page ships zero client-side JavaScript. The entire link-in-bio page, themes, animations, hover effects, QR codes, embed widgets, renders server-side as pure HTML/CSS. No React runtime, no hydration, no framework JS. The visitor downloads HTML + CSS + their fonts. Page loads in under 300ms. That's it. Feature gap vs. the competition (what pushed me to build this): Feature Linktree LinkStack LittleLink Shako LinkBreeze Price $15/mo Free Free Free Free Admin Panel ✅ Slow ❌ ❌ ✅ Fast Multi-Page Paid ❌ ❌ ❌ ✅ Migration Wizard ❌ ❌ ❌ ❌ ✅ Built-in Analytics Paid Basic ❌ ❌ ✅ Full External Analytics ✅ ✅ ❌ ❌ ✅ Email Capture Paid ❌ ❌ ❌ ✅ Embed Widgets Paid ❌ ❌ ❌ ✅ Link Thumbnails Paid ❌ ❌ ❌ ✅ Link Scheduling Paid ❌ ❌ ❌ ✅ Themes Paid Limited CSS only Config ✅ Full Token System + Import/Export Custom CSS ❌ ❌ ✅ ❌ ✅ Language Closed PHP HTML Astro TypeScript Docker Deploy N/A Complex Simple Simple One command License Closed AGPL MIT GPL MIT Live demo (read-only): https://linkbreeze-demo.omnirise.dev/alex Admin demo: https://linkbreeze-demo.omnirise.dev/login (demo / demo1234) Repo: https://github.com/Manak-hash/LinkBreeze I'd genuinely appreciate feedback, bug reports, or feature suggestions. What's missing compared to what you'd expect from a self-hosted tool like this?

Builder Journal · Mars Environmental Dynamics Analyzer (MEDA) Virtual Sensor Recovery Ten times in a row I predicted what my next submission would score before I uploaded it. The worst miss was 0.0025 on a number around nineteen. I took that as confirmation that the physics underneath was correct. It was confirmation that I can do arithmetic. Two days before this competition closed I pointed a review at my own endgame, expecting notes about the code. It came back with three errors and none of them were in the code. All three were in my reasoning, and all three had the same shape: I had run something that felt like a measurement and was not one. This is the fourth entry in this series and the one I would keep if I had to burn the other three. The models are competition-specific. This part is not. The competition in one breath Perseverance carries an environmental station called MEDA. Some of its surface pressure readings are missing, and the competition is to reconstruct them. Scored on mean squared error. The wrinkle is the split. Training covers sols 1 through 100, when pressure is climbing toward its seasonal peak. Test covers sols 201 through 300, when it is falling hard toward the aphelion minimum. Sols 101 through 200 do not exist in either file. Every prediction is outside the range the model was fit on. The first entry covers the first submission, which contained no machine learning at all and took the top of the board at 61.04. Six weeks and seven versions later the public score was 18.99. Almost everything in between was selected by one signal. Not cross-validation. Cross-validation here can only hold out sols from the rising limb, so it is structurally blind to the regime I am scored on. The leaderboard was the only thing that could see the falling limb, so the leaderboard picked every scalar that mattered: the residual shrink, the blend weight, a constant seasonal offset, a diurnal scaling. Hold onto that. It becomes the joke about four hundred words from

Although no one captured a direct image of the impact, scientists have detected signals confirming the rocket hit the lunar surface. Orbiters could provide the first images of the crash site in the coming days.

Baseboard management controllers from the world's biggest manufacturers are a security mess.

X head of product Nikita Bier is stepping down and says he will move into a role as an advisor, writing that "it's time to pass the torch and demote myself to my natural state: a poster." He shared the update just over a month after celebrating his one-year anniversary on the job, and just […]

The serial entrepreneur joins the e-commerce company as CPO to lead its AI agents.

SpaceX won't build large cell towers but plans small base stations across US.

Bridging Design and Code to Empower Local Businesses As a full-stack developer specializing in JavaScript and React, one of the most exciting ventures I'm currently on is building ready-made websites and Next.js templates through Softchic. This isn't just about coding; it's about deeply understanding the needs of businesses, particularly within the vibrant and rapidly evolving Nigerian market, and translating those into high-performance, beautiful web solutions. Why Next.js? Performance, SEO, and Developer Experience My choice of Next.js as the primary framework for these templates was deliberate: Performance: Server-side rendering (SSR) and static site generation (SSG) capabilities are crucial. In areas where internet speeds might vary, a fast-loading website isn't just a nice-to-have; it's essential for user retention and conversion. SEO: For businesses looking to establish a strong online presence, robust SEO capabilities out-of-the-box mean our templates provide a solid foundation for discoverability. Developer Experience: Building with Next.js allows for efficient development, leveraging the power of React while simplifying routing, data fetching, and API routes. This means faster iteration and higher quality templates. The Nigerian Market: Unique Challenges, Immense Opportunity Crafting templates specifically for the Nigerian market presents a fascinating set of considerations: Design Aesthetics: Understanding local preferences in terms of color palettes, layouts, and user flows is critical. It's not just about what looks good globally, but what resonates locally. This is where my dual role as creative director for promotional materials comes into play – applying that eye for design directly to the templates. Mobile-First Mentality: A significant portion of internet users in Nigeria access the web via mobile devices. Every template is meticulously designed with a mobile-first approach to ensure optimal responsiveness and user experience on smaller screens. Aff

It's as much about grounded characters and well-timed emotional beats as CGI-filled action extravaganzas.

Disney is interested in "price-sensitive" streaming customers.

Once, I had some questions about why SpaceX, Elon Musk's healthiest company, acquired xAI, his sickliest one. Now I have some questions about why we're calling the whole thing SpaceX. Look, what we have here, by revenue, is primarily a telecom company and a company that rents compute, according to SpaceX's first quarterly earnings statement […]