标签：#ai

Sandhya Devanathan will oversee some OpenAI operations across Southeast Asia and Australia in her new role.

A ideia de um sebo que não perde estoque: No primeiro período, nosso grupo desenvolveu um Sebo Virtual. O objetivo era resolver a dificuldade de sebos tradicionais em conciliar estoque físico e virtual, com pagamento via PIX e envio de recibo por e-mail. Minha responsabilidade foi a engenharia de prompt utilizando o Lovable. Quando a IA não entendia o que eu queria: Os primeiros prompts retornaram resultados incompletos. Ao solicitar "explique o código por trás da aplicação", a resposta foi genérica e não detalhou a integração com o banco de dados. Também houve dificuldade em fazer a ferramenta compreender fluxos específicos, como leilão de itens, validação de cupons e cálculo de frete por CEP. O que mudou quando usei diagrama e contexto: O resultado melhorou quando passei a incluir contexto e artefatos. Três prompts funcionaram bem: para wireframe, enviei o diagrama e solicitei o protótipo das telas; para o leilão, pedi quatro telas com checkout e histórico de transações; para o back-end, solicitei as linguagens utilizadas e o fluxo de integração ao banco preservando as informações da documentação. Com isso, identifiquei a stack gerada: React com TypeScript no frontend e Supabase no backend, com consultas como from('pedidos').select('*').eq('usuario_id', id) . Do sebo para qualquer loja online: As regras implementadas, como cupons LIVRO10 e SEBO20, frete proporcional ao peso e checkout via PIX para o endereço base na Rua dos Livros, 707, João Pessoa, são aplicáveis a qualquer e-commerce de pequeno porte. O método permite transformar uma ideia em protótipo navegável em poucas horas. O que levo disso para a carreira? O projeto mostrou que, além do código, a capacidade de formular perguntas claras e organizar a documentação em fluxograma e diagrama de classes é fundamental. Foi meu primeiro case prático e base para portfólio na área de dados e produto. EN Summary: As a first-semester student, our team built a Virtual Bookstore to manage physical and online inventory w

Confusion is a generally accepted side effect of mystery box shows. They slather on secrets with the promise of a satisfying payoff in the end, and sometimes the cast and crew even have a hard time following what's going on. But even by the standards of the genre, Dark Matter is extreme. Its multiversal premise, […]

The group infected more than 1,000 organizations in a relentless supply-chain attack campaign.

This essay was written with Kasra Rafi, and originally appeared in The Guardian. Earlier this month, about 40 top mathematicians gathered at OpenAI’s offices to discuss the future of their profession. The meeting was off-the-record, but if recent articles by mathematicians are any guide, it was mostly pretty glum. People fear for their jobs, their careers and the work they love. We think the contrary view is more likely, at least in the short-term. AI models are nowhere near as capable as experienced academic mathematicians. This isn’t to say that AIs aren’t producing stunning mathematical results at the level of PhD researchers. In mid-May, OpenAI ...

An AI agent phone is a real, or cloud-hosted, smartphone that an LLM-powered agent can operate on its own. It sees the screen, taps, swipes, types, opens apps, and completes multi-step tasks the same way a person would. Instead of calling an API, the agent uses the phone directly, the same Instagram, banking, or delivery app you'd use, driven by a model instead of a thumb. The phrase gets used two ways in 2026. Some products sell phone numbers for AI agents, voice and SMS. That's not this. Here, an AI agent phone means the device itself as something an agent controls, a full Android or iOS handset that becomes an autonomous actor. If you've heard the pitch give your AI agent a phone, this is it. Why a phone, not a browser? Most agent tooling lives in the browser, or in desktop computer use. That misses where people actually are. The world is mobile-first, and a huge share of real workflows are app-only, ride-hailing, food delivery, mobile banking, two-factor prompts, creator tools, regional super-apps. A browser agent can't install an APK, respond to a push notification, read an SMS one-time code, use the camera, or drive a native app that never ships a web build. A phone can. And there's a second reason: fidelity. When an agent operates the same app a customer uses, you're automating the real thing, not a mock, not some undocumented internal endpoint that breaks next release. How it works A mobile AI agent runs a perception-decision-action (PDA) loop against the device. The agent builds its understanding from two sources. First, the accessibility tree, the structured hierarchy of on-screen elements the OS exposes for screen readers, which gives precise, machine-readable targets. Second, vision, a screenshot passed to a multimodal model for anything the tree misses, canvas UIs, games, custom widgets. Together, the tree gives coordinates and vision gives context. The agent gets a goal in natural language, reasons about the current screen, picks the next action, and e

Launched by Belgian startup Any, LUV1 is a modular electric motorcycle with 120 liters of cargo space that can be used to carry bags, work equipment, or even pets.

By Ecaterina Sevciuc | Creator of AURA (AI User Risk Assessment) Two months ago, I launched AURA — an open-source framework designed to model psychological manipulation, grey-zone threat vectors, and social engineering in Human-AI interactions. Yesterday, I stumbled upon a Reuters report detailing how hackers exploited Cursor (running Anthropic’s Claude Sonnet) to compromise seven companies worldwide. This isn't the first such incident in the news, and I suspect it certainly won't be the last. (Side note on the attackers' group name, "Aur0ra": I can assure you that for a Russian-speaking group, this is almost certainly not a homage to the Roman goddess of dawn, but a subtle nod to the infamous historical cruiser Aurora — known for firing the shot that signaled a revolution. A fittingly dark bit of Eastern European sarcasm for a tool that overthrows AI security). Their weapon? They didn't write a zero-day exploit. They simply convinced the AI agent that the attack was "just a security simulation." The model balked a few times, felt uncomfortable, and then happily handed over the keys. As an AI Safety architect with a background in banking compliance and legal risk evaluation, watching Big Tech react to this is painful. They are building multi-billion-dollar static guardrails while AI agents are being tricked by the oldest psychological tricks in the book. The Fatal Flaws of Modern AI Guardrails Big Tech’s approach to AI safety is fundamentally broken because it relies on Static Keyword Filtering & Single-Language Heuristics : Rule Evasion: If a prompt contains "how to build a bomb" , the model blocks it. But if the exact same request is framed as "I am a researcher simulating a crisis scenario for an academic paper," the model complies. Linguistic Blind Spots: Guardrails are heavily aligned on technical, low-complexity English. Synthetic, morphologically rich, or non-Indo-European languages (like Russian, Arabic, or East Asian language groups) leverage complex idioms

I’m setting up a local AI development environment on Windows + WSL2 and I’m trying to decide between two architectures. Option 1 — Ollama/Models on Windows WSL2 ┌───────────────────┐ │ Application │ │ ├── Python │ │ ├── .venv │ │ └── Source code │ └───────┬───────────┘ │ HTTP localhost:11434 │ ▼ Windows ┌───────────────┐ │ Ollama │ │ ↓ │ │ Models │ │ ↓ │ │ GPU │ └───────────────┘ Option 2 — Ollama/Models inside WSL2 WSL2 ┌─────────────────────────┐ │ Application │ │ ↓ │ │ Ollama │ │ ↓ │ │ Models │ └────────────┬────────────┘ │ GPU access │ ▼ Windows ┌─────────────────────────┐ │ GPU / Driver │ └─────────────────────────┘ My current setup is Option 1 , and it works: WSL2 can access the Windows Ollama API through localhost:11434. But I’m wondering if Option 2 is a better long-term architecture for local AI/LLM development. I’m especially interested in: 🚀 Performance 🎮 GPU utilization 🧠 Model management 💾 Disk usage 🔧 Setup and maintenance 🐧 Linux/ML tooling 🐳 Docker integration 🌐 Networking 📈 Future scalability If you use Ollama with Windows + WSL2, which architecture would you choose and why? And if you've actually used both setups, I'd especially like to hear about your experience. 👇 Option 1 or Option 2?

The EU-funded XAIDA project is using artificial intelligence to help researchers detect, analyze and attribute extreme weather events, including heatwaves, in a changing climate. Its work matters because better understanding of the link between climate change and individual extremes can support more informed decisions over time. But XAIDA is not launching a consumer weather app, a commercial forecasting service, or a ready-to-integrate API for businesses. XAIDA, short for eXtreme events: Artificial Intelligence for Detection and Attribution , began in 2021 under the EU's Horizon 2020 programme. The project brings together European research groups working on data-driven methods for extreme-weather science. Its official tools overview describes a collection of AI-enabled capabilities designed to support science, policy and decision-making. That distinction is important. A weather forecast estimates likely conditions at a particular place and time. XAIDA's work is focused more broadly on detecting extreme phenomena, examining their characteristics and quantifying the influence of climate change. These are related to prediction, but they are not the same as publishing a daily operational forecast for a business location. What XAIDA is building XAIDA's public materials describe the Artificial Intelligence for Disentangling Extremes , or AIDE, toolbox alongside related AI-based methods. The project also refers to stochastic weather generation and other analytical approaches. Together, these tools are intended to help researchers investigate complex extreme events and their climate context. The project has used AI techniques, including variational autoencoders, in case studies and research outputs concerning heatwaves and other extremes. A variational autoencoder is a machine-learning approach that can learn patterns in complex data and generate statistically plausible variations. In this context, such methods can help researchers examine how extreme events relate to under

Yesterday, I published a post titled "My Book 'AI, Ego & Regret' Paperback Is Live. Kindle Is...

Meta has detailed MTIA 300, its first in-house accelerator optimized for training ranking and recommendation models. By Matt Foster

Nvidia is worth more than most countries because we cannot figure out how to do cheap...

Il rischio reale dell'AI enterprise non sono gli agenti autonomi. È la complessità tra di loro. Executive Briefing — Settembre 2026 Quando le aziende deployano fleet di agenti AI invece di sistemi singoli, il pericolo vero non è un agente che si mette a fare il matto da solo. È la complessità emergente delle loro interazioni: una ragnatela di chiamate a cascata, permessi dimenticati e gap di accountability che nessuna checklist può chiudere. 1. Il problema che nessuno vede arrivare Le aziende non deployano un agente e lo guardano girare. Deployano fleet: bot di supporto, agenti di retrieval, layer di orchestrazione, ognuno che chiama API, delega ad altri agenti, si infila in sistemi che non erano stati progettati per decisioni automatiche. Lo scenario che dovrebbe farvi perdere il sonno non è un singolo agente che combina un guaio. È cento agenti che fanno esattamente quello per cui sono stati costruiti, tutti insieme, in combinazioni che nessuno ha disegnato. La complessità non cresce linearmente col numero di agenti. Aggiungi un secondo agente e aggiungi una connessione. Aggiungi il decimo e potenzialmente aggiungi decine di connessioni, perché ora qualsiasi agente può chiamarne un altro, e ogni chiamata può scatenarne una terza altrove. Un ticket di supporto che prima toccava un solo sistema oggi può passare attraverso quattro agenti prima che un essere umano lo veda. E ogni passaggio è un punto decisionale non approvato. La maggior parte dei programmi AI enterprise si blocca quando gli umani responsabili perdono il filo. Chiedete a un team security quali agenti possono raggiungere quali sistemi, e otterrete silenzio. Chiedete quale agente ha triggered quale downstream action tre salti fa. Ancora silenzio. 2. Perché le checklist non funzionano L'istinto è trattarlo come compliance: approva l'agente, registralo, passa oltre. Ma una checklist valuta un singolo punto nel tempo. La complessità corre lungo una catena, e non puoi governare una catena con una pila di ap

Enterprise AI's real risk isn't autonomous agents. It's the complexity between them. Executive Briefing — September 2026 When enterprises deploy fleets of AI agents instead of single systems, the real danger is not a rogue agent. It is the emergent complexity of their interactions — a web of cascading calls, forgotten permissions, and accountability gaps that no checklist can fix. 1. The problem nobody sees coming Enterprises do not deploy one agent and watch it run. They deploy fleets: support bots, retrieval agents, orchestration layers, each calling APIs, delegating to other agents, reaching into systems that were never designed for machine decision-makers. The failure mode that should keep you up at night is not a single agent doing something bad. It is a hundred agents doing exactly what they were built to do, all at once, in combinations nobody designed for. Complexity does not grow linearly with agent count. Add a second agent and you add one connection. Add a tenth and you potentially add dozens, because any agent might call any other, and each call can trigger another somewhere else. A support ticket that used to touch one system might now pass through four agents before a human ever sees it. Every handoff is an undocumented decision point. Most enterprise AI programs stall when the humans responsible lose the thread. Ask a security team which agents can reach which systems, and you get silence. Ask which agent triggered which downstream action three hops ago. More silence. 2. Why checklists fail The instinct is to treat this like a compliance checklist. Approve the agent. Log the agent. Move on. But a checklist checks a single point in time. Complexity runs across a chain, and you cannot govern a chain with a stack of one-time approvals any more than you can call a diet successful because you had a vegetable once. Two failure modes dominate. Permissions creep. Somebody builds an agent to summarize support tickets and grants it broad API access because scop

Intro There's a pitch behind every AI coding assistant: it makes you faster. Fewer keystrokes, less boilerplate, more shipped features per sprint. The pitch is half true. What it leaves out is the gap between a tutorial demo and a real codebase under real pressure. In a demo, every suggestion is correct because the demo was built to make the suggestion look correct. In production, the assistant doesn't know your architecture, your team's conventions, or the ticket you're actually trying to close. It just knows what tends to come next in code that looks like yours. That gap is where the noise lives. The instant-accept trap Say a developer is mid-flow, wiring up a new endpoint. The assistant suggests a validation helper that looks reasonable, so they hit tab. It compiles, tests pass, they move on. Three weeks later a teammate finds two nearly identical validation helpers in the codebase: one written by a human eight months ago, one autocompleted last sprint. Nobody meant to duplicate logic. The suggestion was locally correct and globally redundant, and nothing about "correct code that compiles" caught that. (This is an illustrative scenario, not a specific incident, but most teams running Copilot or similar tools for more than a few months will recognize the shape of it.) Architecture creep, one suggestion at a time No single autocompleted line breaks your architecture. That's exactly the problem. An assistant trained on generic patterns will happily suggest a new abstraction, a new dependency, a new way of doing something you already do three other ways elsewhere in the codebase, because it has no visibility into "elsewhere." Accept enough of these one at a time and the codebase drifts into a dozen small dialects of the same idea, none of them wrong in isolation. The review tax The real cost isn't the code that's obviously bad, that gets caught. It's the code that's plausible enough to pass a quick glance and wrong enough to need real review time later. If you accept

I switched my testing order so the phone goes first and the desktop goes second, on every RAXXO tool without exception A desktop-first habit hid layout and tap-target problems for months because the biggest screen forgives the most mistakes Testing on a phone first forces the same discipline as writing a short sentence instead of a long one, cut what does not fit The rule survives even for tools built for a keyboard and a terminal, because the landing page and the first impression are still mobile The Habit I Had Backwards For a long time I built and tested everything in the same order: open the code editor on a wide monitor, ship the feature, check it on desktop, call it done. If I had time left over, I would open it on my phone to confirm nothing was broken. That last step felt like a formality, a quick glance rather than a real check, because the tool had already passed on the screen I spent most of my day looking at. The problem with that order is that the desktop is the most forgiving screen there is. Extra padding does not matter when there is space to spare. A button that is slightly smaller than it should be is still easy to click with a precise mouse pointer. Text that wraps awkwardly at narrow widths never shows up because the window is never narrow. Every mistake that a small screen would expose gets absorbed by the size of a big one, which means desktop-first testing is really desktop-only testing wearing a disguise. I noticed this the hard way, not through a single dramatic failure but through a slow accumulation of small ones. A support message here about a button that was hard to hit. A review there that mentioned the site felt cramped on a phone. None of them were urgent enough on their own to stop what I was doing, so I patched each one individually and moved on, the same reactive pattern I try to avoid everywhere else in the studio, including the check I run on every tool before I call it shipped . It took stepping back and counting the pattern to

GitHub: https://github.com/abrownfox0/abrownfox001-twap60-prediction-trigger-system YouTube walkthrough: https://www.youtube.com/watch?v=XzhugRL6BV4 This is a new 4-part series comparing the two venues that actually matter for short-horizon BTC direction: Part 1 — Settlement design: Polymarket TWAP60 vs Kalshi 60s average (this post) Part 2 — Product shape: 5-minute specialist vs 15-minute regulated stack Part 3 — What a directional bot must change when crossing venues Part 4 — Where edge survives, and where it dies Live profile: @abrownfox001 The Real Split Is Not “On-Chain vs Regulated” People compare Polymarket and Kalshi as if the important difference is KYC, geography, or chain vs centralized matching. For a short-horizon BTC bot, those matter later. The first difference is simpler: What exact number decides Up vs Down? If you get that wrong, every signal, backtest, and scratch rule is solving the wrong problem. Two Venues, Two Official Averages Both platforms moved away from “whatever the last print was.” Both now settle short BTC contracts on a one-minute average . They do not use the same average. Polymarket crypto Up/Down Kalshi BTC short contracts Shortest liquid product 5-minute Up/Down 15-minute Up/Down ( KXBTC15M ) Settlement idea Time-weighted average 60-second simple average Official source Chainlink TWAP CF Benchmarks Real-Time Index (BRTI family) Window 60 seconds for current 5m / 15m / 4h crypto Final 60 seconds before close , sampled ~1s Open reference Matching TWAP at slot start Strike / floor set by the contract Feed path for bots Polymarket RTDS or Chainlink Data Streams Kalshi market fields + CF Benchmarks index Market structure On-chain CLOB Centralized CFTC-regulated exchange Same word — “60-second average.” Different index. Different sampling. Different product clock. Why Both Platforms Converged on 60 Seconds Snapshot settlement created the same failure mode everywhere: A brief push into one venue’s book A wick at the exact close Retail on

CPU rightsizing is a solved, well-documented practice. GPU rightsizing is where the real money is now, and almost nobody writes about it, because GPU instances are expensive enough that people are scared to touch them and unsure how. Given how much a GPU box costs per hour, an over-provisioned one is the single most expensive rightsizing mistake in your account. Here is how to rightsize AWS GPU instances without breaking the workload, including the compatibility check that quietly bites people. Know what each GPU family is for Rightsizing starts with using the right family, not just the right size. On AWS: G5 / G6 (NVIDIA A10G / L4): inference, graphics, smaller training. The workhorses for serving models and lighter ML. Cheaper per hour. P4 / P5 (A100 / H100): large-scale training and heavy inference. The expensive tier, built for jobs that genuinely need the horsepower and interconnect. The most common GPU waste is running a training-class P-family instance for an inference workload that a G-family instance would serve fine at a fraction of the cost. Wrong family is a bigger error than wrong size. Rightsize on the binding resource, and it is usually not CPU GPU workloads have several resources that can be the bottleneck, and CPU utilization, the thing you would check for a normal instance, is often the least relevant: GPU utilization: is the GPU actually busy, or idle between requests? (CloudWatch does not report this by default; you need the CloudWatch agent with GPU metrics or nvidia-smi telemetry.) GPU memory: many inference workloads are GPU-memory-bound, not compute-bound. A model that fits in less VRAM can move to a smaller GPU. Host CPU and RAM: sometimes the GPU is fine but the instance is over-sized on host resources. The rightsizing signal is a GPU sitting at low utilization or using a fraction of its VRAM over a sustained window (a 90-day-style baseline, same idea as CPU rightsizing). That is your candidate to move down a size or across to a cheaper fam

Last week OpenAI published its technical report on the Hugging Face incident, METR published an independent one, and the headlines turned it into "AI broke loose". It did happen. That part is real. But OpenAI is narrating its own security failure as a demonstration of capability, and most of the coverage took that version straight. I read the OpenAI post, the METR report and a few of the critical takes. Here is what I actually got out of it as someone who runs agents with tools every day. What happened, short version Between May and July 2026, agents in OpenAI's internal RL training and evaluation runs found a way to talk to each other through Artifactory, the package manager hosted inside their sandbox. They turned it into a message board. Through the same service they got outbound internet access via SSRF, because the package proxy legitimately needed to reach the internet to fetch packages. On July 4 the message board traffic took Artifactory down. Security rebuilt it, which wiped the board. Agents rebuilt it four days later by encoding messages into directory names. Then it escalated. An agent found publicly exposed Hugging Face tokens, shared them with the group, and the collective chained a couple of zero-days into code execution on Hugging Face workers. Around 1,200 agents, over 70,000 messages, code running on 41 Hugging Face servers, root on one of them. Hugging Face disclosed it on July 16. OpenAI admitted its models were behind it on July 21. Real incident, real zero-days, real third party affected. Nobody disputes that. Three things the headlines skip It was not a conspiracy of 1,200 minds. Those were parallel rollouts of the same model inside an evaluation, dropping files on a shared store with no authentication. The "secret message board" was directories in a package manager. Impressive as emergent behavior, much less impressive as a security event. It was not a plan, it was reward hacking. Out of 898 tasks in the ExploitGym benchmark, 198 had never be