标签：#REI

Labor Day already? Say it ain’t so. You can still chase summer a while longer with these great deals on outdoor gear at the REI Labor Day Sale.

Surging demand for AI training data is driving rapid growth for the startup and its rivals.

An extortion gang known for targeting transportation companies and private equity firms has taken credit for a breach at Uber Freight.

In this episode of the podcast, members of the InfoQ editorial staff and friends of InfoQ will discuss current trends in the cloud and DevOps domains as part of our annual trends report. These reports provide InfoQ readers with a high-level overview of key topics to watch. This podcast offers a chance to hear our raw conversation and the stories shared by our expert practitioners. By Daniel Bryant, Matt Saunders, Shweta Vohra, Steef-Jan Wiggers, Mark Silvester

InfoQ editorial staff and friends of InfoQ are discussing the current trends in the domain of Cloud and DevOps as part of the process of creating our annual 2026 trends report. By Steef-Jan Wiggers, Matt Saunders, Shweta Vohra, Daniel Bryant, Mark Silvester

Johannes Schickling explores how he moved beyond the traditional three-tier web stack to a local-first approach. He shares his experience transitioning from Prisma to developing Overtone—a music curation app leveraging client-side event sourcing and SQLite. The discussion highlights the trade-offs between event sourcing and CRDTs. By Johannes Schickling

The neolab is betting that automating routine computer tasks will soon outpace coding as AI's biggest use case.

Airbus selected Scaleway as its sovereign cloud provider after a tender that scored protection against non-European extraterritorial legislation alongside technical capability. Airbus frames it as complementing multi-cloud, not exiting AWS. Practitioners note the pattern is spreading past hyperscalers to small US SaaS vendors, and that sovereignty claims still require verifiable controls. By Steef-Jan Wiggers

Recap. Part 1 framed the supervision problem. Part 2 architected the four-model system on AWS and counted the hardware. Part 3 put the gate on the page in fifteen lines of PyTorch. Now the question that decides whether any of it mattered: does the gate actually earn its keep - and what would it cost to know? This is the close. A verification design, the stability instrumentation that separates a real evaluation from a lucky one, and the FinOps reality that is the reason this whole series is a blueprint instead of a benchmark. SDAR makes two claims, not one It's tempting to reduce the paper to a single number. The reported gains over plain GRPO are real and worth stating - their numbers, not mine : roughly +9.4% on ALFWorld, +7.0% on Search-QA, +10.2% on WebShop accuracy. But the headline percentage is the less interesting claim. The one that matters is the second: SDAR avoids the training instability that naive GRPO+OPSD falls into . The whole point of the gate Part 3 was to keep the teacher's noisy rejections from destabilizing training. So a verification that only measures final task success has tested half the paper. You have to measure stability too - and most reproductions don't. That two-pronged claim dictates the experiment. The verification: three arms, not two You cannot prove SDAR's thesis with a before/after. You need three training runs, identical except for the supervision: Arm What it is What it proves A. GRPO Plain agentic RL, no teacher The baseline the +9.4% is measured against B. Naive GRPO+OPSD Teacher distillation, no gate The instability SDAR claims to fix C. SDAR Gated distillation Part 3 Beats A on score and beats B on stability Drop arm B and you can show SDAR beats GRPO, but you've quietly deleted the paper's actual contribution - there's no evidence the gate did anything a plain teacher wouldn't. Arm B is the expensive arm nobody wants to run and the one that makes the result credible. Metrics that actually test the claim Metric What it mea

A panel on data ownership challenged the definition of "ownership," arguing it must extend beyond simple account control to include structural independence, interoperability, and community governance. Speakers like Zenna Fiscella, Paul Frazee, Boris Mann, and Robin Berjon emphasised the need for shared standards, unbundled platforms, and better tools to support user sovereignty. By Olimpiu Pop

Founded in 2024, Prime Intellect’s goal is to give organizations capabilities to train their own agentic systems without relying on frontier AI labs.

Cycle recently introduced a separate EU-based control plane, allowing European customers to keep platform management data and telemetry within Europe. The new offering is designed to improve compliance, operational isolation, and responsiveness for European organizations. By Renato Losio

Insiders say Sam Altman is in active talks with the Trump administration.

Ahnii! If you let an AI agent modify your application, the agent needs more than a text editor. Raw str_replace on a PHP file passes a lot of tests and still breaks things an hour later in production, because the tool has no idea what the file actually represents. Bimaaji is the Waaseyaa package that gives agents a structured path from "I want to add a field to this entity" to a reviewable patch that a community's sovereignty rules have already vetted. This post walks through what shipped in waaseyaa/bimaaji and why each piece exists. Prerequisites: familiarity with Waaseyaa's package layout, PHP 8.4+, and the idea that an application has more state than the filesystem (routes, entities, introspection metadata). Why not just let the agent edit files The failure mode you want to avoid: an agent reads a prompt like "add a published_at field to the Post entity," does a reasonable-looking edit to Post.php , and leaves the rest of the app inconsistent. The migration is missing. The JSON:API resource doesn't expose the field. The admin panel still doesn't know it exists. The sovereignty profile that was supposed to block the change on a local-only deployment never got consulted. Each of those is a different subsystem. A good agent can write a correct edit to any one of them. What a filesystem-level tool cannot do is ensure the edit is coordinated across all of them and is allowed under the community's posture. Bimaaji separates that problem into three stages: introspect, propose, patch. The pipeline The package description (from packages/bimaaji/composer.json ) spells it out: application graph introspection and agent-safe mutation for Waaseyaa. The flow is: Introspection → ApplicationGraph → MutationRequest → Validator → PatchGenerator → PatchSet An agent reads the graph, submits a structured mutation request, a validator checks it against sovereignty rules, and the patch generator returns reviewable diffs. Nothing touches the filesystem until a human (or a higher-level w

The Samsara Tracking Label hides a BLE system that can offer real-time location in a disposable package.

The spy law known as Section 702, which authorizes the NSA and FBI's warrantless surveillance, will all but certainly expire on Friday for the first time.

https://www.youtube.com/watch?v=VD9zEKQEJxo 这视频深入拆解了人工智能强化学习之父、图灵奖得主理查德·萨顿（Richard Sutton） 在2026年5月共同发表的一篇仅有7页、零算法、零跑分的哲学立场论文。这篇论文提出了 “行动认知 AI”（Enactive Artificial Intelligence，简称 Enactive AI）的概念，并在科技界和资本圈引发了巨大震动（甚至让红杉、英伟达、谷歌联合下注了11亿美元成立新公司）。 视频从 核心概念、哲学脉络、理论内在矛盾、认知科学质疑 以及 产业界的三路对赌 五个维度，极其详细地复盘了视频的所有核心内容： 一、 什么是“行动认知 AI”（Enactive AI）？ 视频强调，全网很多地方都把 Enactive （行动认知/生成认知）和 Generative （生成式 AI，如 GPT、Sora）混淆了，但两者的底层逻辑恰恰相反 [ 00:50 ]： 生成式 AI（Generative AI）： 核心是 续写和预测 。通过已有画面或文本，被动地去预测下一帧、下一个词长什么样 [ 01:07 ]。 行动认知 AI（Enactive AI）： 核心是 在互动中现生成认知 。认知不是大脑被动接收信号并建立静态世界模型，而是“你动了手，世界才向你显现” [ 01:47 ]。 > 举例： 人去拿杯子，不是眼睛先拍下一张静态照片让大脑去死算距离、角度 [ 01:53 ]，而是手往前探的过程中，随着角度、光影的实时动态变化，杯子的形状和可抓取性才在动作里一点点“长出来” [ 01:59 ]。 感知和行动硬死在一起，无法拆分。 这套理论源自认知科学中的 自创生（Autopoiesis）与自主性（Autonomy） [ 02:21 ]。它认为智能体应该像生物一样自我维持、组织，由内在生存需求去塑造感知，而不是一个干等着外部指令输入输出的机器 [ 02:24 ]。 二、 萨顿为什么要发这篇哲学论文？ 萨顿并不是一时性起，这是他为了对抗当前“大模型路线”打出的最后一张哲学底牌： 2019年《苦涩的教训》： 主张人类手写规则干不过堆算力、让机器自己学的通用方法 [ 02:47 ]。 2024年《大世界假设》： 真实世界远比静态内部模型复杂，智能体必须在运行中实时学习 [ 02:59 ]。 2025年《经验时代》： 人类数据是有限的，AI 必须靠自己生成自己的经验长大的 [ 03:12 ]。 2025年9月： 直指整个 AI 行业走错路，大模型堆数据去超智是死路一条 [ 03:19 ]。 这篇论文补上了最后一把火： 之前的论证全是算力、数据和复杂度的“机械账” [ 03:25 ]。而这一次，他第一次把强化学习（RL） 和 认知科学（行动认知）接在了一起，从本体论层面证明： 大模型路走不通，认识世界这件事本身，就只能通过行动和互动的经验来发生 [ 03:39 ]。 为此，2026年初论文共作者创办了 Ineffable Intelligence 公司，号称要造出完全不需要人类数据、靠自己学习的 AI，直接拿到了红杉、英伟达、谷歌 11 亿美元的巨额融资（估值 51 亿美元） [ 03:55 ]。 三、 论文隐藏的两大致命致命逻辑“回旋镖” 视频话锋一转，指出萨顿借来的这套哲学地基里，埋着两根砸中他自己的“大柱子”： 柱子 1：砸中了萨顿的“奖励假设”（自相矛盾） [ 04:35 ] 强化学习的号称教条： 奖励假设（Reward Hypothesis），即所有目标、意图都可以写成“最大化外部给定的标量分数” [ 04:53 ]。David Silver 甚至喊出“奖励就够了” [ 05:13 ]。 行动认知哲学的教条： 自主性（Autonomy），即什么是好坏、成败，标准必须从智能体随时会散架的“物理组织和生存危机”中自发长出来，不能由外部权威操控 [ 05:27 ]。 裂缝： 标准强化学习的奖励函数（Reward Function）是人类设计者用代码硬塞进去的（他律） [ 05:55 ]；而生物判断好坏是为了顶住熵增、维持结构不崩（自主） [ 06:11 ]。论文里作者自己也承认：强化学习的评估标准依然由外部奖励定义 [ 06:38 ]。 内驱动机能救场吗？ 比如好奇心驱动或求知驱动。视频认为不能，因为诸如“优化预测误差”的总结优化目标，依然是人类在架构层死死规定好的，根本不是智能体出于生存忧关的自发需求。没有真正的生命威胁，就没真正的意义生成 [ 07:12 ]。 柱子 2：砸中了萨顿自己的《苦涩的教训》 [ 07:49 ] 萨顿当年痛骂：研究者总忍不住把人类以为的思考结构（比如语法树、手工特征检测器）硬塞进 AI 架构里，这长期必被碾压 [ 08:13

Martin Kleppmann discusses the urgent need for technological sovereignty in modern infrastructure. Exploring the shifting landscape of global tech dependencies, he shares how engineering leaders can leverage multi-cloud architecture, de facto API standardization, the AT Protocol, and local-first development paradigms to reclaim user agency and build highly resilient systems. By Martin Kleppmann

Em 2023, eu arrumei as malas e saí de Recife, minha terra natal, para morar no interior do Mato Grosso. O motivo? Cursar Ciência da Computação. Eu sabia que enfrentaria um choque cultural, o calor do Centro-Oeste e a saudade de casa. O que eu não imaginei é que o maior desafio seria a sensação constante de que eu era uma fraude no meio da minha própria sala de aula. Se você estuda ou trabalha com tecnologia, provavelmente já passou por isso. Você senta na primeira semana de aula e parece que metade da turma já programa desde os 12 anos, fala termos técnicos que parecem outra língua e discute sobre ferramentas que você nem sabia que existiam. Hoje, na reta final da graduação, focando meus estudos em Back-end e Dados, quebrando a cabeça com Go, Python e SQL, eu olho para trás e vejo o quanto essa cobrança silenciosa quase me paralisou. Se você está sentindo que entrou no curso errado ou que todo mundo corre a 100 km/h enquanto você ainda está engatinhando, pega um café e vem ler este papo reto sobre como lidar com a tal da Síndrome do Impostor. O "efeito vitrine" e a falsa ilusão de que somos os únicos perdidos Na faculdade de TI, a comparação é quase inevitável. A gente abre o LinkedIn e vê o colega conseguindo estágio internacional; abre o GitHub e vê códigos impecáveis; na sala de aula, sempre tem quem responda às perguntas do professor antes mesmo dele terminar de falar. O erro está em achar que o ritmo do outro deve ditar o seu. Na Computação, as pessoas vêm de bagagens completamente diferentes. Quem já sabia programar antes da faculdade pode ter facilidade em Lógica de Programação, mas talvez sofra tanto quanto você quando chegar a hora de aprender Álgebra Linear ou Teoria da Computação. Sentir-se perdida não significa incompetência; significa apenas que você está no processo de aprender algo complexo. E adivinha? Todo mundo ali está tentando descobrir como as coisas funcionam, mesmo quem finge que sabe tudo. O que me ajudou a virar a chave Eu não acordei um dia

About this series. I'm going to take a fresh paper - Self-Distilled Agentic Reinforcement Learning (SDAR, arXiv:2605.15155 ) - and architect it end to end on AWS: the system design, the actual gate code, the evaluation plan, and a brutally honest cost model. What I'm not going to do is wave a benchmark number around. Reproducing a paper like this costs thousands in GPU time, and I'd rather show you the machinery than a screenshot you can't audit. The design is the deliverable. This is Part 1. A small, infuriating problem Picture an LLM agent working a web-shopping task. It reads the goal, searches, clicks a category, filters, opens a product, compares, adds to cart - twelve steps in all. At the end, it bought the wrong thing. So you do what reinforcement learning tells you to do: you score the trajectory. Reward = 0. Bad agent. Now answer this: which of the twelve steps was actually wrong? Maybe step 3, the search query, was fine and step 9, a filter choice, doomed everything. Maybe steps 1–11 were brilliant and step 12 fat-fingered the wrong button. Your single scalar reward has no idea. It punishes all twelve equally, including the eight that were correct. That's the supervision problem in agentic RL, and it's the thing this whole series is about. Why "just use RL" isn't enough for agents RL has become the default way to post-train LLM agents. The catch is that the reward usually lands at the trajectory level - one number for the entire multi-step episode. For a single-turn task ("answer this question"), that's tolerable; the action and the outcome are close together. For a long-horizon agent - ten, twenty, fifty turns of searching, calling tools, and reacting to an environment - it's a disaster of credit assignment . The signal is too coarse to tell the model which decisions earned the reward and which torpedoed it. You can throw more episodes at it and let statistics sort the credit out eventually. But "eventually" on a 30-turn task burns a lot of expensive comp