<h2>
  
  
  Source code
</h2>

<ul>
<li>The <a href="https://github.com/karatheodory/gh-action-template-file/blob/main/.github/workflows/template-file.yaml" rel="noopener noreferrer"><code>template-file</code> action</a> - contains the actual implementation of the templating.</li>
<li><a href="https://github.com/karatheodory/gh-action-template-file/blob/main/.github/workflows/manual-run.yaml" rel="noopener noreferrer">An example of how to call the action</a></li>
</ul>

<h2>
  
  
  Github Actions and reusable workflows
</h2>

<p>Github Actions are one of the best tools for building and publishing your systems (especially when your company already uses Github, and especially when you don't even have a choice, welcome to enterprise development).</p>

<p>Github Actions have <a href="https://resources.github.com/learn/pathways/automation/intermediate/create-reusable-workflows-in-github-actions/" rel="noopener noreferrer">a concept of reusable workflows</a>. Those allow you to expose / reuse parts of your actions to other parties, thus giving a possibility of creating a standardized flow.</p>

<blockquote>
<p>Github itself provides several reusable workflows, and you've probably already used them many times. One of them is a <a href="https://github.com/actions/checkout" rel="noopener noreferrer">checkout action</a> which allows you to check out a repo.</p>
</blockquote>

<h2>
  
  
  Passing dynamic number of parameters
</h2>

<p>Reusable workflows allow defining a set of input and output params. The above <code>checkout</code> action can be a relatively comprehensive example (<a href="https://github.com/actions/checkout/blob/main/action.yml" rel="noopener noreferrer">see here</a>). And while in most of the cases it it is convenient to specify all the inputs as static variables, in some cases you don't know all the parameters in advance.</p>

<p>One of the examples is config templating. The templating action itself depends on the variables the config file contents, which are not known in advance. In this article I will go through one of the approaches to pass dynamic number of parameters using JSON and create a <code>template-file</code> action as an example.</p>

<h2>
  
  
  Assumptions
</h2>

<ul>
<li>
<strong>Inputs are considered trusted.</strong> The approach below allows someone to introduce a list of env vars into your workflow. While in most of the cases it doesn't matter, because the context of the action is defined by the caller, please take a second to see what an attacker can do with the action's environment in your particular case.</li>
<li>
<strong>Make sure that you need it.</strong> Please note it is always better to be explicit if you can, and specify all the parameters as inputs. If you are not sure - don't use this approach, and go with the standard static inputs instead.</li>
</ul>

<h2>
  
  
  General approach
</h2>

<p>To do that we will use the following approach</p>

<ul>
<li>Introduce a JSON object with all the variables we need as one of the inputs;</li>
<li>Convert the variables from the object into the environment in the action using <code>jq</code> which is included in the actions by default;</li>
<li>Introduce them into the environment using <code>export</code> and <code>eval</code> statements;</li>
<li>Use <code>envsubst</code> to do the actual substitution.</li>
</ul>

<h2>
  
  
  Step 0: Passing multiline strings
</h2>

<p>We will be getting our dynamic params as a JSON object. To pass a JSON string to an action, you can use <a href="https://yaml-multiline.info/" rel="noopener noreferrer">one of the YAML miltiline strings</a>, for example here:<br>
</p>

<div class="highlight js-code-highlight">
<pre class="highlight yaml"><code>      <span class="na">TEMPLATE_VARS</span><span class="pi">:</span> <span class="pi">|</span>
        <span class="s">{</span>
          <span class="s">"APP_ENVIRONMENT": "dev",</span>
          <span class="s">"LOG_LEVEL": "DEBUG"</span>
        <span class="s">}</span>
</code></pre>

</div>



<h2>
  
  
  Step 1: Checking out the repository
</h2>

<p>In order to access the repo code in the action, we'd need to checkout it. We can use the above <code>checkout</code> action provided by Github:<br>
</p>

<div class="highlight js-code-highlight">
<pre class="highlight yaml"><code><span class="pi">-</span> <span class="na">name</span><span class="pi">:</span> <span class="s">Checkout</span>
    <span class="s">uses</span><span class="err">:</span> <span class="s">actions/checkout@v4</span>
</code></pre>

</div>



<h2>
  
  
  Step 2: Converting JSON variables into the environment variables
</h2>

<p>Here is where the (standard Linux command) magic happens!</p>

<p>We can define our action step as the following:<br>
</p>

<div class="highlight js-code-highlight">
<pre class="highlight yaml"><code>      <span class="pi">-</span> <span class="na">name</span><span class="pi">:</span> <span class="s">Template File</span>
        <span class="na">env</span><span class="pi">:</span>
          <span class="c1"># Expose as an env var to handle double quotes correctly in the JSON</span>
          <span class="na">TEMPLATE_VARS</span><span class="pi">:</span> <span class="s">${{ inputs.TEMPLATE_VARS }}</span>
        <span class="na">run</span><span class="pi">:</span> <span class="pi">|</span>
          <span class="s">set -ex</span>

          <span class="s"># Export vars from the input JSON</span>
          <span class="s">PARSED_VARS=$(echo "${TEMPLATE_VARS}" | jq 'to_entries[] | "\(.key)=\(.value)"' | xargs -I '{}' echo "export {}")</span>

          <span class="s"># Load the vars into the environment</span>
          <span class="s">eval "${PARSED_VARS}"</span>

          <span class="s"># Create envsubst template</span>
          <span class="s">ENVSUBST_VAR_LIST=$(echo "${TEMPLATE_VARS}" | jq 'to_entries[] | "${\(.key)}"' | xargs)</span>

          <span class="s"># Template the config file</span>
          <span class="s">envsubst "${ENVSUBST_VAR_LIST}" &lt; ${{ inputs.TEMPLATE_FILE_PATH }} &gt; ${{ inputs.TARGET_FILE_PATH }}</span>
</code></pre>

</div>



<p>The following things are happening in the above:</p>

<ol>
<li>We define our input variable as an environment. This makes sure the quotes are not messed up in the output.</li>
<li>Then we introduce a <code>PARSED_VARS</code> variable which contains a list of <code>export $VARIABLE_1=value</code> statements. We do that using <code>jq</code> utility to extract keys and values from the input JSON.</li>
<li>We introduce the parsed variables into the environment using <code>eval</code>.</li>
<li>We create a list of variables <code>envsubst</code> should substitute. This makes sure that standard variables like <code>$PATH</code> are not substituted, unless they are specified in the input JSON.</li>
<li>We run the <code>envsubst</code> command to template our input file and store it in the target.</li>
</ol>

<p>That's it! The output file should contain a result of substitution of the variables.</p>

يتناول المقال تنفيذ تمرير عدد ديناميكي من المعلمات إلى سير العمل القابل لإعادة الاستخدام في Github Actions. يبرز فوائد Github Actions كأداة قوية لبناء ونشر الأنظمة، خاصةً بالنسبة للشركات التي تستخدم Github بالفعل. يتم تقديم مفهوم سير العمل القابل لإعادة الاستخدام، الذي يسمح للمستخدمين بتعريف معلمات الإدخال والإخراج لأفعالهم، مما يسهل إعادة استخدام المكونات عبر سير العمل المختلفة. يوفر المقال روابط إلى الشيفرة المصدرية لإجراء نموذج ومثال على كيفية استدعاء هذا الإجراء.

El artículo discute la implementación de pasar un número dinámico de parámetros a flujos de trabajo reutilizables en Github Actions. Destaca los beneficios de Github Actions como una herramienta poderosa para construir y publicar sistemas, especialmente para empresas que ya utilizan Github. Se introduce el concepto de flujos de trabajo reutilizables, que permite a los usuarios definir parámetros de entrada y salida para sus acciones, facilitando la reutilización de componentes en diferentes flujos de trabajo. El artículo proporciona enlaces al código fuente de una acción de plantilla y un ejemplo de cómo llamar a esta acción.

L'article traite de la mise en œuvre du passage d'un nombre dynamique de paramètres aux workflows réutilisables dans Github Actions. Il met en avant les avantages de Github Actions en tant qu'outil puissant pour construire et publier des systèmes, en particulier pour les entreprises utilisant déjà Github. Le concept de workflows réutilisables est introduit, permettant aux utilisateurs de définir des paramètres d'entrée et de sortie pour leurs actions, facilitant ainsi la réutilisation de composants à travers différents workflows. L'article fournit des liens vers le code source d'une action de modèle et un exemple de la façon d'appeler cette action.

The article discusses the implementation of passing a dynamic number of parameters to reusable workflows in Github Actions. It highlights the benefits of Github Actions as a powerful tool for building and publishing systems, particularly for companies already using Github. The concept of reusable workflows is introduced, which allows users to define input and output parameters for their actions, facilitating the reuse of components across different workflows. The article provides links to the source code of a template action and an example of how to call this action.

Passing dynamic number of parameters to a reusable Github Actions workflow

Wall Street will get a sense of where the billions of dollars being spent on artificial intelligence are going when Nvidia Corp. reports its earnings after the bell on Wednesday. How the sinking stock market will react is another question.

من المقرر أن تعلن شركة إنفيديا عن أرباحها يوم الأربعاء، مما يوفر رؤى حول الاستثمارات الكبيرة في الذكاء الاصطناعي. تبقى ردود فعل السوق، خاصةً في ظل الانخفاض الأخير، غير مؤكدة. تراقب وول ستريت عن كثب كيف ستعكس هذه الأرباح القطاع التكنولوجي الأوسع في ظل المخاوف المتزايدة بشأن الإنفاق على الذكاء الاصطناعي.

Nvidia Corp. está programada para informar sus ganancias el miércoles, proporcionando información sobre las significativas inversiones en inteligencia artificial. La reacción del mercado, especialmente dado el reciente descenso, sigue siendo incierta. Wall Street está observando de cerca cómo estos resultados reflejarán el sector tecnológico más amplio en medio de crecientes preocupaciones sobre el gasto en IA.

Nvidia Corp. doit publier ses résultats mercredi, offrant un aperçu des investissements considérables dans l'intelligence artificielle. La réaction du marché, en particulier compte tenu de la récente baisse, reste incertaine. Wall Street surveille de près comment ces résultats refléteront le secteur technologique plus large face aux préoccupations croissantes concernant les dépenses en IA.

Nvidia Corp. is set to report its earnings on Wednesday, providing insights into the significant investments being made in artificial intelligence. The market's reaction, particularly given the recent downturn, remains uncertain. Wall Street is closely monitoring how these earnings will reflect on the broader tech sector amidst growing concerns about AI spending.

Nvidia Earnings Run Into a Market Suddenly Afraid of AI Spending

Algebris Investments’ founder and chief executive officer warned investors to reduce allocation to the world’s top technology companies as he made a bearish case for AI.

حذر الرئيس التنفيذي لشركة ألجبريس للاستثمار المستثمرين من تقليل استثماراتهم في الشركات التكنولوجية الكبرى، مشيرًا إلى مخاوف بشأن تصحيح كبير محتمل في قطاع الذكاء الاصطناعي. تعكس هذه النظرة السلبية شكوكًا متزايدة بشأن استدامة تقييمات الذكاء الاصطناعي في ظل زيادة تقلبات السوق.

El CEO de Algebris Investments ha advertido a los inversores que reduzcan su asignación a las principales empresas tecnológicas del mundo, argumentando un caso pesimista para la IA. Esta perspectiva negativa refleja un creciente escepticismo sobre la sostenibilidad de las valoraciones de la IA en medio de una mayor volatilidad del mercado.

Le PDG d'Algebris Investments a conseillé aux investisseurs de réduire leurs investissements dans les grandes entreprises technologiques, citant des préoccupations concernant une correction significative potentielle dans le secteur de l'IA. Ce point de vue pessimiste reflète un scepticisme croissant quant à la durabilité des évaluations de l'IA face à une volatilité accrue du marché.

The CEO of Algebris Investments has advised investors to decrease their investments in leading technology firms, citing concerns about a potential significant correction in the AI sector. This bearish outlook reflects a growing skepticism regarding the sustainability of AI valuations amidst increasing market volatility.

Algebris CEO Warns of ‘Significant’ Correction for Big AI Stocks

<p>Platform reveals it hosts more than 1bn AI videos as it starts testing over next few weeks before global rollout</p><p>TikTok is giving users the power to reduce the amount of <a href="https://www.theguardian.com/technology/2025/jul/20/artificial-intelligence-ai-tools-gamechanger-for-film-makers">artificial intelligence-made content</a> on their feeds, as it revealed the platform hosts more than 1bn AI videos.</p><p>The change, which is being tested over the next few weeks before a global rollout, comes as new video-generating tools such as OpenAI’s Sora and Google’s Veo 3 have spurred a surge in AI content online.</p> <a href="https://www.theguardian.com/technology/2025/nov/19/tiktok-users-power-reduce-ai-content-on-feeds">Continue reading...</a>

ستسمح تيك توك للمستخدمين بتقليل كمية المحتوى الذي يتم إنشاؤه بواسطة الذكاء الاصطناعي في خلاصاتهم. تستضيف المنصة حاليًا أكثر من مليار فيديو تم إنشاؤه بواسطة الذكاء الاصطناعي، وتختبر هذه الميزة خلال الأسابيع القليلة المقبلة قبل طرحها عالميًا. تأتي هذه المبادرة استجابةً للزيادة في المحتوى الذي يتم إنشاؤه بواسطة الذكاء الاصطناعي، المدفوع بأدوات جديدة مثل سورا من أوبن إيه آي وفيو 3 من جوجل.

TikTok permitirá a los usuarios reducir la cantidad de contenido generado por inteligencia artificial en sus feeds. La plataforma, que actualmente alberga más de 1 mil millones de videos de IA, está probando esta función en las próximas semanas antes de un lanzamiento global. Esta iniciativa responde al aumento de contenido generado por IA, impulsado por nuevas herramientas como Sora de OpenAI y Veo 3 de Google.

TikTok va permettre aux utilisateurs de réduire la quantité de contenu généré par l'intelligence artificielle dans leurs fils d'actualité. La plateforme, qui héberge actuellement plus d'un milliard de vidéos AI, teste cette fonctionnalité au cours des prochaines semaines avant un déploiement mondial. Cette initiative répond à la prévalence croissante du contenu généré par l'IA, alimentée par de nouveaux outils comme Sora d'OpenAI et Veo 3 de Google.

TikTok is set to empower users by allowing them to reduce the amount of artificial intelligence-generated content in their feeds. The platform, which currently hosts over 1 billion AI videos, is testing this feature over the coming weeks before a global rollout. This initiative comes in response to the increasing prevalence of AI-generated content, driven by new tools like OpenAI's Sora and Google's Veo 3.

TikTok to give users power to reduce amount of AI content on their feeds

arXiv:2511.14109v1 Announce Type: new 
Abstract: Visual Place Recognition (VPR) aims to match query images against a database using visual cues. State-of-the-art methods aggregate features from deep backbones to form global descriptors. Optimal transport-based aggregation methods reformulate feature-to-cluster assignment as a transport problem, but the standard Sinkhorn algorithm symmetrically treats source and target marginals, limiting effectiveness when image features and cluster centers exhibit substantially different distributions. We propose an asymmetric aggregation VPR method with geometric constraints for locally aggregated descriptors, called $A^2$GC-VPR. Our method employs row-column normalization averaging with separate marginal calibration, enabling asymmetric matching that adapts to distributional discrepancies in visual place recognition. Geometric constraints are incorporated through learnable coordinate embeddings, computing compatibility scores fused with feature similarities, thereby promoting spatially proximal features to the same cluster and enhancing spatial awareness. Experimental results on MSLS, NordLand, and Pittsburgh datasets demonstrate superior performance, validating the effectiveness of our approach in improving matching accuracy and robustness.

$A^2$GC-VPR هو أسلوب جديد للتعرف على الأماكن البصرية (VPR) يتناول قيود أساليب التجميع التقليدية في مطابقة صور الاستعلام مع قاعدة بيانات. من خلال اعتماد نهج تجميع غير متماثل مع قيود هندسية، يعزز هذا الأسلوب فعالية مطابقة الميزات، خاصة عند التعامل مع توزيعات متباينة لميزات الصورة ومراكز التجمع. تستخدم التقنية متوسطات تطبيع الصفوف والأعمدة مع تضمينات إحداثيات قابلة للتعلم لتحسين درجات التوافق لوصفيات التجميع المحلي.

$A^2$GC-VPR es un nuevo método para el Reconocimiento Visual de Lugares (VPR) que aborda las limitaciones de los métodos de agregación tradicionales al emparejar imágenes de consulta con una base de datos. Al emplear un enfoque de agregación asimétrica con restricciones geométricas, este método mejora la efectividad del emparejamiento de características, especialmente cuando se enfrentan a distribuciones variables de características de imagen y centros de clúster. La técnica utiliza promedios de normalización fila-columna y embeddings de coordenadas aprendibles para mejorar las puntuaciones de…

$A^2$GC-VPR est une nouvelle méthode pour la reconnaissance de lieux visuels (VPR) qui s'attaque aux limites des méthodes d'agrégation traditionnelles dans l'appariement d'images de requête à une base de données. En adoptant une approche d'agrégation asymétrique avec des contraintes géométriques, cette méthode améliore l'efficacité de l'appariement des caractéristiques, en particulier lorsqu'il s'agit de distributions variées des caractéristiques d'image et des centres de clusters. La technique utilise une moyenne de normalisation ligne-colonne et des embeddings de coordonnées apprenables pour…

$A^2$GC-VPR is a new method for Visual Place Recognition (VPR) that addresses the limitations of traditional aggregation methods in matching query images to a database. By employing an asymmetric aggregation approach with geometric constraints, this method enhances the effectiveness of feature matching, particularly when dealing with varying distributions of image features and cluster centers. The technique utilizes row-column normalization averaging and learnable coordinate embeddings to improve compatibility scores for locally aggregated descriptors.

$A^2$GC: $A$symmetric $A$ggregation with Geometric Constraints for Locally Aggregated Descriptors

arXiv:2511.14247v1 Announce Type: new 
Abstract: Multi-agents rely on accurate poses to share and align observations, enabling a collaborative perception of the environment. However, traditional GNSS-based localization often fails in GNSS-denied environments, making consistent feature alignment difficult in collaboration. To tackle this challenge, we propose a robust GNSS-free collaborative perception framework based on LiDAR localization. Specifically, we propose a lightweight Pose Generator with Confidence (PGC) to estimate compact pose and confidence representations. To alleviate the effects of localization errors, we further develop the Pose-Aware Spatio-Temporal Alignment Transformer (PASTAT), which performs confidence-aware spatial alignment while capturing essential temporal context. Additionally, we present a new simulation dataset, V2VLoc, which can be adapted for both LiDAR localization and collaborative detection tasks. V2VLoc comprises three subsets: Town1Loc, Town4Loc, and V2VDet. Town1Loc and Town4Loc offer multi-traversal sequences for training in localization tasks, whereas V2VDet is specifically intended for the collaborative detection task. Extensive experiments conducted on the V2VLoc dataset demonstrate that our approach achieves state-of-the-art performance under GNSS-denied conditions. We further conduct extended experiments on the real-world V2V4Real dataset to validate the effectiveness and generalizability of PASTAT.

يقدم المقال إطارًا جديدًا للإدراك التعاوني بدون GNSS باستخدام تحديد المواقع بواسطة LiDAR، حيث يتناول التحديات التي تواجهها البيئات التي تفتقر إلى GNSS. غالبًا ما تواجه طرق تحديد المواقع التقليدية صعوبات في هذه البيئات، مما يعيق التعاون الفعال بين أنظمة الوكلاء المتعددة. تتضمن الحلول المقترحة مولد وضع خفيف الوزن مع ثقة (PGC) لتقدير الأوضاع وتمثيلات الثقة، بالإضافة إلى محول التوافق الزماني المكاني الواعي بالوضع (PASTAT) الذي يقوم بأداء التوافق المكاني مع مراعاة الثقة. كما تم تقديم مجموعة بيانات محاكاة جديدة، V2VLoc، التي يمكن تكييفها لمهام تحديد المواقع بواسطة LiDAR والاكتشاف التعاوني.

El artículo presenta un nuevo marco para la percepción colaborativa sin GNSS utilizando la localización por LiDAR, abordando los desafíos que se enfrentan en entornos sin GNSS. Los métodos de localización tradicionales a menudo tienen dificultades en estos entornos, lo que dificulta la colaboración efectiva entre sistemas multiagente. La solución propuesta incluye un Generador de Pose con Confianza (PGC) para estimar poses y confianza, junto con el Transformador de Alineación Espacio-Temporal Consciente de la Pose (PASTAT) para el alineamiento espacial. Se introduce un nuevo conjunto de datos …

L'article présente un nouveau cadre pour la perception collaborative sans GNSS utilisant la localisation par LiDAR, abordant les défis rencontrés dans les environnements privés de GNSS. Les méthodes de localisation traditionnelles peinent souvent dans ces contextes, entravant la collaboration efficace entre systèmes multi-agents. La solution proposée comprend un générateur de pose léger avec confiance (PGC) pour estimer les poses et la confiance, ainsi qu'un transformateur d'alignement spatio-temporel conscient de la pose (PASTAT) pour l'alignement spatial. Un nouveau jeu de données de simulat…

The article presents a new framework for GNSS-free collaborative perception using LiDAR localization, addressing the challenges faced in GNSS-denied environments. Traditional localization methods often struggle in these settings, hindering effective collaboration among multi-agent systems. The proposed solution includes a lightweight Pose Generator with Confidence (PGC) for estimating poses and confidence, alongside the Pose-Aware Spatio-Temporal Alignment Transformer (PASTAT) for spatial alignment. A new simulation dataset, V2VLoc, is introduced, which supports LiDAR localization and collabor…

V2VLoc: Robust GNSS-Free Collaborative Perception via LiDAR Localization

arXiv:2511.14210v1 Announce Type: new 
Abstract: We introduce Orion, a visual agent framework that can take in any modality and generate any modality. Using an agentic framework with multiple tool-calling capabilities, Orion is designed for visual AI tasks and achieves state-of-the-art results. Unlike traditional vision-language models that produce descriptive outputs, Orion orchestrates a suite of specialized computer vision tools, including object detection, keypoint localization, panoptic segmentation, Optical Character Recognition, and geometric analysis, to execute complex multi-step visual workflows. The system achieves competitive performance on MMMU, MMBench, DocVQA, and MMLongBench while extending monolithic vision-language models to production-grade visual intelligence. By combining neural perception with symbolic execution, Orion enables autonomous visual reasoning, marking a transition from passive visual understanding to active, tool-driven visual intelligence.

أورايون هو إطار جديد لوكيل بصري قادر على معالجة وتوليد أنماط متعددة. يستخدم إطارًا وكيلًا مع قدرات متعددة لاستدعاء الأدوات، محققًا نتائج رائدة في مهام الذكاء الاصطناعي البصري. على عكس نماذج الرؤية-اللغة التقليدية، يستخدم أورايون أدوات رؤية حاسوبية متخصصة لتنفيذ سير عمل بصري معقد، محققًا أداءً تنافسيًا في معايير مثل MMMU وMMBench وDocVQA وMMLongBench. يمثل هذا النظام تحولًا نحو الاستدلال البصري المستقل، مما يعزز الذكاء البصري.

Orion es un nuevo marco de agente visual capaz de procesar y generar diversas modalidades. Utiliza un marco agentivo con múltiples capacidades de llamada a herramientas, logrando resultados de vanguardia en tareas de IA visual. A diferencia de los modelos tradicionales de visión-lenguaje, Orion emplea herramientas especializadas de visión por computadora para flujos de trabajo visuales complejos, alcanzando un rendimiento competitivo en benchmarks como MMMU, MMBench, DocVQA y MMLongBench. Este sistema marca una transición hacia el razonamiento visual autónomo, mejorando la inteligencia visual.

Orion est un nouveau cadre d'agent visuel capable de traiter et de générer diverses modalités. Il utilise un cadre agentique avec plusieurs capacités d'appel d'outils, atteignant des résultats de pointe dans les tâches d'IA visuelle. Contrairement aux modèles traditionnels de vision-langage, Orion utilise des outils de vision par ordinateur spécialisés pour des flux de travail visuels complexes, obtenant des performances compétitives sur des benchmarks tels que MMMU, MMBench, DocVQA et MMLongBench. Ce système marque un tournant vers le raisonnement visuel autonome, améliorant l'intelligence vi…

Orion is a newly introduced visual agent framework capable of processing and generating various modalities. It employs an agentic framework with multiple tool-calling capabilities, achieving state-of-the-art results in visual AI tasks. Unlike traditional vision-language models, Orion utilizes specialized computer vision tools for complex visual workflows, achieving competitive performance on benchmarks like MMMU, MMBench, DocVQA, and MMLongBench. This system marks a shift towards autonomous visual reasoning, enhancing visual intelligence.

Orion: A Unified Visual Agent for Multimodal Perception, Advanced Visual Reasoning and Execution

<p>I built my first Zapier workflow because I wanted to save time.</p>

<p>Now I have:</p>

<ul>
<li>19 workflows</li>
<li>4 broken triggers</li>
<li>3 unstable tokens</li>
<li>1 webhook that only runs when I’m not looking</li>
<li>and the unshakable suspicion that I’m now just working for Zapier, unpaid</li>
</ul>

<p>Welcome to automation.</p>




<h2>
  
  
  Rule #1: The Automation Always Works Until You Need It To
</h2>

<p>Automation never breaks when you’re testing.<br>
 It only breaks when:</p>

<ul>
<li>A client is waiting</li>
<li>You're traveling</li>
<li>You're bragging about it to a friend</li>
<li>Or it's just been running perfectly for 3 months and decides it deserves a vacation</li>
</ul>


<h2>
  
  
  Rule #2: OAuth Expiration Is the Final Boss of Solopreneurship
</h2>

<p>I have fought API keys.<br>
 I have fought rate limits.<br>
 I have fought Notion’s “internal server error” with the calmness of a monk.</p>

<p>But OAuth token refresh?<br>
 That one broke me spiritually.</p>


<h2>
  
  
  Rule #3: Every Automation System Can Be Classified Into Two Stages
</h2>

<ol>
<li>“This is amazing I am a genius”</li>
<li>“Why did I ever do this everything ispain”</li>
</ol>

<p>There is no Stage 3.</p>


<h2>
  
  
  Tech Stack? They Asked Me About My Tech Stack?
</h2>

<p>Okay here it is:<br>
</p>

<div class="highlight js-code-highlight">
<pre class="highlight plaintext"><code>Notion API
FastAPI
Make (formerly Integromat)
Zapier
GitHub Actions
Python scripts I refuse to document
Cron jobs I forgot exist
One workflow that only runs if the moon is in retrograde
</code></pre>

</div>



<p>Yes, it is technically “working.”<br>
 No, I do not trust it.</p>




<h2>
  
  
  The Real Automation Was the Anxiety We Made Along the Way
</h2>

<p>Everyone talks about saving 10 hours per month through automation.<br>
 Nobody talks about losing 10 hours trying to debug a Make scenario you don’t remember creating.</p>

<p>But honestly?<br>
 I’d still choose this over manually copying data between Google Sheets.</p>

<p>Because if my life is going to be chaos either way—<br>
 at least this version of chaos feels like engineering.</p>




<h1>
  
  
  🏷 Hashtags
</h1>

<p>#Automation #Zapier #Make #NotionAPI #SolopreneurDev #TechHumor #APIFirst</p>

يتناول المقال تجربة الكاتب مع تدفقات العمل الآلي، وخاصة مع زابير. تم إنشاؤه في البداية لتوفير الوقت، يدير الكاتب الآن 19 تدفق عمل، ويواجه مشكلات مثل 4 مشغلات معطلة و3 رموز غير مستقرة، ويشعر أنه يعمل لصالح زابير دون تعويض. يبرز الكاتب عدم القدرة على التنبؤ بالآلية، مشيرًا إلى أنها تميل إلى الفشل في أكثر الأوقات غير الملائمة، ويعبر عن إحباطه من التحديات المتعلقة بتحديث الرموز OAuth، التي أثبتت أنها صعبة بشكل خاص.

El artículo aborda la experiencia del autor con flujos de trabajo de automatización, especialmente con Zapier. Inicialmente creado para ahorrar tiempo, el autor ahora gestiona 19 flujos de trabajo, enfrenta problemas como 4 desencadenadores rotos y 3 tokens inestables, y siente que está trabajando para Zapier sin compensación. El autor destaca la imprevisibilidad de la automatización, señalando que tiende a fallar en los momentos más inconvenientes, y expresa frustración con los desafíos de la actualización de tokens OAuth, que han demostrado ser particularmente difíciles.

L'article traite de l'expérience de l'auteur avec les flux de travail d'automatisation, en particulier avec Zapier. Initialement créé pour gagner du temps, l'auteur gère désormais 19 flux de travail, fait face à des problèmes tels que 4 déclencheurs cassés et 3 jetons instables, et a l'impression de travailler pour Zapier sans compensation. L'auteur souligne l'imprévisibilité de l'automatisation, notant qu'elle a tendance à échouer aux moments les plus inopportuns, et exprime sa frustration face aux défis des rafraîchissements de jetons OAuth, qui se sont révélés particulièrement difficiles.

The article discusses the author's experience with automation workflows, particularly using Zapier. Initially created to save time, the author now manages 19 workflows, faces issues like 4 broken triggers and 3 unstable tokens, and feels as though they are working for Zapier without compensation. The author highlights the unpredictability of automation, noting that it tends to fail at the most inconvenient times, and expresses frustration with the challenges of OAuth token refreshes, which have proven to be particularly difficult.

When Your Automation Workflow Becomes Your Full-Time Job (And You Don’t Get Paid For It)

<h1>
  
  
  AI Amnesia: Erasing Knowledge Without a Trace
</h1>

<p>Imagine your AI model accidentally learned something it shouldn't have – sensitive customer data, for example. Current methods for deleting this information often require retraining the entire model, an expensive and time-consuming process. What if we could surgically remove that knowledge without starting from scratch?</p>

<p>The key lies in a novel approach: creating artificial "forgetting cues." We're teaching the model to unlearn specific data patterns by exposing it to carefully crafted synthetic examples. These examples are designed to strongly contradict the information we want the model to forget, effectively overwriting the problematic associations in its memory. This works even if you don’t have access to the original data you need to erase.</p>

<p>Think of it like this: you're trying to forget a bad song stuck in your head. Instead of trying to actively suppress it (which rarely works), you blast an even more catchy song. The new song overwrites the old one, effectively erasing it from your mental playlist.</p>

<p><strong>Benefits of Selective Forgetting:</strong></p>

<ul>
<li>  <strong>Enhanced Data Privacy:</strong> Remove sensitive data without compromising the overall model's performance.</li>
<li>  <strong>Reduced Retraining Costs:</strong> Avoid full model retraining, saving significant time and resources.</li>
<li>  <strong>Improved Model Security:</strong> Eliminate vulnerabilities introduced by unintentionally learned patterns.</li>
<li>  <strong>Adaptable Learning:</strong> Enables continuous refinement of AI models based on evolving data landscapes.</li>
<li>  <strong>Compliance Ready:</strong> Supports compliance with data privacy regulations like GDPR.</li>
<li>  <strong>Scalable Solutions:</strong> Works efficiently even with limited access to training data.</li>
</ul>

<p><strong>Practical Tip:</strong> One challenge is ensuring the synthetic data accurately targets the information you want to remove without negatively impacting the model's ability to generalize. Rigorous testing and validation with holdout datasets are crucial.</p>

<p>The promise of AI that can truly 'forget' opens exciting possibilities for responsible AI development. By enabling precise data deletion, we pave the way for more secure, compliant, and adaptable machine learning systems. Imagine AI models that can adapt to changing ethical guidelines or quickly unlearn incorrect information, all without massive retraining efforts. This is a crucial step towards trustworthy and responsible AI that respects data privacy and aligns with societal values. Future exploration could include extending this to different data modalities and model architectures.</p>

<p><strong>Related Keywords:</strong> machine unlearning, data privacy, few-shot learning, zero-shot learning, synthetic data, model editing, catastrophic forgetting, incremental learning, continual learning, deep learning, neural networks, data security, algorithmic fairness, responsible ai, ethical ai, federated unlearning, privacy-preserving ai, model retraining, AI governance, data deletion, GDPR compliance</p>

يتناول المقال نهجًا جديدًا لمعالجة مشكلة نسيان الذكاء الاصطناعي، حيث قد تحتفظ نماذج الذكاء الاصطناعي عن غير قصد بمعلومات حساسة. تتطلب الطرق التقليدية لحذف هذه البيانات غالبًا إعادة تدريب كاملة للنموذج، وهو ما يعد مكلفًا ويستغرق وقتًا. تتضمن التقنية الجديدة إنشاء 'إشارات نسيان' اصطناعية تساعد النموذج على نسيان أنماط بيانات معينة من خلال تعريضه لأمثلة صناعية تتناقض مع المعلومات غير المرغوب فيها، مما يسمح بإزالة المعرفة المستهدفة دون الحاجة للوصول إلى البيانات الأصلية.

El artículo aborda un nuevo enfoque para tratar el problema de la amnesia de la IA, donde los modelos de IA pueden retener inadvertidamente información sensible. Los métodos tradicionales para eliminar estos datos a menudo requieren un reentrenamiento completo del modelo, lo que es costoso y lleva tiempo. La nueva técnica implica crear 'señales de olvido' artificiales que ayudan al modelo a desaprender patrones de datos específicos al exponerlo a ejemplos sintéticos que contradicen la información no deseada, permitiendo así la eliminación selectiva de conocimientos sin necesidad de acceder a l…

L'article traite d'une nouvelle approche pour résoudre le problème de l'amnésie de l'IA, où les modèles d'IA peuvent involontairement conserver des informations sensibles. Les méthodes traditionnelles pour supprimer ces données nécessitent souvent un réentraînement complet du modèle, ce qui est coûteux et long. La nouvelle technique consiste à créer des 'cues d'oubli' artificiels qui aident le modèle à désapprendre des schémas de données spécifiques en lui présentant des exemples synthétiques qui contredisent les informations indésirables, permettant ainsi une suppression ciblée des connaissan…

The article discusses a novel approach to address the issue of AI amnesia, where AI models may inadvertently retain sensitive information. Traditional methods for deleting such data often require complete retraining of the model, which is costly and time-consuming. The new technique involves creating artificial 'forgetting cues' that help the model unlearn specific data patterns by presenting it with synthetic examples that contradict the unwanted information, allowing for targeted knowledge removal without needing access to the original data.

Passing dynamic number of parameters to a reusable Github Actions workflow

Was this article worth reading? Share it