Dixa

dixa

Have you ever had that conversation with a parent when you try to face-time, and you’re the one face-timing whilst they’re the one giving you a close up of their ear, muttering words of confusion and frustration? As annoying as getting a close up of your loved one’s ear can be, I think you would agree that it’s important to power through those bizarre movements and help talk them through what they’re doing wrong so they can actually benefit themselves from the magic of modern technology. I think this same loving rule of thumb should be applied to automation and the ability to code something for yourself. The Decision To some, the ability to code can seem like magic. If you’re the person who knows how to code, that point in your show and tell where someone can’t believe what you’ve been able to do, you have to make a decision... Do you accept the praise and astonishment? Cementing your name as a modern day Merlin, or do you help to explain how actually, it’s not that hard once you know where to look? Why Accepting Merlin Status Benefits No One By explaining how something can be done in such a way that you bring as many people along for the ride with you, you will improve things for everyone, not just yourself. That explanation could help plant a seed of an idea in that other persons mind and have them googling that little bit quicker or tackling a problem in a slightly more effective way. That same person could come back to you, the coder-extraordinaire, and pose a question or challenge in a way you’ve never considered, giving you a sudden bout of motivation and a new teammate to help                                crack that million dollar question. Final Thoughts Technology is not magic. It’s built by people, for people, to do things that look like magic. It’s important that when faced with a misplaced ear, you gently explain where the camera is on the phone. Showing other people how they can do what you’ve done, helping to inspire curiosity in others, benefits everyone and actually helps to improve things long into the future. Previously published at https://www.automateall.co.uk/MagicBehindCode

The Empathic Importance of Explaining The Magic Behind Your Code in Laymen's Terms

The world is a much more connected place than it was 20 years ago. With the change came new expectations. Downtimes can cripple businesses. There is an expectation of privacy and security. If we post on Facebook and the post gets lost due to any kind of issue, people will not tolerate it. Data integrity and durability are more critical than ever. As an organization, we must consider multiple technological tenets—security, availability, latency, durability, integrity, reliability, and privacy. We don’t get to choose from this list; we just sort it by importance based on business needs. Business continuity Disaster can come in many flavors, ranging from natural disasters to hardware failures, misconfiguration, and cyberattacks. The problem with cyberattacks is that we risk having our data stolen and wonder if the rest of the data is trustworthy. Maybe the attack was a sleight of hand, and the primary objective was to introduce malicious data that would haunt the organization forever. After the security incident, we need to recover the data and restore the service to the level our customers expect. One of the issues we face is reconstructing the original picture and configuration. We need to understand how all the pieces work together and their interdependency. Data restoration is more than just restoring a highly scalable web service snapshot. Objective At this point, our primary objective should be to ensure that our services are available and reliable by whatever means necessary. In the context of dependency management, we need to have a comprehensive data and dependency catalog, including the relationships between the entities. This provides us with the necessary insight into what exists. We also need regular operations testing and robust disaster recovery through automation and manual runbooks. We need to cultivate a culture of disaster recovery fire drills and identify fallbacks. Maturity Model Like an individual who enters college goes through freshman, sophomore, junior, and senior years before graduating, businesses need to follow a slow, progressive approach. It’s impractical for a company or organization to go from no disaster recovery to a fully-fledged, perfectly fleshed-out plan. The road to success is incremental, and each step builds on the previous step. Maturity models examine an organization’s current state from any specific perspective, such as disaster recovery, and help the organization reach the desired state or capability. As the organization progresses toward the end goal, it goes through various checkpoints, usually called Levels. Most Maturity Models have four levels, starting from Level 1 up to Level 4. Why are we talking about maturity models? One step in such a disaster recovery maturity model requires that the organization understand dependencies among various entities and data objects. Recovery usually requires restoring the data and bringing back critical systems in a specific incremental manner, Implementation Based on the business needs, understanding dependency and its relationship can be somewhere in the middle levels of the maturity models. In a 4-level system, it is most likely in Level 2 or Level 3. The other approach is to split the Dependency Management into the same number of milestones as the number of levels. Suppose we are using a 4-level maturity model. In that case, we can split up dependency management to span all 4 Levels, starting with the most straightforward steps and going all the way to the most complicated, expensive, and involved steps. Going forward, we will only discuss the dependency management aspect of this maturity model. Level 1 - Discovering and Understanding Dependencies For many organizations, this can be the most daunting task. If dependency management was never considered necessary, there might be no institutional knowledge. We will start by discussing technical dependencies and then discuss process dependencies. Regarding technical dependencies, we need to identify the direct dependencies first. If the code is written in JavaScript and runs on a cloud service, then the direct dependencies might be JavaScript, NodeJS, AWS Lambda, etc. If we use npm for package management, the packages listed in the dependency file are direct dependencies. Next, we have transitive dependencies, which are pulled automatically from direct dependencies. In practice, this forms a directed acyclic graph. Another aspect of technical dependencies is the cloud architecture. Let’s take an example of AWS—we might use EC2, S3, API Gateway, CloudFront, ELB, etc. If the cloud resources were deployed using tools like CloudFormation or CDK, then it would be trivial to determine the dependencies. If the resources were created manually, the dependency diagram must be made manually by reading the source code and drawing conclusions. Process dependency falls under the non-technical category and defines how different stakeholders work together to move forward. It can also outline the release process or runbooks to invoke during an incident. Level 2 - Setup Continuous Dependencies Validation Dependencies can change since there are too many cooks in the kitchen. The organization's continuous integration tool should enumerate and track the direct code and library dependencies. Every time the code is pushed, a snapshot of current pinned dependencies should be saved to recreate the specific circumstances or restore to a known good state. Like the Continuous integration concept, the non-technical aspects should have some automation. Alerts should be created on metric, and using tools like PagerDuty, appropriate people should be notified of issues, and proper escalation channels should be created. It should be tested occasionally to avoid the known problem called Dogs Not Barking - a situation when you wonder if you are not getting any alerts because nothing is wrong or the alerting logic is broken. Some dependencies can be validated by using the Scream test. Under proper notification and preparation, turn off specific dependencies and see what fails. You may find a Load Bearing Mac Mini. Collect all of this configuration and manage it through some kind of provisioning tool like Chef and Terraform Level 3 - Fixing dependencies by adding, updating, or removing Once we have all the dependencies figured out, take some time and pat yourself on the back. Now, back on the issue, the work isn’t done. During the quest, you realize that some dependencies are causing more problems than they solve. One approach would be to update the dependencies. For example, if we are using old versions of OpenSSL, updating them to newer versions can solve many of the warnings raised by the build system and even help us pass any security compliance. Sometimes, the specific versions of two separate packages are incompatible. The solution is to find versions that can work together. This involves changing the version and might involve downgrading one or more packages. Sometimes, just removing a package or dependency solves the problem. It could have been a dead dependency with no further need but causing warnings and spamming the application logs. Under some circumstances, we might have to add a dependency to bring harmony. Suppose a transitive dependency is causing the problem, but we cannot update the direct dependency. In that case, explicitly adding the fixed version of the transitive dependency as a direct dependency can be an acceptable solution. Level 4 - Manual and Automated Disaster Recovery Drills This is the highest form of maturity in disaster recovery. We need runbooks that outline the roles and responsibilities of individuals involved. The runbook should also define the interdependencies and criticality of each step. The disaster recovery process should be practiced regularly by simulating a disaster and identifying possible sources of friction that get introduced when the plan is set in motion. The gold standard should be automated disaster recovery under human supervision. Automating the disaster recovery workflow ensures that humans do not make mistakes that go unnoticed. Automating has an additional advantage in that we can now run stress tests and simulation disasters by turning off certain network parts in a controlled manner to validate whether our automated disaster recovery is performing as expected. Conclusion Dependency management plays a critical role in disaster recovery, which is essential after a security incident. By following a structured and methodological approach, organizations can build their understanding progressively and control the dependencies. Each maturity model level builds on the previous one and strengthens the organization's security posture. While it can be easy to write off dependency management as a superfluous technical exercise, it is primarily about creating a comprehensive framework that brings together computer systems, people, and processes to work harmoniously. The systematic approach ensures that when disaster strikes, even in the form of a security breach, we can safely recover the data and be confident in the integrity of the said data. The world is a much more connected place than it was 20 years ago. With the change came new expectations. Downtimes can cripple businesses. There is an expectation of privacy and security. If we post on Facebook and the post gets lost due to any kind of issue, people will not tolerate it. Data integrity and durability are more critical than ever. As an organization, we must consider multiple technological tenets—security, availability, latency, durability, integrity, reliability, and privacy. We don’t get to choose from this list; we just sort it by importance based on business needs. Business continuity Disaster can come in many flavors, ranging from natural disasters to hardware failures, misconfiguration, and cyberattacks. The problem with cyberattacks is that we risk having our data stolen and wonder if the rest of the data is trustworthy. Maybe the attack was a sleight of hand, and the primary objective was to introduce malicious data that would haunt the organization forever. After the security incident, we need to recover the data and restore the service to the level our customers expect. One of the issues we face is reconstructing the original picture and configuration. We need to understand how all the pieces work together and their interdependency. Data restoration is more than just restoring a highly scalable web service snapshot. Objective At this point, our primary objective should be to ensure that our services are available and reliable by whatever means necessary. In the context of dependency management, we need to have a comprehensive data and dependency catalog, including the relationships between the entities. This provides us with the necessary insight into what exists. We also need regular operations testing and robust disaster recovery through automation and manual runbooks. We need to cultivate a culture of disaster recovery fire drills and identify fallbacks. Maturity Model Like an individual who enters college goes through freshman, sophomore, junior, and senior years before graduating, businesses need to follow a slow, progressive approach. It’s impractical for a company or organization to go from no disaster recovery to a fully-fledged, perfectly fleshed-out plan. The road to success is incremental, and each step builds on the previous step. Maturity models examine an organization’s current state from any specific perspective, such as disaster recovery, and help the organization reach the desired state or capability. As the organization progresses toward the end goal, it goes through various checkpoints, usually called Levels. Most Maturity Models have four levels, starting from Level 1 up to Level 4. Why are we talking about maturity models? One step in such a disaster recovery maturity model requires that the organization understand dependencies among various entities and data objects. Recovery usually requires restoring the data and bringing back critical systems in a specific incremental manner, Implementation Based on the business needs, understanding dependency and its relationship can be somewhere in the middle levels of the maturity models. In a 4-level system, it is most likely in Level 2 or Level 3. The other approach is to split the Dependency Management into the same number of milestones as the number of levels. Suppose we are using a 4-level maturity model. In that case, we can split up dependency management to span all 4 Levels, starting with the most straightforward steps and going all the way to the most complicated, expensive, and involved steps. Going forward, we will only discuss the dependency management aspect of this maturity model. Level 1 - Discovering and Understanding Dependencies For many organizations, this can be the most daunting task. If dependency management was never considered necessary, there might be no institutional knowledge. We will start by discussing technical dependencies and then discuss process dependencies. Regarding technical dependencies, we need to identify the direct dependencies first. If the code is written in JavaScript and runs on a cloud service, then the direct dependencies might be JavaScript, NodeJS, AWS Lambda, etc. If we use npm for package management, the packages listed in the dependency file are direct dependencies. Next, we have transitive dependencies, which are pulled automatically from direct dependencies. In practice, this forms a directed acyclic graph. Another aspect of technical dependencies is the cloud architecture. Let’s take an example of AWS—we might use EC2, S3, API Gateway, CloudFront, ELB, etc. If the cloud resources were deployed using tools like CloudFormation or CDK, then it would be trivial to determine the dependencies. If the resources were created manually, the dependency diagram must be made manually by reading the source code and drawing conclusions. Process dependency falls under the non-technical category and defines how different stakeholders work together to move forward. It can also outline the release process or runbooks to invoke during an incident. Level 2 - Setup Continuous Dependencies Validation Dependencies can change since there are too many cooks in the kitchen. The organization's continuous integration tool should enumerate and track the direct code and library dependencies. Every time the code is pushed, a snapshot of current pinned dependencies should be saved to recreate the specific circumstances or restore to a known good state. Like the Continuous integration concept, the non-technical aspects should have some automation. Alerts should be created on metric, and using tools like PagerDuty, appropriate people should be notified of issues, and proper escalation channels should be created. It should be tested occasionally to avoid the known problem called Dogs Not Barking - a situation when you wonder if you are not getting any alerts because nothing is wrong or the alerting logic is broken. Some dependencies can be validated by using the Scream test . Under proper notification and preparation, turn off specific dependencies and see what fails. You may find a Load Bearing Mac Mini . Collect all of this configuration and manage it through some kind of provisioning tool like Chef and Terraform Scream test Scream test Load Bearing Mac Mini Load Bearing Mac Mini Level 3 - Fixing dependencies by adding, updating, or removing Once we have all the dependencies figured out, take some time and pat yourself on the back. Now, back on the issue, the work isn’t done. During the quest, you realize that some dependencies are causing more problems than they solve. One approach would be to update the dependencies. For example, if we are using old versions of OpenSSL, updating them to newer versions can solve many of the warnings raised by the build system and even help us pass any security compliance. Sometimes, the specific versions of two separate packages are incompatible. The solution is to find versions that can work together. This involves changing the version and might involve downgrading one or more packages. Sometimes, just removing a package or dependency solves the problem. It could have been a dead dependency with no further need but causing warnings and spamming the application logs. Under some circumstances, we might have to add a dependency to bring harmony. Suppose a transitive dependency is causing the problem, but we cannot update the direct dependency. In that case, explicitly adding the fixed version of the transitive dependency as a direct dependency can be an acceptable solution. Level 4 - Manual and Automated Disaster Recovery Drills This is the highest form of maturity in disaster recovery. We need runbooks that outline the roles and responsibilities of individuals involved. The runbook should also define the interdependencies and criticality of each step. The disaster recovery process should be practiced regularly by simulating a disaster and identifying possible sources of friction that get introduced when the plan is set in motion. The gold standard should be automated disaster recovery under human supervision. Automating the disaster recovery workflow ensures that humans do not make mistakes that go unnoticed. Automating has an additional advantage in that we can now run stress tests and simulation disasters by turning off certain network parts in a controlled manner to validate whether our automated disaster recovery is performing as expected. Conclusion Dependency management plays a critical role in disaster recovery, which is essential after a security incident. By following a structured and methodological approach, organizations can build their understanding progressively and control the dependencies. Each maturity model level builds on the previous one and strengthens the organization's security posture. While it can be easy to write off dependency management as a superfluous technical exercise, it is primarily about creating a comprehensive framework that brings together computer systems, people, and processes to work harmoniously. The systematic approach ensures that when disaster strikes, even in the form of a security breach, we can safely recover the data and be confident in the integrity of the said data.

Dependency Management is Critical for Disaster Recovery After a Security Incident

The massive, catastrophic flash flood that occurred in Spain on October 29 was so horrific and devastating that it left the world aghast at its destructive nature. Over 200 people died, dozens went missing, and whole towns were covered in mud and debris, especially in the eastern Valencia region. In the United States alone, in 2023, 25 weather or climate disasters [caused at least $1 billion in losses and the [deaths](https://www.ncei.noaa.gov/access/billions/#:~:text=In%202023%20(as%20of%20December,and%201%20winter%20storm%20event) of 482 people, according to the National Centers for Environmental Information (NCEI). According to the US Government Accountability Office (GAO), “Severe storms, hurricanes, floods, and wildfires kill hundreds of people and cause billions of dollars in damage in the U.S. each year.” In fact, [according](https://www.ncei.noaa.gov/access/billions/#:~:text=In%202023%20(as%20of%20December,and%201%20winter%20storm%20event) to the National Oceanic and Atmospheric Administration (NOAA), “the U.S. has sustained 400 weather and climate disasters since 1980 where overall damages/costs reached or exceeded $1 billion (including CPI adjustment to 2024). The total cost of these 400 events exceeds $2.785 trillion.” India suffers from a similar problem. The frequency of flooding is getting worse due to the rapid growth of cities and climate change. Numerous other countries, especially in Africa, have seen similar catastrophes caused by unprecedented flooding. Widespread flooding in West and Central Africa, according to the World Health Organization, “has affected more than 4 million people, caused extensive damage to homes, schools, health facilities and swaths of farmland, heightening the risk of water-borne diseases, food insecurity and malnutrition.” Human-induced climate change “is likely to have significantly worsened the situation”, according to scientists at the Red Cross Climate Centre. Many experts now agree that with the advancement of artificial intelligence, this technology can and should play a bigger role in predicting those events and saving lives. And Northeastern University faculty experts predict it will. In an interview last year, Auroop R. Ganguly, professor of civil and environmental engineering and director of AI4CaS (AI for Climate and Sustainability) said, “In the next 12 months, we are going to see more and more efforts where data-driven systems and artificial intelligence come together.” One example is the Earth-2 platform, which is able to simulate global weather patterns in great detail. Taiwan is already using a similar tool to precisely predict the path and impact of typhoons. AI4CaS serves as a focus area within Northeastern University’s Institute for Experiential AI. The GAO found that AI weather forecasting models can help predict these catastrophic weather events. Machine learning specifically “can help improve model accuracy, increase warning times, and reduce uncertainty” and while some of these types of models are currently in use, “there are challenges to developing and adopting them more widely.” Google recently released an AI model that it claims can generate accurate weather forecasts at scale — while being cheaper than conventional physics-based forecasting. The "Scalable Ensemble Envelope Diffusion Sampler" (SEEDS) model is designed similarly to popular large language models (LLMs) like ChatGPT. SEEDS generates many ensembles — or multiple weather scenarios — much quicker and cheaper than traditional predicting models can. The team described its findings in a paper published March 29 in the journal Science Advances. In yet another effort to get ahead of life-threatening weather patterns, the University of Illinois and IBM have joined forces to work together on an AI model capable of anticipating heavy rainfall and flash floods in the Appalachian Mountains. Such models can triple or quadruple the warning time currently available to meteorologists and scientists. In August, NVIDIA Research, announced that it has developed a new AI model for predicting extreme weather events, called StormCast, part of a larger effort by NVIDIA to improve climate research through AI. These new AI models are financially beneficial since they can process information far quicker than older models. As extreme weather events, like flooding, become more frequent and intense, it is becoming increasingly important to be able to predict, understand, and prepare for these events. AI models will assist in predicting weather catastrophes and will help reduce not just major damage, but will also save thousands of lives. Such AI technologies will help scientists make faster and more accurate predictions, which can ultimately save lives and reduce the cost of disasters. In September, Smithsonian Magazine interviewed computer scientist and meteorologist Amy McGovern. She said that while AI will not replace human meteorologists in the near future, “We’re developing things to give them more options so that they can focus on turning the forecasts into actionable information for people to make better decisions.” “Humans are a lot better at that than AI is right now,” she said. The massive, catastrophic flash flood that occurred in Spain on October 29 was so horrific and devastating that it left the world aghast at its destructive nature. Over 200 people died, dozens went missing, and whole towns were covered in mud and debris, especially in the eastern Valencia region. In the United States alone, in 2023, 25 weather or climate disasters [caused at least $1 billion in losses and the [deaths](https://www.ncei.noaa.gov/access/billions/#:~:text=In%202023%20(as%20of%20December,and%201%20winter%20storm%20event) of 482 people, according to the National Centers for Environmental Information (NCEI). According to the US Government Accountability Office (GAO), “Severe storms, hurricanes, floods, and wildfires kill hundreds of people and cause billions of dollars in damage in the U.S. each year.” kill hundreds of people In fact, [according](https://www.ncei.noaa.gov/access/billions/#:~:text=In%202023%20(as%20of%20December,and%201%20winter%20storm%20event) to the National Oceanic and Atmospheric Administration (NOAA), “the U.S. has sustained 400 weather and climate disasters since 1980 where overall damages/costs reached or exceeded $1 billion (including CPI adjustment to 2024). The total cost of these 400 events exceeds $2.785 trillion .” The total cost of these 400 events exceeds $2.785 trillion India suffers from a similar problem. The frequency of flooding is getting worse due to the rapid growth of cities and climate change. The frequency of flooding is getting worse Numerous other countries, especially in Africa, have seen similar catastrophes caused by unprecedented flooding. Widespread flooding in West and Central Africa , according to the World Health Organization, “has affected more than 4 million people, caused extensive damage to homes, schools, health facilities and swaths of farmland, heightening the risk of water-borne diseases, food insecurity and malnutrition.” Widespread flooding in West and Central Africa Human-induced climate change “is likely to have significantly worsened the situation”, according to scientists at the Red Cross Climate Centre . Red Cross Climate Centre Many experts now agree that with the advancement of artificial intelligence, this technology can and should play a bigger role in predicting those events and saving lives. And Northeastern University faculty experts predict it will. In an interview last year, Auroop R. Ganguly, professor of civil and environmental engineering and director of AI4CaS (AI for Climate and Sustainability) said, “In the next 12 months, we are going to see more and more efforts where data-driven systems and artificial intelligence come together.” One example is the Earth-2 platform , which is able to simulate global weather patterns in great detail. Earth-2 platform Taiwan is already using a similar tool to precisely predict the path and impact of typhoons. AI4CaS serves as a focus area within Northeastern University’s Institute for Experiential AI. The GAO found that AI weather forecasting models can help predict these catastrophic weather events. Machine learning specifically “can help improve model accuracy, increase warning times, and reduce uncertainty” and while some of these types of models are currently in use, “there are challenges to developing and adopting them more widely.” Google recently released an AI model that it claims can generate accurate weather forecasts at scale — while being cheaper than conventional physics-based forecasting. The " Scalable Ensemble Envelope Diffusion Sampler " (SEEDS) model is designed similarly to popular large language models (LLMs) like ChatGPT. Scalable Ensemble Envelope Diffusion Sampler SEEDS generates many ensembles — or multiple weather scenarios — much quicker and cheaper than traditional predicting models can. The team described its findings in a paper published March 29 in the journal Science Advances . Science Advances In yet another effort to get ahead of life-threatening weather patterns, the University of Illinois and IBM have joined forces to work together on an AI model capable of anticipating heavy rainfall and flash floods in the Appalachian Mountains. joined forces Such models can triple or quadruple the warning time currently available to meteorologists and scientists. In August, NVIDIA Research, announced that it has developed a new AI model for predicting extreme weather events, called StormCast, part of a larger effort by NVIDIA to improve climate research through AI. announced These new AI models are financially beneficial since they can process information far quicker than older models. As extreme weather events, like flooding, become more frequent and intense, it is becoming increasingly important to be able to predict, understand, and prepare for these events. AI models will assist in predicting weather catastrophes and will help reduce not just major damage, but will also save thousands of lives. Such AI technologies will help scientists make faster and more accurate predictions, which can ultimately save lives and reduce the cost of disasters. In September, Smithsonian Magazine interviewed computer scientist and meteorologist Amy McGovern. She said that while AI will not replace human meteorologists in the near future, “We’re developing things to give them more options so that they can focus on turning the forecasts into actionable information for people to make better decisions.” Smithsonian Magazine interviewed “Humans are a lot better at that than AI is right now,” she said.

Will AI Ever Be Able to Predict Disasters and Save Lives?

Don’t Panic! How Clear Communication and Strategy Can Prevent Cybersecurity Disasters

Diagramming Tools: When Should You Use No Code, DSL, or Code?

Diagrams are fantastic tools for explaining complex systems. They act as a visual map, allowing viewers to grasp how various components interact and connect in a single glance. However, static diagrams sometimes lack the ability to convey intricate processes fully. Imagine presenting a complex system diagram – you can narrate, draw focus to specific elements, and guide your audience through the visual representation. However one of the primary goals of diagrams is to provide a clear explanation without needing a constant narration. Animating diagrams, particularly the flow between different components, bridges the gap between static images and detailed narration. Animations act as a visual guide, directing viewers’ attention and highlighting the order of operations within the system. Imagine seeing a building’s piping system statically on paper (which is still an amazing way to give information about, well the piping in a building). But imagine the drawing animates and shows the direction of the flow of water or sewer system. Increasing the information signal using animations, amplifies the impact of the system diagram. I recently ended up using Drawing.io that has easy way to do this. Most of the time the connections between components of the diagrams are the pieces we want to animate. I looked for this in a lot of other tools I used in the past. Both proper diagramming tools like lucid charts, mermaid, excalidraw, and miro. figma/figjam and general presentation apps like Keynote, and PowerPoint. None have easy animation options. The way I started doing easy animations, is to use my existing drawings, import them as static jpeg/png or SVG into draw.io then draw a connector line over my existing lines, give it a white color and make it thick, then check it’s “animation” option to animate white dashes lines that animate over my existing connector lines, making them animated. Beyond Animation, I like draw.io because it’s a fantastic drawing tool. It’s open-source and free to use and contribute to. Works cross-platform. It has a lot of drawing elements, and you can import your own png, svg icons to your library. You can install draw.io on your desktop (on Mac, using homebrew, as simple as brew install –cask drawio), use the online version for free, or signup for cloud plans to collaborate with your team. Diagrams are fantastic tools for explaining complex systems. They act as a visual map, allowing viewers to grasp how various components interact and connect in a single glance. However, static diagrams sometimes lack the ability to convey intricate processes fully. Imagine presenting a complex system diagram – you can narrate, draw focus to specific elements, and guide your audience through the visual representation. However one of the primary goals of diagrams is to provide a clear explanation without needing a constant narration. Animating diagrams, particularly the flow between different components, bridges the gap between static images and detailed narration. Animations act as a visual guide, directing viewers’ attention and highlighting the order of operations within the system. Imagine seeing a building’s piping system statically on paper (which is still an amazing way to give information about, well the piping in a building). But imagine the drawing animates and shows the direction of the flow of water or sewer system. Increasing the information signal using animations, amplifies the impact of the system diagram. I recently ended up using Drawing.io that has easy way to do this. Most of the time the connections between components of the diagrams are the pieces we want to animate. I looked for this in a lot of other tools I used in the past. Both proper diagramming tools like lucid charts, mermaid, excalidraw, and miro. figma/figjam and general presentation apps like Keynote, and PowerPoint. None have easy animation options. The way I started doing easy animations, is to use my existing drawings, import them as static jpeg/png or SVG into draw.io then draw a connector line over my existing lines, give it a white color and make it thick, then check it’s “animation” option to animate white dashes lines that animate over my existing connector lines, making them animated. Beyond Animation, I like draw.io because i t’s a fantastic drawing tool. It’s open-source and free to use and contribute to. Works cross-platform. It has a lot of drawing elements, and you can import your own png, svg icons to your library. i You can install draw.io on your desktop (on Mac, using homebrew, as simple as brew install –cask drawio), use the online version for free, or signup for cloud plans to collaborate with your team.

Using Draw.io to Animate Systems Diagrams

What started as a tool I developed to simplify visualizing complex database schemas quickly grew into a platform that aims to help developers and data analysts alike. The vision was simple: to create an intuitive, keyboard-friendly database design tool that anyone could use, without the friction of cumbersome clicks or endless UI navigation. The tool supports both open-source languages like ERDL and DBML, giving users the flexibility to choose their preferred modeling language for designing database diagrams. Whether you're a fan of DBML's expressive syntax or prefer the simplicity of ERDL, the tool seamlessly integrates both. How It Works: User-Friendly UI/UX: Just open DatabaseDiagram.com and start typing your schema—your diagram appears instantly, making design fast and easy.
The Toolbar: Manage diagrams, import/export schemas, and share your designs with a click.
Code Editor: Type out your database schema and see your diagram take shape in real-time. The syntax is straightforward—just type, and the diagram is generated.
Visual Diagram: Zoom, pan, and navigate through your diagram effortlessly. Creating Tables and Fields: You can define tables and fields directly in the code editor. Here's the syntax to help you get started: Creating Relationships in ERDL: < one-to-many. e.g: Customer.id < Order.Id
many-to-one. e.g: Order.Id > Customer.id
one-to-one. e.g: Customer.id - Address.id
<> many-to-many. e.g: authors.id <> books.id eg. CustomerID int FK > Customer.CustomerID Import and Export Options Whether you're working with PostgreSQL or MySQL, you can import your existing scripts to instantly visualize your database. Export your diagrams to PDF, PNG, or SQL scripts to share with your team or integrate into your workflow. I’m constantly working on improving DatabaseDiagram.com, and feedback from the community has been invaluable in shaping its evolution. If you love the tool, spread the word and share your feedback! What started as a tool I developed to simplify visualizing complex database schemas quickly grew into a platform that aims to help developers and data analysts alike. The vision was simple: to create an intuitive, keyboard-friendly database design tool that anyone could use, without the friction of cumbersome clicks or endless UI navigation. visualizing complex database schemas The tool supports both open-source languages like ERDL and DBML, giving users the flexibility to choose their preferred modeling language for designing database diagrams. Whether you're a fan of DBML's expressive syntax or prefer the simplicity of ERDL, the tool seamlessly integrates both. ERDL How It Works: How It Works: User-Friendly UI/UX: Just open DatabaseDiagram.com and start typing your schema—your diagram appears instantly, making design fast and easy. The Toolbar: Manage diagrams, import/export schemas, and share your designs with a click. Code Editor: Type out your database schema and see your diagram take shape in real-time. The syntax is straightforward—just type, and the diagram is generated. Visual Diagram: Zoom, pan, and navigate through your diagram effortlessly. User-Friendly UI/UX: Just open DatabaseDiagram.com and start typing your schema—your diagram appears instantly, making design fast and easy. DatabaseDiagram.com The Toolbar: Manage diagrams, import/export schemas, and share your designs with a click. Code Editor: Type out your database schema and see your diagram take shape in real-time. The syntax is straightforward—just type, and the diagram is generated. Visual Diagram: Zoom, pan, and navigate through your diagram effortlessly. Creating Tables and Fields: Creating Tables and Fields: You can define tables and fields directly in the code editor. Here's the syntax to help you get started: Creating Relationships in ERDL: Creating Relationships in ERDL: < one-to-many. e.g: Customer.id < Order.Id many-to-one. e.g: Order.Id > Customer.id one-to-one. e.g: Customer.id - Address.id <> many-to-many. e.g: authors.id <> books.id < one-to-many. e.g: Customer.id < Order.Id many-to-one. e.g: Order.Id > Customer.id one-to-one. e.g: Customer.id - Address.id <> many-to-many. e.g: authors.id <> books.id eg. CustomerID int FK > Customer.CustomerID Import and Export Options Whether you're working with PostgreSQL or MySQL, you can import your existing scripts to instantly visualize your database. Export your diagrams to PDF, PNG, or SQL scripts to share with your team or integrate into your workflow. Import and Export Options I’m constantly working on improving DatabaseDiagram.com , and feedback from the community has been invaluable in shaping its evolution. If you love the tool, spread the word and share your feedback! DatabaseDiagram.com

Designing Database Diagrams Effortlessly with Open-Source ERDL Language

The leading open oracle network, DIA, is taking a bold step forward in the decentralized data market with the launch of 'Lumina.' Lumina addresses the ongoing demand for fully decentralized, trustless data infrastructure in web3. This innovative design also enhances the utility of DIA’s native token, $DIA, by integrating it into a new system that rethinks the way oracles operate. DIA is a decentralized oracle network that provides data feeds for tokens, liquid staking tokens (LST), real-world assets (RWA), random numbers, and other information to all blockchains. The architecture of DIA  is permissionless, enabling nodes and stakers to engage in decentralized data sourcing that is safeguarded by cryptographic and crypto-economic methods. All essential oracle tasks are conducted on DIA's Ethereum layer 2 rollup, named 'Lasernet,' guaranteeing reliable calculations and complete auditability. Powered by Lasernet: A Rollup-based, Modular Oracle Architecture At the heart of Lumina’s advanced design is its modular architecture, which revolutionizes oracle operations from data sourcing to verification, storage, processing, and delivery. All key functions take place on DIA’s native L2 rollup, ‘Lasernet,’ leveraging Ethereum’s latest scaling technologies for optimal performance. This approach provides flexibility and scalability, ensuring oracles remain transparent, secure, and decentralized. Lumina’s architecture uses staking incentives to motivate a distributed network of nodes to collect trade data from over 100 centralized and decentralized exchanges. The native $DIA token secures these contributions, while Zero Knowledge components verify the raw data and computations. With Lasernet managing both data storage and processing, the system ensures on-chain integrity and immutability, sending data where it’s needed via a decentralized messaging protocol. DIA is known as a flexible and transparent cross-chain oracle in the industry. By integrating with over 50 L1 and L2 chains, DIA provides data support for over 200 decentralized applications (dApps) with customizable data feeds for on-chain and off-chain purposes. Zygis Marazas, DIA’s Head of Product, emphasized: "Lumina represents four years of web3 innovation, consolidating it all into one product. It is the cornerstone of our vision for a decentralized, trustless oracle network that guarantees security and flexibility for every use case." A Phased Rollout for Full Decentralization DIA’s Lumina will follow a phased rollout, beginning with the launch of the Lasernet testnet and an early bird staking campaign. Next, the Lasernet mainnet will go live in a closed environment, introducing live staking, followed by an open mainnet that will allow broad participation from the web3 community. “This redesign is a paradigm shift for the oracle space,” said DIA co-founder Paul Claudius. “With Lumina, we are realizing our founding vision of a fully decentralized, transparent oracle network that can be seamlessly integrated by our partners.” Don’t forget to like and share the story. Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR The leading open oracle network, DIA , is taking a bold step forward in the decentralized data market with the launch of 'Lumina.' Lumina addresses the ongoing demand for fully decentralized, trustless data infrastructure in web3. This innovative design also enhances the utility of DIA’s native token, $DIA, by integrating it into a new system that rethinks the way oracles operate. DIA DIA is a decentralized oracle network that provides data feeds for tokens, liquid staking tokens (LST), real-world assets (RWA), random numbers, and other information to all blockchains. The architecture of DIA is permissionless, enabling nodes and stakers to engage in decentralized data sourcing that is safeguarded by cryptographic and crypto-economic methods. All essential oracle tasks are conducted on DIA's Ethereum layer 2 rollup, named 'Lasernet,' guaranteeing reliable calculations and complete auditability. DIA Powered by Lasernet: A Rollup-based, Modular Oracle Architecture At the heart of Lumina’s advanced design is its modular architecture, which revolutionizes oracle operations from data sourcing to verification, storage, processing, and delivery. All key functions take place on DIA’s native L2 rollup, ‘Lasernet,’ leveraging Ethereum’s latest scaling technologies for optimal performance. This approach provides flexibility and scalability, ensuring oracles remain transparent, secure, and decentralized. Lumina’s architecture uses staking incentives to motivate a distributed network of nodes to collect trade data from over 100 centralized and decentralized exchanges. The native $DIA token secures these contributions, while Zero Knowledge components verify the raw data and computations. With Lasernet managing both data storage and processing, the system ensures on-chain integrity and immutability, sending data where it’s needed via a decentralized messaging protocol. DIA is known as a flexible and transparent cross-chain oracle in the industry. By integrating with over 50 L1 and L2 chains, DIA provides data support for over 200 decentralized applications (dApps) with customizable data feeds for on-chain and off-chain purposes. Zygis Marazas, DIA’s Head of Product, emphasized: "Lumina represents four years of web3 innovation, consolidating it all into one product. It is the cornerstone of our vision for a decentralized, trustless oracle network that guarantees security and flexibility for every use case." A Phased Rollout for Full Decentralization DIA’s Lumina will follow a phased rollout, beginning with the launch of the Lasernet testnet and an early bird staking campaign. Next, the Lasernet mainnet will go live in a closed environment, introducing live staking, followed by an open mainnet that will allow broad participation from the web3 community. “This redesign is a paradigm shift for the oracle space,” said DIA co-founder Paul Claudius. “With Lumina, we are realizing our founding vision of a fully decentralized, transparent oracle network that can be seamlessly integrated by our partners.” Don’t forget to like and share the story. Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR Vested Interest Disclosure: This author is an independent contributor publishing via our business blogging program . HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR Vested Interest Disclosure: business blogging program business blogging program

DIA’s Lumina to Roll Out in Phases, Pioneering a Decentralized Approach to Data Management

DIA-nın Lumina, Mərkəzləşdirilməmiş Məlumatların İdarə Edilməsinə Qabaqcıl Yanaşaraq Mərhələlərdə Yayılacaq

Lumina de DIA se implementará en fases y será pionera en un enfoque descentralizado para la gestión de datos

Lumina de DIA lanzarase por fases, sendo pioneiro nun enfoque descentralizado para a xestión de datos

DIA Lumina će se uvoditi u fazama, pionir decentraliziranog pristupa upravljanju podacima

DIA の Lumina が段階的に導入され、データ管理への分散型アプローチの先駆者となる

Lumina firmy DIA będzie wdrażana etapami, będąc pionierem zdecentralizowanego podejścia do zarządzania danymi

DIA’s Lumina Fasekunapi Roll Out, Pionero huk Descentralizado Enfoque Datos Gestión kaqpi

Lumina de la DIA se va lansa în etape, deschizând o abordare descentralizată a managementului datelor

Lumina от DIA будет внедряться поэтапно, став пионером децентрализованного подхода к управлению данными

Lumina ya DIA kugirango izamuke mu byiciro, Pioneer uburyo bwo kwegereza ubuyobozi abaturage uburyo bwo gucunga amakuru

Lumina ของ DIA จะเปิดตัวเป็นระยะๆ เพื่อบุกเบิกแนวทางแบบกระจายอำนาจสำหรับการจัดการข้อมูล

DIA'nın Lumina'sı Aşamalar Halinde Piyasaya Sürülecek ve Veri Yönetimine Merkezi Olmayan Bir Yaklaşımın Öncülüğünü Yapacak

It's one o'clock in the morning. My phone starts going crazy. I groan and look over. Sure enough, it's another PagerDuty call. I am on-call this week, and it's the third time in as many days that I have been woken up for the same reason. Hundreds of our customers have complained in unison, for three days in a row, that they have purchased event tickets from us and arrived at the venue only to find that their tickets have already been redeemed. They are being denied entry, and they are understandably very angry. Initial Investigation Showed Some Very Odd Behavior. Customers had purchased tickets, and then a matter of minutes later, their tickets would be redeemed. This was far in advance of the event itself, so it was clearly unexpected. And it wasn't spread evenly across our customer base; the issue was heavily focused within some select groups of university students. We already knew that we had a fundamental issue with our redemption system; it was possible to scan the QR code in the ticket order confirmation email yourself like the door staff would do at the venue, which would redeem the ticket. Unfortunately, the bigger fix for this was still a few weeks away. Even so, why would large numbers of customers within specific university circles redeem or accidentally redeem their own tickets before the event?! None of our early theories made sense, and the issue kept happening. We Went Digging. Eventually, We Found the Issue. The automated virus scanner on specific university .edu email systems were hitting all links in our confirmation emails to check them for viruses, including programmatically scanning all QR codes. Therefore this one process, completely invisible to my team, was automatically redeeming our customers' event tickets before the customer even received the email. Never, Ever Update Persisted Data on Page-Load. By making the ticket QR code hit a webpage that immediately redeems the ticket on page-load, it meant that anything visiting that link, whether human or machine, would inadvertently update our database to mark the ticket as redeemed. There are thousands of unseen automated visitors to our websites that we don't think about most of the time. Here are some examples: Browsers prefetching resources. This is a big one. Let's say you visit https://myblog.com, and there are prominent links on that page to/blog-post-1 and /blog-post-2 . It is very likely that your browser will already be prefetching those pages in the background. If those pages mutate data on page-load, for example, to track visits, just being on the homepage is enough to trigger it, even though you might never click on the links. Google crawlers. Another big one. We recently saw a large number of errors that came from Google crawlers visiting pages and then triggering unauthenticated calls to APIs. Email scanners. Very common in schools, governments, large corporations, and more. Internet archive machines The simple lesson here is to avoid things happening automatically on page-loads as much as possible. Besides the issues that we experienced, unexpected side-effects like this are bad UX. If the action is "Redeem", make it happen when you click a button that says "Redeem". I hope that this was useful - subscribe now to get my Dev Diary in your inbox! It's one o'clock in the morning. My phone starts going crazy. I groan and look over. Sure enough, it's another PagerDuty call. I am on-call this week, and it's the third time in as many days that I have been woken up for the same reason. Hundreds of our customers have complained in unison, for three days in a row, that they have purchased event tickets from us and arrived at the venue only to find that their tickets have already been redeemed . They are being denied entry, and they are understandably very angry. tickets have already been redeemed very angry. Initial Investigation Showed Some Very Odd Behavior. Customers had purchased tickets, and then a matter of minutes later, their tickets would be redeemed. This was far in advance of the event itself, so it was clearly unexpected. And it wasn't spread evenly across our customer base; the issue was heavily focused within some select groups of university students. We already knew that we had a fundamental issue with our redemption system; it was possible to scan the QR code in the ticket order confirmation email yourself like the door staff would do at the venue, which would redeem the ticket. Unfortunately, the bigger fix for this was still a few weeks away. Even so, why would large numbers of customers within specific university circles redeem or accidentally redeem their own tickets before the event?! None of our early theories made sense, and the issue kept happening. We Went Digging. Eventually, We Found the Issue. The automated virus scanner on specific university .edu email systems were hitting all links in our confirmation emails to check them for viruses, including programmatically scanning all QR codes . automated virus scanner .edu programmatically scanning all QR codes Therefore this one process, completely invisible to my team, was automatically redeeming our customers' event tickets before the customer even received the email. before the customer even received the email. Never, Ever Update Persisted Data on Page-Load. By making the ticket QR code hit a webpage that immediately redeems the ticket on page-load, it meant that anything visiting that link, whether human or machine, would inadvertently update our database to mark the ticket as redeemed. anything There are thousands of unseen automated visitors to our websites that we don't think about most of the time. Here are some examples: Browsers prefetching resources. This is a big one. Let's say you visit https://myblog.com, and there are prominent links on that page to/blog-post-1 and /blog-post-2 . It is very likely that your browser will already be prefetching those pages in the background. If those pages mutate data on page-load, for example, to track visits, just being on the homepage is enough to trigger it, even though you might never click on the links. Browsers prefetching resources. This is a big one. Let's say you visit https://myblog.com , and there are prominent links on that page to /blog-post-1 and /blog-post-2 . It is very likely that your browser will already be prefetching those pages in the background. If those pages mutate data on page-load, for example, to track visits, just being on the homepage is enough to trigger it, even though you might never click on the links. Browsers prefetching resources. https://myblog.com /blog-post-1 /blog-post-2 just being on the homepage Google crawlers. Another big one. We recently saw a large number of errors that came from Google crawlers visiting pages and then triggering unauthenticated calls to APIs. Google crawlers. Another big one. We recently saw a large number of errors that came from Google crawlers visiting pages and then triggering unauthenticated calls to APIs. Google crawlers. Email scanners. Very common in schools, governments, large corporations, and more. Email scanners. Very common in schools, governments, large corporations, and more. Email scanners. Internet archive machines Internet archive machines Internet archive machines The simple lesson here is to avoid things happening automatically on page-loads as much as possible. Besides the issues that we experienced, unexpected side-effects like this are bad UX. If the action is "Redeem", make it happen when you click a button that says "Redeem". I hope that this was useful - subscribe now to get my Dev Diary in your inbox!

Dev Diary #3: The Sneaky Issue That Was Silently Ruining Our Database

Authors:
(1) Rafael Rafailo, Stanford University and Equal contribution; more junior authors listed earlier;
(2) Archit Sharma, Stanford University and Equal contribution; more junior authors listed earlier;
(3) Eric Mitchel, Stanford University and Equal contribution; more junior authors listed earlier;
(4) Stefano Ermon, CZ Biohub;
(5) Christopher D. Manning, Stanford University;
(6) Chelsea Finn, Stanford University. Table of Links Abstract and 1. Introduction 2 Related Work 3 Preliminaries 4 Direct Preference Optimization 5 Theoretical Analysis of DPO 6 Experiments 7 Discussion, Acknowledgements, and References Author Contributions A Mathematical Derivations A.1 Deriving the Optimum of the KL-Constrained Reward Maximization Objective A.2 Deriving the DPO Objective Under the Bradley-Terry Model A.3 Deriving the DPO Objective Under the Plackett-Luce Model A.4 Deriving the Gradient of the DPO Objective and A.5 Proof of Lemma 1 and 2 A.6 Proof of Theorem 1 B DPO Implementation Details and Hyperparameters C Further Details on the Experimental Set-Up and C.1 IMDb Sentiment Experiment and Baseline Details C.2 GPT-4 prompts for computing summarization and dialogue win rates C.3 Unlikelihood baseline D Additional Empirical Results D.1 Performance of Best of N baseline for Various N and D.2 Sample Responses and GPT-4 Judgments D.3 Human study details C.2 GPT-4 prompts for computing summarization and dialogue win rates A key component of our experimental setup is GPT-4 win rate judgments. In this section, we include the prompts used to generate win rates for the summarization and dialogue experiments. We use gpt-4-0314 for all our experiments. The order of summaries or responses are randomly chosen for every evaluation. Summarization GPT-4 win rate prompt (S). Which of the following summaries does a better job of summarizing the most \ important points in the given forum post? Post: Summary A: Summary B: FIRST provide a one-sentence comparison of the two summaries, explaining which \ you prefer and why. SECOND, on a new line, state only "A" or "B" to indicate your \ choice. Your response should use the format: Comparison: Preferred: <"A" or "B"> Summarization GPT-4 win rate prompt (C). Which of the following summaries does a better job of summarizing the most \ important points in the given forum post, without including unimportant or \ irrelevant details? A good summary is both precise and concise. Post: <post> Summary A: <Summary A> Summary B: <Summary B> FIRST provide a one-sentence comparison of the two summaries, explaining which \ you prefer and why. SECOND, on a new line, state only "A" or "B" to indicate your \ choice. Your response should use the format: Comparison: <once-sentence comparison and explanation> Preferred: <"A" or "B"> Dialogue GPT-4 win rate prompt. For the following query to a chatbot, which response is more helpful? Query: <the user query> Response A: <either the test method or baseline> Response B: <the other response> FIRST provide a one-sentence comparison of the two responses and explain \ which you feel is more helpful. SECOND, on a new line, state only “A“ or \ “B“ to indicate which response is more helpful. Your response should use\ the format: Comparison: <one-sentence comparison and explanation> More helpful: <“A“ or “B“> This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. Authors: (1) Rafael Rafailo, Stanford University and Equal contribution; more junior authors listed earlier; (2) Archit Sharma, Stanford University and Equal contribution; more junior authors listed earlier; (3) Eric Mitchel, Stanford University and Equal contribution; more junior authors listed earlier; (4) Stefano Ermon, CZ Biohub; (5) Christopher D. Manning, Stanford University; (6) Chelsea Finn, Stanford University. Authors: Authors: (1) Rafael Rafailo, Stanford University and Equal contribution; more junior authors listed earlier; (2) Archit Sharma, Stanford University and Equal contribution; more junior authors listed earlier; (3) Eric Mitchel, Stanford University and Equal contribution; more junior authors listed earlier; (4) Stefano Ermon, CZ Biohub; (5) Christopher D. Manning, Stanford University; (6) Chelsea Finn, Stanford University. Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Preliminaries 3 Preliminaries 4 Direct Preference Optimization 4 Direct Preference Optimization 5 Theoretical Analysis of DPO 5 Theoretical Analysis of DPO 6 Experiments 6 Experiments 7 Discussion, Acknowledgements, and References 7 Discussion, Acknowledgements, and References Author Contributions Author Contributions A Mathematical Derivations A.1 Deriving the Optimum of the KL-Constrained Reward Maximization Objective A.1 Deriving the Optimum of the KL-Constrained Reward Maximization Objective A.2 Deriving the DPO Objective Under the Bradley-Terry Model A.2 Deriving the DPO Objective Under the Bradley-Terry Model A.3 Deriving the DPO Objective Under the Plackett-Luce Model A.3 Deriving the DPO Objective Under the Plackett-Luce Model A.4 Deriving the Gradient of the DPO Objective and A.5 Proof of Lemma 1 and 2 A.4 Deriving the Gradient of the DPO Objective and A.5 Proof of Lemma 1 and 2 A.6 Proof of Theorem 1 A.6 Proof of Theorem 1 B DPO Implementation Details and Hyperparameters B DPO Implementation Details and Hyperparameters C Further Details on the Experimental Set-Up and C.1 IMDb Sentiment Experiment and Baseline Details C Further Details on the Experimental Set-Up and C.1 IMDb Sentiment Experiment and Baseline Details C.2 GPT-4 prompts for computing summarization and dialogue win rates C.2 GPT-4 prompts for computing summarization and dialogue win rates C.3 Unlikelihood baseline C.3 Unlikelihood baseline D Additional Empirical Results D.1 Performance of Best of N baseline for Various N and D.2 Sample Responses and GPT-4 Judgments D.1 Performance of Best of N baseline for Various N and D.2 Sample Responses and GPT-4 Judgments D.3 Human study details D.3 Human study details C.2 GPT-4 prompts for computing summarization and dialogue win rates A key component of our experimental setup is GPT-4 win rate judgments. In this section, we include the prompts used to generate win rates for the summarization and dialogue experiments. We use gpt-4-0314 for all our experiments. The order of summaries or responses are randomly chosen for every evaluation. Summarization GPT-4 win rate prompt (S). Summarization GPT-4 win rate prompt (S). Which of the following summaries does a better job of summarizing the most \ important points in the given forum post? Post: Summary A: Summary B: FIRST provide a one-sentence comparison of the two summaries, explaining which \ you prefer and why. SECOND, on a new line, state only "A" or "B" to indicate your \ choice. Your response should use the format: Comparison: Preferred: <"A" or "B"> Summarization GPT-4 win rate prompt (C). Summarization GPT-4 win rate prompt (C). Which of the following summaries does a better job of summarizing the most \ important points in the given forum post, without including unimportant or \ irrelevant details? A good summary is both precise and concise. Post: <post> Summary A: <Summary A> Summary B: <Summary B> FIRST provide a one-sentence comparison of the two summaries, explaining which \ you prefer and why. SECOND, on a new line, state only "A" or "B" to indicate your \ choice. Your response should use the format: Comparison: <once-sentence comparison and explanation> Preferred: <"A" or "B"> Dialogue GPT-4 win rate prompt. Dialogue GPT-4 win rate prompt. For the following query to a chatbot, which response is more helpful? Query: <the user query> Response A: <either the test method or baseline> Response B: <the other response> FIRST provide a one-sentence comparison of the two responses and explain \ which you feel is more helpful. SECOND, on a new line, state only “A“ or \ “B“ to indicate which response is more helpful. Your response should use\ the format: Comparison: <one-sentence comparison and explanation> More helpful: <“A“ or “B“> This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv

Stories, News & Mentions about DIXA

Dixa’s Stories on HackerNoon