Themis AIis in the AI Infrastructure Industry

Themis AI

themisai

💡 Introduction to Git and GitHub This is an open-source introduction to Git and GitHub guide that will help you learn the basics of version control and start using Git for your SysOps, DevOps, and Dev projects. No matter if you are a DevOps/SysOps engineer, developer, or just a Linux enthusiast, you can use Git to track your code changes and collaborate with other members of your team or open source maintainers. The guide is suitable for anyone working as a developer, system administrator, or a DevOps engineer and wants to learn the basics of Git, GitHub and version control in general. 🚀 Download To download a copy of the ebook use one of the following links: Dark mode Light mode 📘 Chapters About the book Introduction to Git Version Control Installing Git Basic Shell Commands Git Configuration Introduction to GitHub Initializing a Git project Git Status Git Add Git Commit Git Diff Git Log Gitignore SSH Keys Git Push Git Pull Git Branches Git Merge Reverting changes Git Clone Forking in Git Git Workflow Pull Requests Git And VS Code GitHub CLI Git Cheat Sheet Conclusion 🌟 Sponsors This book is made possible thanks to these fantastic companies! 💙 DigitalOcean DigitalOcean is a cloud services platform delivering the simplicity developers love and businesses trust to run production applications at scale. It provides highly available, secure, and scalable compute, storage, and networking solutions that help developers build great software faster. Founded in 2012 with offices in New York and Cambridge, MA, DigitalOcean offers transparent and affordable pricing, an elegant user interface, and one of the largest libraries of open source resources available. For more information, please visit https://www.digitalocean.com or follow on Twitter. @digitalocean If you are new to DigitalOcean, you can get a free $100 credit and spin up your own servers via this referral link here: Free $100 Credit For DigitalOcean 👩‍💻 DevDojo The DevDojo is a resource to learn all things web development and web design. Learn on your lunch break or wake up and enjoy a cup of coffee with us to learn something new. Join this developer community, and we can all learn together, build together, and grow together. Join DevDojo For more information, please visit or follow on Twitter. https://www.devdojo.com @thedevdojo 🕸️ Web Page A web page showcasing the eBook: Introduction to Git and GitHub website The web page was built using , a new `kick-ass` drag-and-drop TailwindCSS page builder! Tails DigitalOcean App Platform The will be hosted on the . DigitalOcean App Platform We will utilize the to deploy to the App Platform. "Deploy to DigitalOcean" Button 📹 Mini Video Course Coming soon 💻 Interactive training You can follow the interactive training online here: Interactive Git training The training was built with Katacoda. You can find the Katacoda repository . here For more information on how to use Katacoda make sure to follow the steps from this tutorial here: How to Use Katacoda to Create Highly Engaging Training Tutorials 👋 About the author My name is Bobby Iliev, and I have been working as a Linux DevOps Engineer since 2014. I am an avid Linux lover and supporter of the open-source movement philosophy. I am always doing that which I cannot do in order that I may learn how to do it, and I believe in sharing knowledge. I think it's essential always to keep professional and surround yourself with good people, work hard, and be nice to everyone. You have to perform at a consistently higher level than others. That's the mark of a true professional. For more information, please visit my blog at , follow me on Twitter and . https://bobbyiliev.com @bobbyiliev_ YouTube In case that you want to support me you can . Buy Me a Coffee here 🔥 PDF Generation Tool The project uses developed by . Ibis Mohamed Said Ibis is a PHP tool that helps you write eBooks in markdown. 🎨 Book Cover The cover for this ebook was created with . Canva.com If you ever need to create a graphic, poster, invitation, logo, presentation – or anything that looks good — give Canva a go. 🔗 Links My blog: https://blog.bobby.sh Free $100 Credit For DigitalOcean: https://m.do.co/c/2a9bba940f39 Join DevDojo: https://devdojo.com?ref=bobbyiliev Ibis: https://github.com/themsaid/ibis/ Canva: https://www.canva.com/ Tails: http://devdojo.com/tails Katacoda: https://www.katacoda.com/ 📖 Other books Introduction to Bash Scripting eBook 🤲 Contributing If you are contributing 🍿 please read the before submitting your pull requests. contributing file

Step-to-Step eBook to Learn the Git and GitHub Basics💡

How to protect your email on Office 365. There’s a constant threat to your organization. Email is one of the most common ways hackers attempt to breach your organization’s security. Over half of the messages received are spam, phishing campaigns, and malicious. Staying up-to-date with the latest email related threats isn’t an easy task. It’s virtually every company uses third-party email hosting with built-in spam protection or opts for third-party spam protection in front of on-premise servers. Like its competitors, Office 365 comes with built-in protection for your environment, Microsoft’s Information Protection team is staying up-to-date and evolving the email security of Office 365 to protect Microsoft customers from threats. The default configuration isn’t always enough to keep your organization secure and compliant. Why is Spam Still a Problem? Source: ICSI Despite years of combating spam, the campaigns are still relentless. To understand why there’s so much spam we have to ask, Since spammers typically work in the shadows it’s a difficult question to answer. Fortunately, the International Computer Science Institute (ICSI) performed an interesting and surprising . to determine how successful these campaigns truly are. By redirecting the botnet to a fake pharmacy website and throwing an error instead of completing a purchase the ICSI was able to provide insight into the spam plague that has troubled us for so long. How successful is a spam campaign? study The ICSI hijacked an existing botnet By emailing 200 million email addresses the fake pharmacy site led to over 10,000 people visiting the site with 28 purchases. This would have led to almost $9500 a day in sales. What’s Phishing? “Phishing is a cybercrime in which a target or targets are contacted by email by someone posing as a legitimate institution to lure individuals into providing sensitive data such as credit card details and passwords. The information is then used to access important accounts and can result in identity theft and financial loss.” — phishing.org 30% of phishing messages get opened by targeted users and 12% of those users click on the malicious attachment or link. — verizon.com How a Phishing Attack Affected U.S. Elections In September of 2015, the FBI contacted the Democratic National Committee (DNC) to inform them that at least one of their computers has been compromised and sending information to Russian based computers. Six months later a campaign chairman receives a phishing email masked as a fake alert stating another user attempted to breach his account and he needs to change his password. The DNC chairman clicks the link that directs his browser to a fake website where he types in his password giving Russian hackers access to his account. The average cost of a phishing attack for a mid-sized company is $1.6 million. On July 22, 2016, WikiLeaks posts emails that were stolen through the phishing attack. For weeks, the emails consumed the media. No one can say for certain how much the phishing attack affected the election but the damage was done. What is Malware? Malware is any form of malicious code that is often distributed via email. It can be distributed using a direct attachment, stored in a Word or PDF document, or can be downloaded once a person clicks a link in an email. The 1st Email Virus In January 1999, a new threat emerged. Codenamed Happy99, it was one of the first known viruses spread through email. The malicious code would attach itself as an email attachment and send an email with fireworks saying Happy New Year. Unknown to the user, the code would install itself on the computer and begin spreading again. Unlike today, the Happy99 virus wouldn’t do anything other than send itself to the next computer. Today, malicious code is used to steal identities, financial information, and other private corporate information. How is Spam Detected in Office 365? Office 365 uses a multi-layered spam protection system. Every email is scanned using a number of different algorithms to detect and block unwanted messages. Content Filters Source: Microsoft Content filters are used to detect words and phrases that are known to be related to spam. It’s a constant struggle because content continues to change. Microsoft handles the content filtering behind the scenes but you can create a transport rule in EOP to add words and phrases your company would like to block. See the section “Custom Content Filtering in Office 365” below to customize your tenant. Connection Filters Screenshot of EAC Connection FIltering Customizations Connection filters block emails from known bad IP addresses. Spammers will typically use botnets and hijack legitimate companies to send spam. Microsoft will detect an IP address that has an infected device and immediately blocks the bad email from your organization. Once the issue is resolved by the sender, Microsoft will remove the IP address from the block lists and allow email to reach the destination. Additionally, your organization can add IP addresses to a block list or allow list from the EAC. See the section “Custom Connection Filtering” below to customize your tenant. Domain Filtering A domain is the second part of your email address, everything after the @ symbol. For example, the domain of john.gruber@gitbit.org is gitbit.org. Domain filtering is a broad term used to determine the authenticity of the email senders. If a particular domain continues to send unwanted or malicious emails, the entire domain may be blocked. Microsoft uses a number of complex algorithms to determine the validity of a domain. For example, many spammers will create a new domain and immediately start sending spam and malicious code. Since most organizations are established and have had an email system for years, Office 365 is more likely to block an email from a new domain. At the same time, using ‘fuzzy-match’ technology Microsoft will review the sender’s domain for suspicious activity. If someone sends an email from micros0ft.com (a zero replaces the O) it will be blocked. See the section “How to Whitelist and Blacklist Senders as Spam” below to customize your tenant. SPF / Spoofing Protection Technically, SPF filtering could be part of the domain or connection filtering based on how it works. In short, a hacker can send an email from their computer and set the from address to another. While the email will appear legitimate it can be sent from another location. To put this in real-world terms, imagine receiving a letter with your mom’s name and address in the return address field. You would immediately assume the letter originated from your mom but in reality, anyone could have sent the letter. An organization can and should configure an SPF record from each domain they own. For example, if I send email from the IP address 1.1.1.1 using my gitbit.org domain I should create a TXT record to inform the world that 1.1.1.1 is the only IP address that is allowed to send gitbit.org email. // Example SPF record
v=spf1 ip4:1.1.1.1/32 Office 365 automatically reviews SPF records of anyone that sends you an email to keep your organization safe. You’ll need to configure the SPF record for your organization when migrating to Office 365 to help protect your outbound email as well. Microsoft will verify your SPF record after you add it to the Office 365 tenant. I haven’t added a section on how to set up your SPF record. Microsoft provides some assistance to their customers but I recommend asking a professional if you haven’t managed SPF records before. Reputation Filtering Bulk emails are a bit different than spam. A bulk email is typically a newsletter or an offer from organizations that you know. They usually appear in your inbox after signing up for a new website. Source: Microsoft Channel 9 Some users want bulk email while others don’t. Microsoft uses a feedback loop to know when to block bulk emails. If Microsoft receives a certain amount of complaints they will start to block a bulk message. See the section “Spam and Bulk Actions” and “Adjusting the Reputation Filter” below to customize the reputation filtering for your organization. Malware / Attachment Filtering Screenshot of EOP’s Malware Policies Malware filtering is fairly straightforward. Microsoft will scan and block emails that contain known malware or suspicious code. “Using multiple anti-malware engines, EOP offers multilayered protection that’s designed to catch all known malware. Messages transported through the service are scanned for malware (viruses and spyware). If malware is detected, the message is deleted.” — Microsoft You can customize the attachment filter by blocking and allowing certain types of attachments. For example, most organizations don’t need VBS files so you may want to add it to your blocked attachment list. Microsoft does block the most dangerous attachments from your organization out of the box. See the section “Protecting Office 365 from Malicious Attachments” below to customize the reputation filtering for your organization. Defending an Organization from Email Threats Microsoft has designed Office 365 to support all verticals and organizations. Since every organization’s requirements are different, Microsoft has provided a number of customization options to help you tailor your tenant to your specific needs. Most of these customizations are configurable using the . Exchange Admin Center (EAC) Accessing the Exchange Admin Center Most of the customizations are provided through the Exchange Admin Center (EAC). Only administrators will be able to find and change options so be sure the account you use to log in is an admin in your Office 365 tenant. Go to https://outlook.office365.com/ecp Login with your admin credentials. Customizing Office 365 SPAM Filtering Customizing the SPAM filtering can be done from . EAC > protection > spam filter Screenshot of Office 365 Exchange Admin Center spam filter Typically, there is only one policy used across the organization called default. Double-clicking the policy will open the spam filter options window. The spam settings cover a number of options included reputation filtering, block/allow email or domain, and other advanced options. Spam and Bulk Actions The second tab in the spam filter options window gives the most common and broad setting adjustments. These settings are used to customize the reputation filtering for your organization. screenshot of Office 365’s spam and bulk actions What should happen to spam/bulk email? The first section gives you options on how to handle spam. Microsoft classifies a message as good, spam, and high confidence spam (meaning Microsoft is quite sure it’s spam). By default, spam/bulk email will appear in a users junk folder. Another common option is to move the messages to the . Quarantined messages won’t be delivered to the user’s mailbox. Instead, they’ll be moved to a special place only accessible by administrators. Currently, the quarantine is located in . An administrator will need to release an email from the quarantine prior to the user seeing the message in their mailbox. quarantine EAC > protection > quarantine Adjusting the Reputation Filter Some organizations prefer more bulk emails to arrive in the junk email folder while other organizations would prefer bulk email to arrive in the inbox. Microsoft uses a rating system of 1–9. Nine meaning Microsoft is confident it’s unwanted bulk email while one means it’s probably a good email. Screen to adjust Office 365’s spam confidence level You can adjust the bulk email rating for your organization to meet your needs. How to Whitelist and Blacklist Senders as Spam The next two tabs (block lists, and allow lists) allow you to define your organizations black and white lists. Whitelisting an email address or domain will allow all mail from the sender through the spam filter. Microsoft refers to the whitelist as Allow List. If you are having difficulty receiving emails from a person or partner that you trust you can add their email addresses or domains to the allow list. Blacklisting an email address or domain will block all mail from the sender through the spam filter. Microsoft refers to the blacklist as Block List. If you’re receiving a lot of spam from a particular company or person you can add them to the block lists. In the screenshot below I’ve marked all email from marketing@contoso.com as spam. Any other email from contoso.com users will travel through the standard spam filtering policy. I’ve also marked all email from the gmail.com domain as spam. The gmail.com policy will mark any email that comes from someone using Gmail as spam. screenshot of Office 365 block lists A lot of CRM systems, forms, and third-party tools will have emails caught in the spam filters. To make sure people receive the emails to their inbox you should add them to the allow list. In the screenshot below I’ve added website-forms@themisdigital.com to the allow list. This will verify all of my email forms from website-forms@themisdigital.com go directly to a person’s inbox even if Microsoft thinks the email is spam. I’ve also added fabrikam.com to my allow list as a domain. Fabrikam is a fictitious organization I partner with so I’ve verified any emails that come from this domain will arrive at my users’ inbox. Protecting Office 365 from Phishing Attacks Phishing attacks are one of the most dangerous and difficult attacks to protect your organization. The best way to protect your organization from phishing attacks is by training and testing your users. Below is a quick list of things your users should know to help protect against phishing attacks: Don’t trust the Display Name in the sender information. Review the email address. Check the from address carefully. Verify the letter o isn’t replaced with zeros, etc. Don’t trust an email because it has your name or other personal information on it. A lot of information is available through social media and online that may be used to give you false confidence. Hover over links to verify the link is going to the correct URL. When in doubt, don’t click a link. Instead, go directly to the company’s website and log in. Ask the IT department to verify the email prior to taking any action. Protecting Office 365 from Malicious Attachments Out of the box, Microsoft will scan and block emails with attachments that fail a virus check. You can add extensions that you want to block to help keep your organization secure. For example, “.VBS” files are Visual Basic Script files that can be used to pass malicious code through email. Since most organizations will never receive a good VBS file through email it’s a safe bet to block them from entering your organization. Customizing the attachment filtering can be done from the . Double click the policy to view your options. EAC > protection > malware filter Default Screenshot of Office 365 Exchange Admin Center malware filter From the tab, you can enable which will block dangerous file types. Once turned on you can add and remove file types to customize the malware filter for your organization. settings Common Attachment Types Filter screenshot of Office 365’s Common Attachment Types Filter Custom Content Filtering in Office 365 Content filtering is difficult to customize. It’s easy for spammers to change the wording or your IT staff may get too aggressive and block good messages. “Viagra” can typically be blocked in most organizations and it’s common for spammers to send emails about Viagra. To block content we can create a transport rule: EAC > mail flow > rules > plus sign (+) > Create a new rule… Create a rule with the following settings: Name: Block Content (or anything you want) Apply this rule if…: The subject or body includes. Add the word viagra to the list of words. Do the following… Delete the message without notifying anyone. Save your new rule. Transport rules can take up to an hour to take effect. If you click you’ll have a number of other options. After clicking more options you can update the selection with . Doing so will provide the sender with a detailed error so they know to remove the word. More options… Do the following… Reject the message with the explanation Custom Connection Filtering Connection filtering is fairly straightforward. You can allow or block IP addresses through your SPAM filter. To access the connection filtering go to . EAC > protection > connection filter > double-click Default Office 365 Connection Filtering From the connection filtering tab, you add the IP address or range to the filters. Microsoft also provides the ability to enable or disable the “safe list”. The safe list is a range of IP addresses that Microsoft knows are used by safe senders. Any changes may take up to an hour to take effect. There are more options that I haven’t covered. If you’d like to more about a particular feature please let me know in the comments. Thanks,

Introduction to Email Threats & Defense

How Build Your Own AI Confessional: How to Add a Voice to the LLM

কীভাবে আপনার নিজের এআই স্বীকারোক্তিমূলক তৈরি করুন: কীভাবে এলএলএম-এ একটি ভয়েস যুক্ত করবেন

So bauen Sie Ihren eigenen KI-Beichtstuhl: So fügen Sie dem LLM eine Stimme hinzu

Cómo construir su propio confesionario de IA: cómo agregar una voz al LLM

Comment créer votre propre confessionnal d'IA : comment ajouter une voix au LLM

अपना खुद का AI कन्फेशनल कैसे बनाएं: LLM में आवाज़ कैसे जोड़ें

独自の AI を構築する方法 告白: LLM に声を加える方法

자신만의 AI 고백을 구축하는 방법: LLM에 음성을 추가하는 방법

Como construir seu próprio confessionário de IA: como adicionar uma voz ao LLM

Как создать свою собственную исповедь искусственного интеллекта: как добавить голос к LLM

Kendi Yapay Zeka İtirafınızı Nasıl Oluşturabilirsiniz: Yüksek Lisans'a Ses Nasıl Eklenir?

Cách xây dựng chuyên gia AI của riêng bạn: Cách thêm tiếng nói vào LLM

如何构建自己的 AI 忏悔室：如何为法学硕士 (LLM) 添加声音

114 Stories To Learn About Btc

Emphatic words must stand in emphatic positions

No. Thank you and goodbye. Ok, first off, I apologize for the click-bait style title. Every now and then, when I get an idea for a demo, it doesn't work. But , it doesn't work out in a fun and interesting way, and I figure it's a good idea to share it anyway. (Also, there's always the chance that it didn't work out because I did something wrong!) Today's demo is a perfect example of that, I think. sometimes strong Earlier this year, I built a using Alpine.js and the quite excellent . Yesterday, I was thinking about the game, and I wondered - what would happen if I used generative AI to ask for help when playing it? So, for example, I went to and saw this: Blackjack game Deck of Cards API my game In Google's PaLM Makersuite app, I wrote: I'm playing blackjack and the dealer has a six of diamonds showing. I've got a jack of clubs and a jack of hearts. Should I hit or should I stay? And got this response: You have 20, which is a good hand. The dealer has 16, which is below the average. If you hit, you risk getting a card that will bust you. So it's better to stay and hope that the dealer busts. Which, ok, is pretty obvious. You would never hit with two Jacks. But I loved the descriptive response that reinforces the principles of the game. I thought - what if I added a button to the game that lets you ask GenAI for help? Here's how I did it: The Front End On the front end, I kept the UI changes pretty minimal. I began by adding a simple button: Clicking this needs to do two important things. First, it needs to 'translate' the game state into a prompt; then, it needs to pass this to a service that will handle the call. I created a new function, , that handles it: gen AI askForHelp async askForHelp() {
	this.aiHelp = '<i>Asking our AI friend...</i>';

	// first, "translate" the cards to English
	let q = `I'm playing blackjack and the dealer currently has ${this.pcCards[1].value} of ${this.pcCards[1].suit.toLowerCase()}. I've got a ${this.playerCards[0].value} of ${this.playerCards[0].suit.toLowerCase()} and a ${this.playerCards[1].value} of ${this.playerCards[1].suit.toLowerCase()}. Should I hit or should I stay?`;
	console.log(q);

	let resp = await fetch(`https://eo8dkukowewth66.m.pipedream.net?text=${encodeURIComponent(q)}`);
	let help = await resp.json();

	/*
	note that help has line breaks and stuff in and in theory
	we should map to br or something, but im fine ignoring.
	*/
	this.aiHelp = '';
	alert(help);
}, I store the player and PC cards in two variables that contain an array of cards representing their hands. Each card has a 'value,’ which is either a number or a name, like Jack. Each card also has a suit. In theory, that's probably not critical information, but I figured it couldn't hurt. Here's an example prompt generated by a game: I'm playing blackjack and the dealer currently has 8 of clubs. I've got a 7 of spades and a QUEEN of hearts. Should I hit or should I stay? This is passed off to my backend service (more on that in a second), and the result is then displayed in a JavaScript alert, which is 100% horrible UX, but as this was an experiment, I figured it was good enough. If you're curious, this is what the response was: Stay. You have a total of 17, which is a good hand. The dealer's 8 is not a high card, so you are unlikely to lose if you stay. The Back End For my back end, I whipped up a quick workflow. My workflow consisted of literally just three steps. Pipedream The first is my trigger, an HTTP endpoint. You can see the URL in the front-end code above. The second step is the built-in Google PaLM API action. All I had to do was tell it what to use for a prompt: {{steps.trigger.event.query.text}} My final step simply returned the result: export default defineComponent({
  async run({ steps, $ }) {
    await $.respond({
      status: 200,
      headers: {
        "Content-Type":"application/json"
      },
      body: JSON.stringify(steps.generate_text.$return_value[0].candidates[0].output),
    })
  },
}) Time to Win, right? So... yeah. My first few tests worked great, and then... things went off the rails. Here's an example. The dealer had a seven of hearts. I had a queen of spades and an eight of diamonds. When I asked for help, I got: The dealer's upcard is 7, which is a "hard" 17. That means that the dealer must stand regardless of what you do. You have a "soft" 19, which means that you have an ace that can count as either 1 or 11. You should stay. Ok, sure, I stay, but what in the actual frak? I have 19? The dealer has 17? And it gets better. Once when I had 18 showing, I got the following: should I would recommend hitting. With a total of 18, you are only one point away from 21, which is the winning hand in blackjack. The dealer's seven of hearts gives them a total of 17, which is below the 21 threshold. Therefore, there is a good chance that the dealer will bust if they hit, which would give you the win. However, if you stay, you will not be able to improve your hand and will have to hope that the dealer busts. Hitting on 18. I'm no professional gambler, but I'm pretty sure that's bad advice. This got me thinking about my prompt a bit more. I did some tweaks, like lowercasing the suit, as I was worried it may have confused the AI. I iterated through some options in Makersuite, and consistently, it seemed to struggle with basic math at times. Even when I was explicit: I've got a jack of clubs (worth ten) and an eight of hearts (worth 8). It still thought I had 19. I did have better luck with this: I've got a jack of clubs (worth ten) and an eight of hearts (worth 8). My hand total is 18. And I also tried this: I've got a ace of clubs and an eight of hearts. My hand total is 9 or 19. And that seemed to consistently work correctly. So, with that in mind, I went back to my prompt generation and updated my logic there. Here's the updated version: async askForHelp() {
	this.aiHelp = '<i>Asking our AI friend...</i>';

	let totalOb = this.getCount(this.playerCards);
	let totalStr = 'My hand has a total of ';
	if(totalOb.lowCount === totalOb.highCount) {
		totalStr += totalOb.lowCount;
	} else {
		totalStr += `${totalOb.lowCount} or ${totalOb.highCount}`;
	}
	
	// first, "translate" the cards to English
	let q = `I'm playing blackjack and the dealer currently has ${this.pcCards[1].value} of ${this.pcCards[1].suit.toLowerCase()}. I've got a ${this.playerCards[0].value} of ${this.playerCards[0].suit.toLowerCase()} and a ${this.playerCards[1].value} of ${this.playerCards[1].suit.toLowerCase()}. ${totalStr}. Should I hit or should I stay?`;
	console.log(q);

	let resp = await fetch(`https://eo8dkukowewth66.m.pipedream.net?text=${encodeURIComponent(q)}`);
	let help = await resp.json();
	console.log('RESP', help);
	/*
	note that help has line breaks and stuff in and in theory
	we should map to br or something, but im fine ignoring.
	*/
	this.aiHelp = '';
	alert(help);
}, My Blackjack game has a utility function, , that returns an object containing two values , where Aces are treated as one, and , where one Ace at most is considered eleven. With this, I add to my prompt information about the total. getCount lowCount highCount This help quite a bit, but I still sometimes returned delusional responses. did So... maybe consider it your drunk friend at the gambling table offering advice? I'd love to share this version of the demo publicly, and I can share the code, of course, but I'm disabling the Pipedream workflow so that I don't get charged for it. Feel free to look at the front-end code here: https://codepen.io/cfjedimaster/pen/QWYpQjY/905e31ed6d9e3539c041044b308f5f06?editors=1011 As always, let me know what you think! Also published . here

Can I Build a Generative AI Demo to Help Me Win Big in Vegas?

আমি কি আমাকে ভেগাসে বড় জয় করতে সাহায্য করার জন্য একটি জেনারেটিভ এআই ডেমো তৈরি করতে পারি?

Kann ich eine generative KI-Demo erstellen, die mir hilft, in Vegas groß zu gewinnen?

¿Puedo crear una demostración de IA generativa que me ayude a ganar a lo grande en Las Vegas?

Puis-je créer une démo d’IA générative pour m’aider à gagner gros à Vegas ?

क्या मैं वेगास में बड़ी जीत हासिल करने में मदद के लिए एक जेनरेटिव एआई डेमो बना सकता हूं?

ラスベガスで大勝ちするために生成 AI デモを構築できますか?

라스베가스에서 큰 승리를 거두는 데 도움이 되는 생성적 AI 데모를 구축할 수 있습니까?

Posso criar uma demonstração de IA generativa para me ajudar a ganhar muito em Las Vegas?

Могу ли я создать демо-версию генеративного искусственного интеллекта, которая поможет мне выиграть по-крупному в Вегасе?

Vegas'ta Büyük Kazanmama Yardımcı Olacak Bir Üretken Yapay Zeka Demosu Oluşturabilir miyim?

Tôi có thể xây dựng bản trình diễn AI sáng tạo để giúp tôi thắng lớn ở Vegas không?

我可以构建一个生成式 AI 演示来帮助我在维加斯赢得大奖吗？

These events occupied so much time, that June had numbered more than half its days

Mount Olympus

Welcome! If you’re reading this, you’re probably curious about Angular Signals, the latest gem in the framework’s treasure trove. Known for its streamlined approach to state management, Angular Signals can be a game-changer. But where does it truly shine in real-world applications? Let’s embark on a detailed exploration, touching on three use cases and their practical applications. Angular 1. Form Management Angular Forms, whether Reactive or Template-Driven, are essential for capturing user input. Instead of manually subscribing to form control value changes or using , Signals can represent form control values. This ensures that only relevant components or parts of your application react to changes in form values, making form handling more efficient. valueChanges Angular’s Forms module, be it Reactive or Template-Driven, is a core part of many applications, ensuring seamless user data input. With the introduction of Angular Signals, developers can further enhance the efficiency and simplicity of form handling. Reactive Form Handling with Signals Consider a typical reactive form setup; with Angular Signals, you can use the function to represent form control values as signals. This ensures that only relevant components or parts of your application react to changes in form values. toSignal Let’s create an example component with a form that uses Angular Signals: https://gist.github.com/nomadts/bc93b5960161b8713db95ffbb59b7e8b?embedable=true Angular component with form utilizing Signals In the example above, is created using the function. It represents the value of the 'firstName' form control. Now, instead of manually subscribing to the of the form control, we directly use the signal in our template. The div containing the compliment message for the first name will only be rendered or updated when the first name changes. More than that, we can have computed signals that would have an effect on our template. For example, we can have a form field with the date of birth, which is converted to a using . And another that would calculate and show the age based on the date of birth signal. firstNameSignal toSignal valueChanges writable signal toSignal computed signal Benefits Unveiled: Only the parts of your application that rely on a specific signal will update when its value alters. While not a virtual DOM like React, it does offer focused template updates. Selective Rendering: The function has got your back! It subscribes when a component starts and unsubscribes upon its demise, safeguarding against memory leaks. Smart Subscription Handling: toSignal No more murky waters! With Signals, dynamic values within Angular templates become straightforward and intuitive. Crystal-clear Code: In essence, integrating Angular Signals with forms amplifies their performance and paves the way for a more organized, maintainable coding experience. 2. Dynamic Theming Changing the visual theme of an application based on user preferences or system settings is a common requirement. A signal can represent the current theme. When a user switches themes, the signal updates, and only components that depend on the theme need to re-render, instead of the whole application. Dynamic theming enables users to switch between different color schemes or designs on the fly, enhancing user experience. Using Angular Signals, we can achieve efficient and reactive theming solutions that are both scalable and maintainable. Dynamic Theming with Angular Signals Traditionally, you might have handled dynamic theming with and a service that is using BehaviorSubject to store the state. Using Angular Signals, we can streamline this process: ngClass https://gist.github.com/nomadts/9c6a5e2deaf2b507dbc18b9bcafad223?embedable=true Theme service with Signals https://gist.github.com/nomadts/58d6f6da0c938be0249396219e864f64?embedable=true App component that is consuming the theme service Here’s what we achieved: With Angular Signals, we use the function to create a signal for the theme, removing the need for the BehaviorSubject. Simplified Service: signal In the component, we can directly access the theme using the signal. This ensures that our component template reacts instantly to theme changes. Reactive Theme Switch: By using , we can access the current theme directly, ensuring that the theme state is always synchronized. Synchronized State Access: themeService.themeSignal() Angular Signals crafts a path for dynamic theming that’s intuitive, efficient, and free from excessive coding. Bottom line: 3. Live Data Streams Applications like stock trading platforms or live sports scoreboards require real-time data updates. By integrating with RxJS through the function, Angular Signals can directly represent these data streams. This means that as new data is streamed in, only the affected components get updated, ensuring optimal performance. toSignal Modern often deal with dynamic data that constantly updates in real-time. Such live data streams are typically used in stock trading platforms, sports scoreboards, live chat applications, and more. Angular Signals can be incredibly effective in these scenarios, ensuring that data remains reactive and synchronizes with the UI seamlessly. web applications Handling Live Data with Angular Signals Traditionally, managing live data streams might involve services, observables, and complex subscription logic. Now, let’s reimagine this using Angular Signals: https://gist.github.com/nomadts/d32ce0d984a86b631c70bbbf3d40ed9c?embedable=true Chat Service with Signals https://gist.github.com/nomadts/eb4bf05be7ded9613718d1738043162e?embedable=true Chat Component with Signals Notable Enhancements: Gone are the days of . Now, the Signal takes the stage, ensuring on-the-fly updates. Message Management: BehaviorSubject Signals directly in templates react instantly to live data changes, all without the need for async pipes or subscription hassles. Reactive Templates: Less is more! Reducing manual work enhances clarity and maintenance. Trimmed Down Code: To wrap it up, Angular Signals is revolutionizing how we manage live data streams, offering a method that’s simpler and more efficient. Enough for now. Stay tuned for more in our next part, where we’ll delve into User Profiles, E-commerce Cart Management, and Notification Systems. Remember, the world of Angular Signals is vast, and as more developers dive in, its horizons will only expand. Happy coding! Also published . here

A Guide to Angular Signals With Practical Use Cases (Part 1)

THE MANTIS: HER HUNTING

Introduction, or running ahead of the locomotive So, the about the connection between two DAOs showed that the topic of NFTs has not faded in the hearts of readers. Therefore, today we will try to delve deeper into the and uncover the most interesting vector of Web 3.0: the market of programmable assets and, in particular, NFT 2.0. first article forest How to absorb NFT new formats? I understand that not everyone is interested in an academic format as it can be tedious, boring, and lengthy. However, if you are a true , I recommend the following research on the topic (just in case?): netstalker : , , Opportunities & Challenges; Non-Fungible Token (NFT) Overview Evaluation — The Interactive NFTs; NFT 2.0 : Unlocking NFT Liquidity; MetaFi NFTs; Wrapped . A Deep Dive into the Financialization of NFTs; NFTs & DeFi to Wrapped NFTs; A Beginner’s Guide NFT 2.0?; What is of NFT 2.0; A Brief Overview : New Era of NFTs. NFT 2.0 Certainly! Now that you understand that the topic was not created by me, but has been researched by me for the past three years, let's dive deeper. Shall we? Let's do it! Projects Here are some projects that I was able to study before writing this text: Abacus, Arcade, Bailout, Bend DAO, Cally, Cardinal, Charged Particles, Drops, DAO Envelop, Floor DAO, Fuku, Hook, JPEG’d, JpeX, Meta block,s Mimicr,  Mnemonichq, Mycelium, Nested, Nextround, NFT2point, NFTembe,  NFTfi, NFTftrade, NFTperp, NFTx, NIFTY, Options, OpenLand, Opensky, Optic.xyz, PartyBid, Putty, reNFT, Reservoir, Rmrk.app, Solv protocol, Spicy, Stater, SudoSwap, Swap.net, Taker, Themis, Theos, Triber3, TrustNFT, UpShot,  Vinci, YAWWW. "In total, over the course of three years, there have been more than 50 such projects. And I present this list to you for one purpose, so that you can finally understand that this is indeed a , not just a collection of thoughts on the subject. Well, knowing this, let's try to answer the first important question: "what types of NFT 2.0 are there?" Shall we try? Easily! trend NFT 2.0 types It seems that all the articles listed above lack depth and detail in terms of practicality, and therefore realism. Shall we try to fill this gap? Yes, of course! The following terms can be considered as general (collective): (sometimes mixed with NFT 1.5, such as financial NFTs); NFT 2.0 (although sometimes a subset called dNFTs is understood under this term); Interactive NFTs (there is also no standardized approach here); Smart NFTs ; NFTs Programmable There are also a number of rarer categories. There are also several other less common categories. Of course, this list will be open-ended, but here is a list that appears in the overwhelming majority of research, which tries to divide NFTs by functionality: Fractionalization (Fractional, Unic.ly, NFT20, etc.); Landing (TrustNFT, BendDAO, JPEG'd, etc.); Indexes (Envelop, NFTx, BridgeSplit, Index, etc.); Investment DAOs (The LAO, Flamingo, Palm, Pleasr, etc.); Marketplaces (OpenSea, Rarible, Gem, etc.); Derivatives (Envelop, Fuku, Putty, Nifty Options, etc.); Financial NFTs (Axie, Sandbox, StepN, etc.); Rental (Unitbox, Rentable, Prom, reNFT, etc.); Pricing (UpShot, Taker, SudoSwap, etc.); Analytics (Dune, NFTgo, Nansen, etc.) The problem with this approach is that it mixes NFT 1.0 (on OpenSea, for example), NFT 1.5 (NFTx), and NFT 2.0 (Envelop, JPEG'd, etc.) together. Moreover, there are many projects that will always be at the intersection of these categories: . example Let's try to categorize specifically NFT 2.0. Dynamic NFT (dNF) Example: . chain.link/education-hub/dynamic-nft-use-cases Many GameFi projects are developing this type of NFT (such as through controlled smart contract updates, through NFT rental cycles across different services, and so on), but the most interesting vector appears to be at Chainlink: verifiable data change is crucial for industries such as weather forecasting or insurance, and even more so if they are combined. SharkRace Unitbox Therefore, the essence of dynamic NFTs can be defined as follows: these are NFTs whose metadata can change depending on external (or occasionally internal) conditions. There are several types of dynamic NFTs: Sports-related: The Association; LaMelo Ball; Art-related: Jackson NFT; ; Ogcrystals ; Botanicatrees NFTofDAY; Game-related: Aavegotchi; SharkRace; Major League Baseball (MLB) Real-world sector: The Eternal Pump; Regenerative Resources (RRC); Gold Change NFT; Many others. However, a more nuanced categorization of dNFTs can be based on a deeper basis: , where metadata changes through an oracle (Chainlink and its analogs) dNFT , where metadata changes through joint creation and multisig signing (NFTof.day); cNFT , where metadata affects gameplay and changes on-chain rather than off-chain entities. gdNFT As for , these are so-called "community NFTs" that are created gradually and by different actors, and then changes are sent to the blockchain. This is a kind of synthetic act of creation. cNFT In this sense, the evolution from static NFTs (sNFT) to dynamic ones (dNFT) can be achieved in several ways: through cNFT; through changeable metadata (including with Verifiable Randomness Function (VRF)); by adjusting smart contract variables (SharkRace). So, a (dNFT), or a "living NFT", is an NFT that can change depending on external conditions. In the static NFT model, each NFT can be distinguished from another using a single token identifier and a unique contract address. Metadata, such as images, video files, or other data, can be attached to it, which means that you can own a token that represents a unique digital object. dynamic NFT If you want to learn more about dNFTs, I recommend this and this , or alternatively, this “ . Let's keep going! article supplement ” material Wrapped NFT (wNFT) Despite the fact that different projects call these NFTs differently, it is still worth acknowledging that they all have the same essence in the process: by wrapping interchangeable tokens and/or coins, as well as other NFTs, assets of a completely different order are created. And the term "wrapping" should be strictly technical and organizational, as we all know: wBTC as a custodial solution for transferring BTC to the Ethereum network; Wrapped assets of first-generation cross-chain bridges; Various derivatives created using this technology; Other assets. Among the projects that specifically create wNFTs, I would highlight: (2020 - idea, 2021 - MVP, 2022 - release); DAO Envelop (2020, 2021, 2022); Charge Particles (2021-2023). Solv Finance It should be noted that the emergence of these small projects has influenced much older and larger services, such as Rarible, which since June 2021 has from an NFT marketplace to a protocol; AAVE has tested its approach to programmable assets; and even in the Ethereum ecosystem, the standard was born. Not to mention much initiatives (but no less at the same time). requalified ERC-3525 smaller important Initially, is the process by which a token is locked on the blockchain through a smart contract, and its equivalent is "recreated" on the destination blockchain. is the replacement of one set of token interaction standards with another set. wrapping Token wrapping As for (wrapped NFTs), these are modernized NFTs with extended functional capabilities such as economic setting, on-chain royalties, rental mechanism, time/cost/event locking, protection against devaluation, and . wNFTs fraud protection system Hence, I call wNFTs - NFT 2.0, because this approach seeks to change the principles and rules of fundraising in particular and some other industries as a whole. An example of a general explanation. Approaches related to wrapping are generally similar, although details can always be distinguished: №01, №02, №03. And on these examples, we clearly do not finish, but we need to take one more step. Forward! example example example Financial NFT (fNFT) This passage discusses financial NFTs, which can be classified as both NFT 1.5 and 2.0 depending on their usage: If NFTs are used as collateral for services, then they fall under the 1.5 category, whereas if they are programmable to create innovative practices, then they are considered 2.0. NFTs are an improved version of NFTs with a new token standard that allows for multidimensional asset attributes to be expressed in token form. Financial Several projects have used NFTs in financial applications, such as for fixed-rate lending, for insurance token coverage, and NIFTEX for financing NFTs using ERC-20. The most well-known implementation of financial NFTs is probably the LP token of Uniswap v3, which allows for transferring NFTs with liquidity, high trading ranges, and uncollected fees, making it easier and also providing protection against vampire mining. 88mph Armor Finance pre-embedded encompasses fractionalization, lending markets, investment DAOs, derivatives, and pricing, among other things. This market is evolving rapidly and has a lot of potential. For instance, the growth of lending markets that utilize NFT mechanics has led Uniswap to enter into negotiations with leading NFT protocols for loans and credits in 2022. NFT x DeFi The reasons are: , the obvious conclusion is that evolution is inevitable. Firstly , we can move from a general statement to specifics: for example, the gradual growth of credit markets involving various NFT mechanics has led, in particular, to Uniswap starting negotiations with leading NFT protocols for loans/credit in 2022 (see ). Secondly link From here, we can discover an obvious discrepancy in indicators. What kind? I will show it now. Let's compare : Graph №01 Where we see a clear increase in the "Utility" category. And in this case, the priority remains explicitly with another category, namely, "PFP", which indicates a discrepancy between existing (widely accepted) and emerging (unnoticed) trends. The data can be verified at this link: . nftgo.io/analytics/market-overview But the question is becoming increasingly irrelevant. A lively and innovative sub-market, known as ( + ), emerged around NFT financing - regardless of how creators want to classify the nature of their projects. NFTFi Non-fungible-token Finance It is hardly surprising that such a market has emerged. In 2021, the trading volume on the NFT market was $17.6 billion. This is a staggering 27% of the sales volume in the world's traditional art markets ($65B) in the same year. You know what happened next. But there is something that not everyone pays attention to. For example, one can imagine that NFTFi is . It allows for more efficient forms of NFT trading - from swapping, hedging, fractionalization, and valuation - to full-fledged, automated rentals. simply DeFi applied to NFT Let's look at the largest sub-sectors in NFTFi: loans/credits. Financial applications and tools have emerged on the market that address NFT issues aimed at increasing and improving certainty in asset pricing, as well as improving the use of funds on NFT markets by combining NFTs and other finances to create a better user experience on the NFT market, collectively referred to as " ". liquidity dAPPs NFTfi Until 2022, almost nothing was heard about NFT lending/borrowing, but it gradually began to appear in conversations. There are two main ways of NFT lending/borrowing: P2P (peer to peer) and P2Pool (peer to pool). Representatives here include protocols such as and . They differ in transaction frequency, volume, and price discovery efficiency, but these are details. The main thing is that the overall market is still relatively small compared to the market for interchangeable tokens. And coins. NFTfi BendDAO Current NFT lending/borrowing remains in the realm of "blue chips"; as for low- and mid-level NFTs, they have not yet gained wide recognition in the market due and . to high price volatility the lack of oracles In short, NFTs are becoming suitable for various financial instruments such as collateral, staking, and leverage for exchange tokens or stocks. And? In the future, when the accumulation of historical trading prices gradually complements the NFT quotation mechanism, the NFT market will become more liquid than ever. \But a couple of important theses are still to come. NFT-based lending The Rocket LP DAO cryptocurrency lending platform once issued a loan secured by the brantly.eth domain name of the Ethereum Name Service (ENS). In turn, the press service of the NFT marketplace X2Y2 announced that from October 9, 2022, users could use their own NFT digital collectibles as collateral to borrow Ethereum (ETH) from other lenders. And do you know what all this is about? That the lending market in the DeFi segment will never be the same again. Interesting? Then here are three more articles: , , . one two three But this market is not the only one going through harsh turbulence. What else? You will find out now! Royalties 2.0 Creators of Ethereum-based NFT collections have earned over $1.8 billion in royalties. Impressive, right? By the time this article is published, the figure will likely be even higher. But if a billion dollars doesn't impress you, then there's not much that can. At the time when the staggering sum of $1,800,000,000 was reached, just 10 organizations accounted for 27% of all royalties earned, while 482 NFT collections accounted earned at that time. for 80% of all royalties Since then, much has changed. The DeGods ecosystem removed royalties from all of its affiliated NFT collections (DeGods, y00ts), even though DeGods founder Frank had defended royalties on Twitter and still believes that royalties are between operators and holders of NFT collections. the best mechanism for aligning incentives Fast-forward to 2023 and the showdown between . Interesting, right? OpenSea and Blur "Some marketplaces have also changed their royalty payouts (or removed them), for example, x2y2, while the largest marketplace, OpenSea, has not". First of all, now everyone is following suit. In addition, the largest Solana NFT marketplace, Magic Eden, made a controversial move by making all royalties on its platform completely optional. Magic Eden's latest move to eradicate royalties is especially noteworthy given that they announced MetaShield, a controversial tool aimed at improving royalty control, back in September 2022. Royalty supporters are worried that normalizing their removal will return the NFT space to the dark ages of traditional creator structures, as happened with Van Gogh. incentivization On the other hand, opponents of royalties claim that mechanisms for compulsory collection are impossible on a chain without serious compromises, which negate many of the advantages of blockchain. permissionless If you're interested, I recommend reading this . And as for us, let's move on! thread Fractionalization There are quite a few services here: Fractional.Art, Teller, and . But personally, I'm interested in why so many are focused on a special feature (fractionalizing indivisible assets), when in reality, it's just a small spectrum of the possibilities of NFT 2.0 as a whole. others Rent There are two fundamentally different vectors: Renting through (w)NFT; Renting actual NFTs. These processes are interconnected but can still be separated as distinct. Prom & Unitox, for example, offer similar but different approaches to solving the integration of these processes. In turn, BendDAO and JPED'd have implemented an efficient liquidity model called "peer-to-pool", which connects supply and demand in a customizable liquidity pool, eliminating the need to make bids and wait for requests. Another example is Cyan, the largest and brightest BNPL protocol. Additional information can be found , , or even . But the furthest in the connection of these mechanics is Unitbox - a project based on the DAO Envelop . here here here protocol And what about us? Let's try to combine all the ingredients together and the NFT 2.0 . taste dish NFT 2.0 as understood by ChatGPT Of course I couldn't help but give the floor to one of the protagonists of the start of 2023. Here's what he thinks of NFT 2.0. NFT 2.0" is not an established or standardized term within the NFT (Non-Fungible Token) space, and there is no widely accepted definition of what it encompasses. However, it generally refers to new and innovative developments in the NFT ecosystem that aim to improve the technology and user experience. Some possible features that could be included in an NFT 2.0 include: Interoperability: The ability for NFTs to interact with each other and other blockchain-based assets, allowing for more complex and integrated applications. Programmability: The ability for NFTs to have embedded functionality, enabling them to perform specific actions or operations when triggered. Sustainability: The ability for NFTs to have a smaller environmental impact, through the use of more energy-efficient or carbon-neutral technologies. Decentralization: The ability for NFTs to be more decentralized, removing the need for centralized authorities or intermediaries. Fractionalization: The ability for NFTs to be divided into smaller parts, allowing for greater accessibility and liquidity. Dynamic and Adaptive properties: The ability for NFTs to change or evolve over time, through smart contracts that allow for dynamic or adaptive properties. It's worth noting that these features are not exclusive to an "NFT 2.0" and many are already being developed and implemented by various NFT projects and platforms. NFT technology is still evolving rapidly, and new features and innovations are likely to emerge as the ecosystem continues to grow and mature. Problems Does this market have problems? Plenty. Firstly, there are standards: what we call them are not actually standards (take this or this , or many others). Secondly, there are no leaders, which is good for business, but for the user... they will have to make an effort not just to understand the trends, but also to "taste them" in practice. Finally, there is a lot of chaff in this market, connecting hype with hype (insignificant NFT collections with liquid markets... tokens that are of little use to anyone, for example). example one But none of these problems is evidence that the market will not continue to develop. Rather, they are a consequence of the of this sphere. evolution Outlook I can't tell you about all the vectors in this article, so I've chosen the ones that are closest to me. Unstoppable NFT Now there are already solutions ( ) where NFT connects with decentralized, tokenized storage, but it is much more interesting when the programmable NFT itself pays for storage of its metadata for IPFS/SWARM and other nodes. So far the solutions require oracle and centralization, but I believe the next 2-3 years will see a move towards decentralization. example ОТС Not using such solutions for trading is strange: the secondary market for tokens is highly in demand and reaches tens of billions of dollars, but there are -fledged solutions yet, and funds are quite conservative in their practices, so here too we will have to wait for at least 2-3, or even 4-5 years. OTC not so many full ZKP For me, this is probably the most promising direction, as bridge hacks (Wormhole, Allbridge, Binance Hub, etc.) and CEXs are all about liquidity centralization, and achieving its normal decentralization can be done through -NFTs, which require different approaches: both oracle software and oracle (humans), and improvements to existing concepts. cross verifiers Add in and you get a lot of interesting things. By the way, if this topic turns out to be in demand, I'll talk about it, but for now that's all. cross/multi-chain mechanics Bye!

NFT 2.0: the Era of Programmable Assets

BOOK II

BOOK I

BANDA ORIENTAL AND PATAGONIA

BY PROFESSOR HUXLEY

THE GREEKS AND THE PERSIANS

What can stop a startup from taking off? The idea, the team, and the product are the core of any startup, which grows in sales and takes off into space. To get off the ground, we need a massive acceleration. Whoever is first is the fastest to achieve growth. In this article, I will tell you . how you can speed up the process of product creation and how you can create an admin panel in the shortest possible time What do I need a CMS for? In our company, , we attract developers from different IT sectors to our web platform and suggest they go through our automated system of hard skills assessment. They must pass quizzes and coding tasks. After that, an HR specialist is assigned to the candidate. He conducts an interview with him during which his experience, grade (junior/middle/senior), and soft skills are evaluated. As a result of the interview, HR prepares a detailed CV for the candidate. Insquad Thus, my team of developers and I had the task of creating a CMS system that would allow: Track each candidate’s status and all relevant information in real-time Fill out and edit each developer’s resume Group, filter, sort, and show data in different formats to track statistics and search Manage users with role-based access In addition to the points mentioned above, the system . Since a startup is constantly testing new hypotheses and everything is changing very quickly, you need flexibility. For example, add a column and formula at any time to calculate HR professionals’ KPIs. Besides, creating a simple and intuitive interface is difficult. The CMS needs to implement new features quickly and be flexible, understandable, and scalable. must be flexible with a user-friendly interface What were our solutions? Continuing to process data manually is not our option; we must develop quickly. A customized solution is expensive and time-consuming for a startup like ours. While researching, I identified possible solutions to speed up the creation of the system: Use some libraries, for example, or . This solution has disadvantages: long and expensive to develop; it is not a flexible solution at all) react-admin admin.js An out-of-the-box solution, for example, . It allows you to configure a universal interface for your data storage and has a bunch of integrations. This service provides a dashboard and a set of widgets. It’s a quick solution but not as flexible as I’d like it to be. The big disadvantage of this system is that you have to write a CRUD API to integrate with it. Any addition or change of data involves changes to your working database and changing your API. Forestadmin Airtable is a cloud service with the most versatile interface. It is more flexible than the previous one, because the data is stored on Airtable. The necessary data can be added at any time. For full integration, you only need to implement a synchronization service with your own internal database. Why Airtable? So we settled on Airtable. It acts as a CMS for data processing. How does it differ from Excel spreadsheets and Google sheets? In addition to its user-friendly interface, it is not only about working with spreadsheets. The View functionality in Airtable allows you to show data from a table in various formats: gallery, timeline, calendar, grid, kanban, Gantt chart, etc. Default tools such as sorting, grouping, filtering and views are implemented very simply. Anyone, even a non-programmer, can easily master them. Pros Airtable provides excellent CMS functions for processing and visualizing data Data can be divided into different workspaces, bases and tables For each table, you can customize views Role-based access setup for team members The service is intuitive and easy to learn for employees who are not familiar with programming and databases There are a lot of integrations As a result, you get a large set of features and save time and money. The paid version of Airtable allows you to use the REST API, gives you access to a whole marketplace of integrated services and automations. Automation is a very useful feature of Airtable, which allows you to write JavaScript scripts and automate various processes. In addition to that, Airtable is a stable company with a large capitalization. Cons Now consider the disadvantages of this service: airtable is paid per user A limited number of entries. For example, if you use a Pro subscription, you will be limited to 50k per base You can’t fine-tune access to each field of the table. Everyone has access to the entire workspace at once. API has a limit of 5 requests per second Hack #1 Here’s how I got around this limitation. The data are duplicated in your database and Airtable. It can change for two reasons: Changing data on the side of our server, which updates the data in the private database and Airtable Editing data by the HR department in Airtable To get around the limits in the first case, I wrote a proxy function that buffers all requests and, on timeout, updates the data in Airtable with one batch request. The code of this function: import * as Airtable from 'airtable';
import { RecordData, FieldSet, Table } from 'airtable';

enum Action {
  CREATE,
  UPDATE,
  DELETE,
}

class AirtableProxy {
  private timeout = 1000 * 3; // 3 second delay
  private airtable: Table<FieldSet>;
  private batchCreate: { fields: Partial<FieldSet> }[] = [];
  private batchUpdate: RecordData<Partial<FieldSet>>[] = [];
  private batchDelete: string[] = [];

  constructor() {
    this.airtable = new Airtable({
      apiKey: process.env.AIRTABLE_API_KEY,
    })
      .base(process.env.AIRTABLE_BASE_ID)
      .table(process.env.AIRTABLE_TABLE_NAME);
  }

  private async _registerTimer(action: Action) {
    if (action === Action.CREATE) {
      setTimeout(async () => {
        const data = this.batchCreate;
        this.batchCreate = [];
        await this.airtable.create(data);
      }, this.timeout);
    }

    if (action === Action.UPDATE) {
      setTimeout(async () => {
        const data = this.batchUpdate;
        this.batchUpdate = [];
        await this.airtable.update(data);
      }, this.timeout);
    }

    if (action === Action.DELETE) {
      setTimeout(async () => {
        const data = this.batchDelete;
        this.batchDelete = [];
        await this.airtable.destroy(data);
      }, this.timeout);
    }
  }

  public async create(data: Partial<FieldSet>) {
    const setTimerFlag = !this.batchCreate.length;
    this.batchCreate.push({ fields: data });
    if (setTimerFlag) this._registerTimer(Action.CREATE);
  }

  public async update(data: RecordData<Partial<FieldSet>>) {
    const setTimerFlag = !this.batchUpdate.length;
    this.batchUpdate.push(data);
    if (setTimerFlag) this._registerTimer(Action.UPDATE);
  }

  public async destroy(data: string) {
    const setTimerFlag = !this.batchDelete.length;
    this.batchDelete.push(data);
    if (setTimerFlag) this._registerTimer(Action.DELETE);
  }
}

export default new AirtableProxy(); It is essential to know that the official Airtable for Node.js has built-in logic to handle the speed limit. library Hack #2 To get around the limits in the second option — editing data in the Airtable, we wrote a separate service that transfers all the changes from the Airtable to our private database every few minutes. Link to the repository: https://github.com/AbrSergey/serverless-airtable-sync-cron Overall, I found Airtable suitable for creating a full-fledged CRM system for a project. And the presence of a convenient API and multiple integrations speeds up the project’s launch and saves a lot of money, which is crucial for a startup. Conclusion Since the IT market evolves quickly and unpredictably, a startup needs time to grow and transform at the same speed. Airtable allows you to make a product from scratch in a tight timeframe. And with the several listed hacks, you can build a complex flexible solution. I hope the information I have provided above is helpful and will simplify startup developers' life. If you are scaling your development team, contact us at to scale your team 4x faster with remote startup developers. Insquad

How to Use Airtable as a CMS for a Startup

The Essays of Adam Smith: Part VII, Section II, Chapter I

Stories, News & Mentions about THEMISAI

Themisai’s Stories on HackerNoon