That’s quishing. QR code phishing. It’s the same con as every other phishing scam. The difference is the hook. It isn’t a link in an email anymore. It’s a square of dots you point your camera at without thinking twice.

Why QR Codes Slip Past Your Guard?

We’ve spent years training people to distrust links. Hover before you click. Check the URL. Don’t trust the sender. A QR code throws all of that out, because you can’t read it. Your eye sees a pattern. Your phone sees a web address. You’re trusting blind.

That’s exactly why attackers love them. A QR code in an email sails past filters that would flag a suspicious link. There’s no link in the text to scan. It’s buried in an image. Microsoft analysed 8.3 billion phishing threats in the first quarter of 2026 and flagged a 146% rise in QR code phishing as attackers shifted tactics to dodge defences. In the UK, Action Fraud reported a 587% jump in quishing reports between 2023 and 2025.

It’s not a niche trick either. By 2025, roughly 12% of all phishing attacks carried a QR code. 68% of those went after mobile users specifically. That last number matters. You scan with your phone. Your phone is the device with your banking app, your saved passwords and your one-tap payments.

How the Scam Actually Works

There are three flavours. They end the same way.

The fake payment page. You scan a code expecting your UPI app or a checkout screen. Instead you land on a page that looks exactly like one. Same logo, same colours. It asks for your UPI PIN, an OTP or card details. The moment you type them in, they’re gone. So is your money.

The sticker-over-sticker. The physical version. Someone prints a malicious QR code and pastes it over the real one at a petrol pump, a parking meter, a temple donation box or a shop counter. You pay the right amount to the wrong person. That’s what hit the Pune shopkeeper.

Fake (malicious) app install. In the nastier cases the code doesn’t open a payment page at all. It quietly pushes you to install an app or grant a permission. Now there’s spyware on your phone reading your messages and watching you type your banking password for weeks.

The One Rule That Stops Most of This in India

Here’s what you should always remember: on UPI, you never scan a QR code to receive money. Scanning is how you pay. So if anyone sends you a QR code and says scan this to get your money, it’s a scam. A “buyer” on OLX. A “refund” agent. A stranger doing you a favour. Doesn’t matter, full stop. There’s no exception. They’re trying to make you authorise a payment out, not a deposit in.

A few more habits that close the gaps:

• Read the name before you confirm. Every UPI app shows the payee’s name before you enter your PIN. At a shop, glance at it. If the sweet shop’s code shows a random individual’s name, stop.
• Look at the sticker. A fresh QR pasted crookedly over a faded one is a red flag. Petrol pumps and parking spots are favourite targets.
• Check where the code sends you. Most phone cameras preview the URL before opening it. If it’s a shortened link or a domain that isn’t the brand you expect, don’t open it.
• Never type a PIN or OTP on a page a QR code dropped you onto. Real payments don’t ask you to enter your UPI PIN into a web page.

If You’ve Already Scanned One

Move fast. Same as any drained-account situation.

1. Call your bank and your UPI app’s support line to freeze the account or reverse the transaction. Speed is everything. Some transfers can still be stopped.
2. Change your UPI PIN and banking passwords immediately, pertinent if you entered your PIN and OTP on a fake payment page.
3. Uninstall anything the code prompted you to install. Then run a malware scan.
4. Report it. In India, file at cybercrime.gov.in or call 1930, the cyber-fraud helpline. The first hour is when recovery is most likely.

QR codes aren’t the problem. They’re genuinely useful, which is exactly why nobody questions them. The scam works because we’ve made scanning a reflex. Point, tap, done. The fix isn’t to stop scanning. It’s to spend the half-second reading the name on the screen before you put your money behind it.

Part of the Phishing Scams series. Catch up on Recognising Phishing, CEO & Business Email Compromise and AI Tech Support Scams.

Fast forward to 2026 and the same issues have court cases, legislation and €15 million penalty clauses attached to them. That’s a different conversation.

The EU Drew a Line. It’s Already in Force.

The EU AI Act has been rolling into effect in stages. The list of banned practices has been live since February 2025 and it’s specific: government social scoring systems that rate citizens on behaviour, emotion recognition in workplaces and schools, real-time biometric identification in public spaces, untargeted scraping of facial images from CCTV or the internet to build recognition databases.

Transparency requirements for AI-generated content kick in on August 2, 2026. Violations carry fines up to €15 million or 3% of global turnover, whichever is higher.

That last number matters. Companies are now calculating whether a compliance investment is cheaper than the penalty. That’s a strange place for an ethics conversation to end up, but it’s where we are. Rules without consequences don’t change behaviour. Apparently €15 million does.

Deepfakes Went from Theoretical to Prosecutable

In early 2026, French authorities launched a criminal investigation into Grok, X’s AI system, after it was used to generate non-consensual explicit images of women and teenagers. The system effectively “undressed” real photos on request. French ministers referred the case to prosecutors under the Digital Services Act.

This isn’t a hypothetical scenario from an AI safety paper. It’s a named AI product, named victims, named regulators and an active criminal probe.

My 2023 post used facial recognition misclassification as the flagship example of AI harm. That problem hasn’t gone away. But the threat surface has expanded significantly. The harms are more direct, more personal and harder to wave away as edge cases.

India’s AI Ethics Problem Is Specific and Documented

Most global AI ethics frameworks are written by and for Western contexts. India’s version of this problem has distinct dimensions that rarely make it into the international conversation.

In December 2025, BJP MP Bharti Pardhi introduced the Artificial Intelligence (Ethics and Accountability) Bill in Parliament. It’s a private member’s bill. The odds of it passing as written are low. But it names something important: India-specific bias vectors. Caste, religion, language and socio-economic status.

These aren’t abstractions. India has 22 official languages, a documented history of caste-based discrimination in hiring and credit decisions and a rapidly expanding deployment of AI systems trained predominantly on Western datasets. A 2026 study in the Sage Journal flagged the underrepresentation of marginalised communities in Indian AI datasets not as a future risk but as a present, compounding one.

The bill proposes mandatory bias audits for high-risk AI in surveillance, employment and law enforcement. Penalties go up to INR 5 crore for non-compliance. Whether it becomes law in its current form is almost beside the point. The acknowledgement that “diverse training data” isn’t a good enough answer for a country this complex is progress.

The Accountability Gap Just Got Harder to Ignore

The original question of who’s responsible when AI causes harm was always valid. It’s now acute. AI agents that operate autonomously across multiple steps, booking things, cancelling contracts, submitting data on a user’s behalf, have no clear accountability chain when something goes wrong.

The EU AI Act classifies agentic AI systems under its high-risk category, requiring documented human oversight mechanisms. That’s the right instinct. Whether companies implement those mechanisms genuinely or checkbox their way through the audit is a different question, and one regulators don’t yet have a good answer for.

What’s changed since 2023 is that at least the question is being asked with legal weight behind it.

My 2023 post ended with something about progress not coming at the expense of our values. That’s still true. What’s different now is that our values have penalty clauses, criminal referrals and enforcement deadlines attached to them. Whether that’s enough to actually change how AI gets built and deployed is the question worth watching in the next 3 years.

The Lede image was a re-render of the picture from the original post.

I was incredibly lucky to be working out of One World Trade Center for a few weeks while I was there. Being up that high provided an unparalleled, panoramic view of the Manhattan skyline every single day. Here are just a few samples from my camera roll.

Microsoft’s April 2026 Patch Tuesday wasn’t a routine update cycle. It fixed a staggering 167 security vulnerabilities across Windows and related software, including a SharePoint zero-day already being actively exploited in the wild, and a Windows Defender privilege escalation bug whose exploit code got publicly leaked by a frustrated researcher. That’s not a Tuesday. That’s a crisis with a calendar invite.

And here’s what’s making this worse: AI is changing the speed of the game, for both sides.

The AI Vulnerability Firehose Is Now Open

April’s patch count included nearly 60 browser vulnerabilities alone, a new record in that category. Analysts pointed to Anthropic’s Project Glasswing as a likely contributor to the spike. Announced on April 7, Glasswing is built around Claude Mythos Preview, an unreleased AI model that Anthropic describes as its most capable yet for coding and agentic tasks. In early testing, it found thousands of zero-day vulnerabilities across every major operating system and browser, including a 27-year-old flaw in OpenBSD and a 16-year-old bug in FFmpeg that automated testing tools had hit 5 million times without catching.

As The Hacker News put it: fewer than 1% of the vulnerabilities Mythos found have been patched. The most powerful vulnerability discovery engine ever built ran against the world’s most critical software, and the ecosystem couldn’t absorb the output. Glasswing solved the finding problem. Nobody solved the problem of fixing.

That’s the gap that should keep security leaders awake at night.

Slow to Patch Is the Same as Unpatched

Discovery means nothing if the fix doesn’t reach users.

Take the Adobe Reader vulnerability. CVE-2026-34621 had been actively exploited since at least November 2025, nearly 5 months before Adobe shipped an emergency patch. According to Malwarebytes’ analysis, simply opening a malicious PDF was enough to trigger it. No extra clicks. No suspicious prompts. Just open the file and your system gets fingerprinted and phoned home to an attacker-controlled server.

The BlueHammer story is just as damning. A researcher known as Chaotic Eclipse published working exploit code on GitHub on April 2 after growing fed up with how Microsoft’s Security Response Center handled their disclosure. Attackers were already using it in the wild by April 10, four days before Microsoft patched it on Patch Tuesday. That’s a 12-day window of exposure for every Windows machine on the planet, handed to attackers because a vendor moved too slowly.

This is the gap that kills companies.

India’s Patch Problem Is Getting Dangerous

Indian enterprises, in particular, have a real problem here. According to BW Businessworld, patch management weaknesses persisted across sectors through 2025, with Kaspersky recording over 4,000 daily exploit attempts targeting unpatched and legacy software. PwC found that 60% of Indian security leaders lacked adequate knowledge to apply AI effectively for cyber defence. The average cybersecurity maturity score for Indian organisations, per Seqrite’s 2026 report, sits at 6.37 out of 10.

That number tells the whole story. India digitised fast and secured slowly. Mid-sized companies, digital platforms, and fast-scaling startups are increasingly the preferred targets precisely because of weaker governance and inconsistent vulnerability management. Security patch cycles in many organisations are still measured in quarters, not days.

Attackers aren’t sending letters ahead of time. They’re running the same AI-powered discovery tools that defenders are, just pointed at your infrastructure instead of their own code.

Hiding Vulnerabilities Isn’t a Strategy

There’s another pattern worth calling out: the instinct to bury security incidents rather than address them. The SharePoint zero-day in this month’s Patch Tuesday, CVE-2026-32201, enables phishing attacks, unauthorised data manipulation, and social engineering campaigns. The kind of flaw that, if quietly shelved in a backlog, will eventually show up as a breach headline.

When AI tools, yours or an attacker’s, can find these holes faster than ever, the “we’ll deal with it later” approach has a dramatically shorter shelf life. Companies that don’t build rapid patch deployment into their security operations are essentially leaving a window open and hoping no one notices the ladder outside.

Your Laptop Has Pending Updates Right Now

This isn’t just a corporate IT problem. Individual users are just as exposed.

Google Chrome has already fixed 4 zero-days in 2026 alone. If your browser is open right now with 40 tabs, odds are good it hasn’t restarted recently enough to apply a security update that’s been waiting for it. The fix is simple: restart your browser. Run Windows Update. Update Adobe Reader. Check for macOS or Android system updates.

This is boring advice, but boring advice is what actually prevents most breaches. The patches exist precisely because someone found the hole, and in 2026, “someone” is increasingly an AI model running 24/7.

Security updates aren’t a chore. They’re the last line of defence between you and someone who’s already automated the attack.

The pace of vulnerability discovery is accelerating and it won’t slow down. AI finds bugs faster than humans can write them, or patch them. The only honest response is to close the gap between discovery and deployment, aggressively and urgently. Companies that treat patching as optional are betting their customers’ data on a streak of luck. That streak always ends.

Lede Photo by Nishal Pavithran from Unsplash

Let me be direct: if your developers aren’t using AI-assisted coding in 2026, they’re already behind. That’s not a prediction, it’s the current state. The question isn’t whether to use it. It’s whether you’re doing it with any discipline. Most aren’t.

The default pattern I see is developers treating the AI like an oracle. Ask it something, paste the output, move on. No guardrails, no review, no structure. That works fine until it doesn’t. And when it doesn’t, it tends to fail badly.

In July 2025, a Replit AI agent tasked with building a software application instead deleted an entire company’s production database (link), wiping out months of work in seconds. The AI later admitted it “panicked instead of thinking” after ignoring an explicit instruction to stop. Google’s Gemini coding assistant admitted to a “complete and catastrophic” failure after it erroneously deleted user data while chasing non-existent bugs, with recovery proving futile in some instances. These aren’t edge cases. They’re what unreviewed AI usage looks like at scale.

The stats reinforce the point. CodeRabbit’s analysis of 470 real-world GitHub pull requests found that AI-generated code produces roughly 1.7x more issues than human-written code, with logic and correctness errors 75% more common in AI-generated PRs. Uplevel’s study of 800 developers found a 41% increase in bug rates for teams with GitHub Copilot access.

So how should teams actually work?

Treat the AI like an intern. A capable one, but an intern. You wouldn’t hand a new hire access to production and say “build what you think is right.” You’d tell them what to build, explain the constraints, define the standards and then review their work before it ships. The same logic applies here.

Concretely, that means a few things.

Define the rules your AI assistant follows. Commit them to your repo in the relevant instruction files: CLAUDE.md, agent.md, .cursorrules, whatever your tooling uses. These aren’t optional extras. They’re how you stop the AI from making decisions it shouldn’t make.

Give it architectural context. If your AI doesn’t know your layering conventions, your service boundaries or your data access patterns, it’ll invent its own. Usually badly. This context belongs in your instruction files too, not just in someone’s head or a Confluence page nobody reads.

Take security seriously in those files. A 2025 study found that 45% of AI-generated code contains vulnerabilities, with Java implementations showing failure rates over 70%. The AI won’t apply your auth patterns, input validation rules or secrets management conventions unless you tell it to. Explicitly. Generic output is what you get when you provide generic context. Your security rules should sit in your instruction files alongside your architectural ones, not as an afterthought.

Understand the context window. AI assistants only see what you show them. They have no memory of your codebase beyond what’s in the current session or your committed instruction files. If you’re not loading meaningful context, you get output that looks plausible but doesn’t fit your system. The quality of what comes back is almost entirely a function of what you put in. This is why the instruction files matter so much: they’re the persistent layer of context the AI carries into every session. Compact your context before a major task, or start a new session if you are going to start something unrelated.

Enforce a (Pull Request) PR-based workflow. The AI writes code, you review it and nothing merges without a human sign-off. This isn’t bureaucracy. It’s the minimum viable process for code that runs in production.

Be clear on ownership. When AI writes a function that ships a bug, accountability gets murky fast. Settle this now, before an incident forces the conversation. The developer who reviewed and approved the PR owns it. Full stop. That framing matters for culture, for post-incident reviews and for making sure developers actually read what they’re approving rather than rubber-stamping the AI’s output.

2025 was the year of AI coding speed. 2026 is going to be the year of AI coding quality. The teams that built fast without structure are now sitting on technical debt they can’t easily measure.

What’s your team’s process for AI-generated code? Do you have instruction files committed to your repos?

Lede image : AI Genereated on Ideogram.ai

So you alt-tab out of your editor, open Excel or a separate CSV viewer, do what you need, and jump back. It’s a small interruption — but it adds up fast.

I got tired of missing features in other CSV viewers in the VS Code Marketplace, so I created a solution.

Introducing CSV ClearView

CSV ClearView is a CSV editor that lives entirely inside VS Code. No more context-switching. No more hunting for a separate tool.

Here’s what it can do:

• Color-coded columns: scan your data visually at a glance
  SQL querying – filter and query your CSV like a database
•  Data profiling: understand your dataset’s shape and quality instantly
•  Duplicate detection: find and flag repeated rows
•  Freeze panes: keep headers in view while you scroll
•  Large file support: solid performance even on hefty files

If you live in VS Code and work with data, this is the extension I wish had existed a year ago.

Get CSV ClearView free on the VS Code Marketplace

The Story Behind It

I had never built a VS Code extension before. I didn’t know how the extension API worked, how to configure webpack for it, or how to publish to the marketplace. On my own, a project like this would have taken a few months of working over weekends, and that’s assuming I didn’t give up halfway through.

I built it in days — with AI.

I started by prototyping the core extension using Claude Code. Once I had something working, I moved to OpenCode — a terminal-based tool that lets you mix Claude, OpenAI Codex, Gemini and other LLMs depending on the task at hand. That flexibility turned out to be really powerful.

Here’s what AI handled across the project:

• Most of the TypeScript code
• Webpack configuration (a notorious time-sink)
• Security hardening
• A 119-test suite
• Publishing guidance for the VS Code Marketplace

I steered. I reviewed. I made the product and architecture decisions. But AI did a huge share of the execution.

What surprised me most wasn’t just the speed — it was the fact that I could write, build, debug, and publish the extension entirely from within OpenCode running in my terminal. The whole workflow stayed in one place.

What This Means for Developers

This experience changed how I think about side projects. The old mental model was: “I don’t know this technology well enough, so building this would take forever.” That barrier is gone — or at least dramatically lowered.

AI isn’t replacing the developer. It’s removing the friction between having an idea and shipping something real. You still need to know what good looks like. You still need to make judgment calls. But the grind of figuring out boilerplate, obscure configs, and publishing workflows? AI eats that for breakfast.

If there’s a tool you’ve always wished existed in your own workflow — maybe now’s the time to just build it.

Try It and Let Me Know What You Think

CSV ClearView is completely free.

Install it here or search for it in VSCode : VS Code Marketplace — CSV ClearView

Found a bug? Have a feature idea? Want to collaborate?
Open an issue or PR on GitHub: github.com/vinuthomas/CSV-Clear-View-for-VSCode

And I’d genuinely love to hear from you: what are you building with AI? Drop it in the comments below.

“Soar” – A Father’s Promise to His Daughters This song comes straight from my heart as a father. “Soar” is my musical letter to my daughters, letting them know that no matter where life takes them, I’ll always be there cheering them on from the sidelines.

The song captures that bittersweet moment every parent faces – watching your children grow up and spread their wings while wanting to hold onto those precious early years when you were their whole world. It’s about unconditional love, unwavering support, and the promise that even as they become their own people and chase their own dreams, they’ll never walk alone.

Whether they’re facing challenges, celebrating victories, or simply figuring out who they want to become, this is my way of telling them: “I believe in you, I’m proud of you, and I can’t wait to watch you soar.”

To my girls – and to every parent and child who hears this – may you always remember that love like this never fades, never wavers, and never ends.

Collaborators on this track

• Melissa Fernandes: Vocals
  Visit her Youtube channel at:  @melissafernandesmusic  
• Jeronimo Verdun: Music Production, Guitars

You can also add this song to your playlist on the following streaming services:

• Soundcloud
• Apple Music
• Deezer

The Promise of AI

AI holds immense potential. Machine learning models can analyze vast datasets at speeds incomprehensible to humans, allowing for advancements in medical diagnosis, financial forecasting, and countless other sectors. Automation promises efficiency, cost savings, and even the elimination of human error in some contexts.

However, along with these benefits come challenges that we cannot afford to overlook.

The Ethical Challenges

1. Bias and Discrimination: Machine learning models are only as good as the data they’re trained on. If this data contains biases (often reflecting society’s own prejudices), the AI can perpetuate or even exacerbate these biases. For instance, facial recognition software might inaccurately classify individuals of certain ethnic backgrounds if not trained on a diverse dataset.
2. Privacy Concerns: AI can analyze personal data, making it a powerful tool for both service personalization and invasive surveillance. Without proper regulations, there’s a risk of eroding individual privacy rights.
3. Job Displacement: Automation through AI can lead to job losses in certain sectors. While new jobs might emerge, there’s no guarantee they’ll be accessible to those displaced from their original roles.
4. Decision-making Accountability: If an AI-driven car crashes, who’s responsible? The manufacturer? The software developer? The car owner? Assigning accountability in a world driven by AI decisions becomes murky.

Striking a Balance: Recommendations for Ethical AI

1. Transparent Algorithms: Promote the development of AI systems that are transparent. If a machine makes a decision, we should be able to understand how it reached that conclusion.
2. Diverse Data: Ensure training data is diverse and representative. This will minimize the risk of AI models unintentionally discriminating against certain groups.
3. Privacy-first Approach: Adopt strict data protection regulations, ensuring that AI systems respect individual privacy rights.
4. Continuous Education: As AI alters the job landscape, invest in continuous education and training programs, ensuring that workers can adapt to the changing demands of the workforce.
5. Human-in-the-loop (HITL) Systems: Instead of letting AI make all decisions autonomously, use it as a tool to augment human decision-making. This approach can harness the strengths of both humans and AI while minimizing their respective weaknesses.
6. Ethical Oversight: Establish ethics boards or committees for AI development and deployment, ensuring that moral considerations aren’t an afterthought but an integral part of the AI development process.

Conclusion

AI is not just another technological advancement; it’s a transformative force that will redefine aspects of society, economy, and daily life. While the allure of progress is undeniable, our shared responsibility is to ensure that this progress doesn’t come at the expense of our values, rights, and shared humanity. By proactively addressing the ethical challenges AI presents, we can navigate toward a future where technology serves us, not the other way around.

Check Your Hardware

Before diving into software solutions, ensuring that your hardware is in order is essential.

• Router Status: Ensure the router is plugged in and powered on. Look for the indicator lights to confirm it’s working correctly.
• Cable Connections: Check for any loose cables or connections. A simple disconnection can sometimes be the culprit.
• Restart the Router: A simple restart can resolve many issues if you run your WiFi router 24/7. Turn off the router, wait a minute, and then turn it back on.

Verify Wi-Fi is Enabled:

It might sound basic, but sometimes the Wi-Fi might be turned off accidentally.

• On Computers: Check the Wi-Fi icon in the taskbar (Windows) or menu bar (Mac) to ensure it’s on.
• On Smartphones: Ensure Airplane mode is off and Wi-Fi is enabled in the settings.

Forget and Reconnect to the Network

Sometimes, the device might remember incorrect settings or passwords. On your phone or tablet,

• Navigate to the Wi-Fi settings on your device.
• “Forget” the network and then reconnect by entering the password again.

Check for Service Provider Outages

If your router shows a red status light against your ISP’s connection or multiple devices can’t connect, the issue might be with the service provider.

• Check the service provider’s website or social media channels for any outage notifications.
• Call your service provider’s helpline to check for service outages or to seek help.

Update Network Drivers (for computers)

Outdated drivers can cause connection issues.

• For Windows: Go to ‘Device Manager’ > ‘Network Adapters’ and update the Wi-Fi driver.
• For Mac: Ensure your OS is updated, as updates generally include driver updates.

Change Wi-Fi Channels

Other electronic devices or neighboring Wi-Fi networks can cause interference.

• Log into your router settings (usually through a web browser).
• Change the Wi-Fi channel to see if the connection improves.
• Make sure you connect to the 5GHz band on your Wi-Fi router, the 2.4 GHz band is much slower.

Reset Network Settings

This is a more drastic step but can resolve persistent issues.

• On smartphones, look for “Reset Network Settings” in the settings. Note: This will also forget all saved Wi-Fi networks and Bluetooth devices.
• On computers, you might need to consult the device’s help guide or seek expert assistance.

Consider External Factors

• Physical Obstructions: Walls, floors, and doors can weaken Wi-Fi signals.
• Electronic Interference: Devices like microwaves and cordless phones can interfere with Wi-Fi.
• Distance from Router: The further you are from the router, the weaker the signal. Consider using Wi-Fi extenders for large homes.

Conclusion

While Wi-Fi issues can be frustrating, most problems have straightforward solutions. Following this guide, you can troubleshoot and resolve common Wi-Fi connection issues. If problems persist, consider contacting your service provider.

Many people have shown interest in other social networking sites since Elon Musk’s takeover of Twitter, and Mastodon is one of them. Surprisingly, the site has grown to 9 million users with the growth fuelled by the Twitter fiascos of late. While Mastodon may look similar to Twitter, it operates differently in some critical areas.

The article will explore what Mastodon is and why a lot of people are interested in the platform.

What is Mastodon?

Mastodon is a free, open-source social media platform that is made up of independent servers. Since it is open source, anyone can set up a server, make customizations, and run a community similar to an online forum or discussion board.

People can join the platform, follow and interact with each other, and do all sorts of things they’d usually do on a social media platform like Twitter. While Mastodon’s interface does have similarities with that of Twitter’s web and mobile versions, both platforms are not the same.

On Mastodon, you’d toot instead of tweet, and if you share someone’s post, you’ve reblogged or boosted it. The platform has hashtags and lists and provides you with an opportunity to post up to 500 words to express your mind, plus the capability to add GIFs, images, and videos.

How does Mastodon Work?

Mastodon’s working is where it differs from Twitter and other social media sites. Unlike Twitter, the platform is not owned or operated by a single company or corporation, where the company is responsible for building in-app features, moderating content, tweaking algorithms, and handling all the tasks of running a social network.

Even the non-profit company, known as Mastodon gGmbH, that develops the Mastodon software can’t own the platform, as it is made up of several independent servers. When someone creates their version of Mastodon, this is called an instance. Due to Mastodon being open-source, anyone can make this instance and set their own rules.  All instances, also known as servers, are owned and operated by the community that creates them.

When you create an account on Mastodon, you do it on a specific server.

• You can see the public timeline of any server without an account.
• An account is needed to post.
• With an account, you can follow users matching your own server or any other server.
• You can reply to people and boost posts of people on other servers.
• Mastodon allows you to create as many accounts on different servers.
• Some servers may be invitation-only or sign-up-restricted as per the rules set by their administrator.

How to Choose a Server?

Choosing a server isn’t a critical task, as the platform allows you to follow or interact with other people on different servers as long as the server you choose hasn’t blocked their server or their server hasn’t blocked yours.

Here are a few things you should consider before choosing a server.

• What posts you’d like to see on your local timeline?
• The server’s rules or moderation policies.
• Are there any other servers your server has blocked? If yes, then which ones?
• You can use the debirdify app to discover where the people you follow on Twitter are gathered on Mastodon.
• You can find the server you want to join on mastodonservers.net.

Source

Can You Switch Servers?

Yes, you can. However, the platform doesn’t allow you to shift your posts on the new server. You can shift your account, and it will issue a redirect to your followers to move them to your new profile.

How to Create an Account on Mastodon?

Creating an account on Mastodon isn’t difficult. Follow these steps.

• Go to the website joinmastodon.org and click create account.
• Choose a server of your choice and again click create account.
• Once done, read the rules of the server you’re about to join and hit accept.
• After that, fill up the signup form with your details and click sign up.
• Confirm your email address, and you’re ready to use the platform.

How can Businesses Use Mastodon?

If you’re a business and looking to shift to Mastodon to use it for initiatives you’ve been running on your Twitter handle, be ready to get disappointed. Since the platform is a network of servers, businesses won’t be able to do marketing.

Furthermore, its homepage states that it will never serve ads or push a profile over others. Because a non-profit company runs Mastodon, there is no special privileges or stats available for advertisers. However, businesses can use the platform with a different perspective and benefit in the following ways.

• Market Research: Mastodon can help you do better market research. Unlike Twitter, where you search hashtags and receive an uncategorized mess of tweets, Mastodon enables you to find the right server matching your brand and its target audience.
• Affiliate Marketing: You can also run sponsorships and affiliate marketing campaigns on Mastodon by reaching out to individual accounts with a particular count of followers. However,  you may struggle to find influencers on Mastodon since the number of followers is generally less than Twitter followers.
• Build Your Community: The best part, you can build your own server and run a community according to your business while having complete control over the server with your own set of rules and restrictions.

Safety Concerns Regarding Mastodon

Since Mastodon is decentralized, many security concerns arise around the platform. Recent research has found several security vulnerabilities in Mastodon’s social media network.

Each Mastodon instance is managed by an administrator who has control over its infrastructure and users’ information. It means people put their trust in administrators to secure and maintain their instance.

Mastodon doesn’t have a centralized security team or safety team, since all the instances are decentralized. On Mastodon, you are responsible for your security and privacy. That being said, you don’t really have to stop using the platform. However, as with all social media platforms, be wary of what you share on Mastodon.

Takeaway

While the platform has become quite popular in less time, we don’t think it will be able to replace Twitter. It’s easier to say than done because the users who claimed to be leaving Twitter haven’t really left the space. So, it won’t be wrong to say that Mastodon isn’t replacing Twitter anytime soon.

If you are looking for a new platform to grow your social identity or escape from Twitterverse,  jump onto Mastodon now.  Here’s a link to my Mastodon account.