Calmness is a vital life skill.

It aids communication. It boosts creativity and performance.

I've studied calm people's habits for years. Commonalities:

Have mastered the art of self-humor.

Protectors take their job seriously, draining the room's energy.

They are fixated on positive pursuits like making cool things, building a strong physique, and having fun with others rather than on depressing influences like the news and gossip.

Every day, spend at least 20 minutes moving, whether it's walking, yoga, or lifting weights.

Discover ways to take pleasure in life's challenges.

Since perspective is malleable, they change their view.

Set your own needs first.

Stressed people neglect themselves and wonder why they struggle.

Prioritize self-care.

Don't ruin your life to please others.

Make something.

Calm people create more than react.

They love creating beautiful things—paintings, children, relationships, and projects.

Don’t hold their breath.

If you're stressed or angry, you may be surprised how much time you spend holding your breath and tightening your belly.

Release, breathe, and relax to find calm.

Stopped rushing.

Rushing is disadvantageous.

Calm people handle life better.

Are aware of their own dietary requirements.

They avoid junk food and eat foods that keep them healthy, happy, and calm.

Don’t take anything personally.

Stressed people control everything.

Self-conscious.

Calm people put others and their work first.

Keep their surroundings neat.

Maintaining an uplifting and clutter-free environment daily calms the mind.

Minimise negative people.

Calm people are ruthless with their boundaries and avoid negative and drama-prone people.

Learn to identify them, make habits to avoid them

First, a clarification. This article is aimed to make new programmers aware of their mistakes, train them to detect them, and remind them to prevent them.

I learned from all these blunders. I'm glad I have coding habits to avoid them. Do too.

These mistakes are not ordered.

1) Writing code haphazardly

Writing good content is hard. It takes planning and investigation. Quality programs don't differ.

Think. Research. Plan. Write. Validate. Modify. Unfortunately, no good acronym exists. Create a habit of doing the proper quantity of these activities.

As a newbie programmer, my biggest error was writing code without thinking or researching. This works for small stand-alone apps but hurts larger ones.

Like saying anything you might regret, you should think before coding something you could regret. Coding expresses your thoughts.

When angry, count to 10 before you speak. If very angry, a hundred. — Thomas Jefferson.

My quote:

When reviewing code, count to 10 before you refactor a line. If the code does not have tests, a hundred. — Samer Buna

Programming is primarily about reviewing prior code, investigating what is needed and how it fits into the current system, and developing small, testable features. Only 10% of the process involves writing code.

Programming is not writing code. Programming need nurturing.

2) Making excessive plans prior to writing code

Yes. Planning before writing code is good, but too much of it is bad. Water poisons.

Avoid perfect plans. Programming does not have that. Find a good starting plan. Your plan will change, but it helped you structure your code for clarity. Overplanning wastes time.

Only planning small features. All-feature planning should be illegal! The Waterfall Approach is a step-by-step system. That strategy requires extensive planning. This is not planning. Most software projects fail with waterfall. Implementing anything sophisticated requires agile changes to reality.

Programming requires responsiveness. You'll add waterfall plan-unthinkable features. You will eliminate functionality for reasons you never considered in a waterfall plan. Fix bugs and adjust. Be agile.

Plan your future features, though. Do it cautiously since too little or too much planning can affect code quality, which you must risk.

3) Underestimating the Value of Good Code

Readability should be your code's exclusive goal. Unintelligible code stinks. Non-recyclable.

Never undervalue code quality. Coding communicates implementations. Coders must explicitly communicate solution implementations.

Programming quote I like:

Always code as if the guy who ends up maintaining your code will be a violent psychopath who knows where you live. — John Woods

John, great advice!

Small things matter. If your indentation and capitalization are inconsistent, you should lose your coding license.

Long queues are also simple. Readability decreases after 80 characters. To highlight an if-statement block, you might put a long condition on the same line. No. Just never exceed 80 characters.

Linting and formatting tools fix many basic issues like this. ESLint and Prettier work great together in JavaScript. Use them.

Code quality errors:

Multiple lines in a function or file. Break long code into manageable bits. My rule of thumb is that any function with more than 10 lines is excessively long.

Double-negatives. Don't.

Using double negatives is just very not not wrong

Short, generic, or type-based variable names. Name variables clearly.

There are only two hard things in Computer Science: cache invalidation and naming things. — Phil Karlton

Hard-coding primitive strings and numbers without descriptions. If your logic relies on a constant primitive string or numeric value, identify it.

Avoiding simple difficulties with sloppy shortcuts and workarounds. Avoid evasion. Take stock.

Considering lengthier code better. Shorter code is usually preferable. Only write lengthier versions if they improve code readability. For instance, don't utilize clever one-liners and nested ternary statements just to make the code shorter. In any application, removing unneeded code is better.

Measuring programming progress by lines of code is like measuring aircraft building progress by weight. — Bill Gates

Excessive conditional logic. Conditional logic is unnecessary for most tasks. Choose based on readability. Measure performance before optimizing. Avoid Yoda conditions and conditional assignments.

4) Selecting the First Approach

When I started programming, I would solve an issue and move on. I would apply my initial solution without considering its intricacies and probable shortcomings.

After questioning all the solutions, the best ones usually emerge. If you can't think of several answers, you don't grasp the problem.

Programmers do not solve problems. Find the easiest solution. The solution must work well and be easy to read, comprehend, and maintain.

There are two ways of constructing a software design. One way is to make it so simple that there are obviously no deficiencies, and the other way is to make it so complicated that there are no obvious deficiencies. — C.A.R. Hoare

5) Not Giving Up

I generally stick with the original solution even though it may not be the best. The not-quitting mentality may explain this. This mindset is helpful for most things, but not programming. Program writers should fail early and often.

If you doubt a solution, toss it and rethink the situation. No matter how much you put in that solution. GIT lets you branch off and try various solutions. Use it.

Do not be attached to code because of how much effort you put into it. Bad code needs to be discarded.

6) Avoiding Google

I've wasted time solving problems when I should have researched them first.

Unless you're employing cutting-edge technology, someone else has probably solved your problem. Google It First.

Googling may discover that what you think is an issue isn't and that you should embrace it. Do not presume you know everything needed to choose a solution. Google surprises.

But Google carefully. Newbies also copy code without knowing it. Use only code you understand, even if it solves your problem.

Never assume you know how to code creatively.

The most dangerous thought that you can have as a creative person is to think that you know what you’re doing. — Bret Victor

7) Failing to Use Encapsulation

Not about object-oriented paradigm. Encapsulation is always useful. Unencapsulated systems are difficult to maintain.

An application should only handle a feature once. One object handles that. The application's other objects should only see what's essential. Reducing application dependencies is not about secrecy. Following these guidelines lets you safely update class, object, and function internals without breaking things.

Classify logic and state concepts. Class means blueprint template. Class or Function objects are possible. It could be a Module or Package.

Self-contained tasks need methods in a logic class. Methods should accomplish one thing well. Similar classes should share method names.

As a rookie programmer, I didn't always establish a new class for a conceptual unit or recognize self-contained units. Newbie code has a Util class full of unrelated code. Another symptom of novice code is when a small change cascades and requires numerous other adjustments.

Think before adding a method or new responsibilities to a method. Time's needed. Avoid skipping or refactoring. Start right.

High Cohesion and Low Coupling involves grouping relevant code in a class and reducing class dependencies.

8) Arranging for Uncertainty

Thinking beyond your solution is appealing. Every line of code will bring up what-ifs. This is excellent for edge cases but not for foreseeable needs.

Your what-ifs must fall into one of these two categories. Write only code you need today. Avoid future planning.

Writing a feature for future use is improper. No.

Write only the code you need today for your solution. Handle edge-cases, but don't introduce edge-features.

Growth for the sake of growth is the ideology of the cancer cell. — Edward Abbey

9) Making the incorrect data structure choices

Beginner programmers often overemphasize algorithms when preparing for interviews. Good algorithms should be identified and used when needed, but memorizing them won't make you a programming genius.

However, learning your language's data structures' strengths and shortcomings will make you a better developer.

The improper data structure shouts "newbie coding" here.

Let me give you a few instances of data structures without teaching you:

Managing records with arrays instead of maps (objects).

Most data structure mistakes include using lists instead of maps to manage records. Use a map to organize a list of records.

This list of records has an identifier to look up each entry. Lists for scalar values are OK and frequently superior, especially if the focus is pushing values to the list.

Arrays and objects are the most common JavaScript list and map structures, respectively (there is also a map structure in modern JavaScript).

Lists over maps for record management often fail. I recommend always using this point, even though it only applies to huge collections. This is crucial because maps are faster than lists in looking up records by identifier.

Stackless

Simple recursive functions are often tempting when writing recursive programming. In single-threaded settings, optimizing recursive code is difficult.

Recursive function returns determine code optimization. Optimizing a recursive function that returns two or more calls to itself is harder than optimizing a single call.

Beginners overlook the alternative to recursive functions. Use Stack. Push function calls to a stack and start popping them out to traverse them back.

10) Worsening the current code

Imagine this:

Add an item to that room. You might want to store that object anywhere as it's a mess. You can finish in seconds.

Not with messy code. Do not worsen! Keep the code cleaner than when you started.

Clean the room above to place the new object. If the item is clothing, clear a route to the closet. That's proper execution.

The following bad habits frequently make code worse:

• code duplication You are merely duplicating code and creating more chaos if you copy/paste a code block and then alter just the line after that. This would be equivalent to adding another chair with a lower base rather than purchasing a new chair with a height-adjustable seat in the context of the aforementioned dirty room example. Always keep abstraction in mind, and use it when appropriate.

• utilizing configuration files not at all. A configuration file should contain the value you need to utilize if it may differ in certain circumstances or at different times. A configuration file should contain a value if you need to use it across numerous lines of code. Every time you add a new value to the code, simply ask yourself: "Does this value belong in a configuration file?" The most likely response is "yes."

• using temporary variables and pointless conditional statements. Every if-statement represents a logic branch that should at the very least be tested twice. When avoiding conditionals doesn't compromise readability, it should be done. The main issue with this is that branch logic is being used to extend an existing function rather than creating a new function. Are you altering the code at the appropriate level, or should you go think about the issue at a higher level every time you feel you need an if-statement or a new function variable?

This code illustrates superfluous if-statements:

function isOdd(number) {
  if (number % 2 === 1) {
    return true;
  } else {
    return false;
  }
}

Can you spot the biggest issue with the isOdd function above?

Unnecessary if-statement. Similar code:

function isOdd(number) {
  return (number % 2 === 1);
};

11) Making remarks on things that are obvious

I've learnt to avoid comments. Most code comments can be renamed.

instead of:

// This function sums only odd numbers in an array
const sum = (val) => {
  return val.reduce((a, b) => {
    if (b % 2 === 1) { // If the current number is odd
      a+=b;            // Add current number to accumulator
    }
    return a;          // The accumulator
  }, 0);
};

Commentless code looks like this:

const sumOddValues = (array) => {
  return array.reduce((accumulator, currentNumber) => {
    if (isOdd(currentNumber)) { 
      return accumulator + currentNumber;
    }
    return accumulator;
  }, 0);
};

Better function and argument names eliminate most comments. Remember that before commenting.

Sometimes you have to use comments to clarify the code. This is when your comments should answer WHY this code rather than WHAT it does.

Do not write a WHAT remark to clarify the code. Here are some unnecessary comments that clutter code:

// create a variable and initialize it to 0
let sum = 0;
// Loop over array
array.forEach(
  // For each number in the array
  (number) => {
    // Add the current number to the sum variable
    sum += number;
  }
);

Avoid that programmer. Reject that code. Remove such comments if necessary. Most importantly, teach programmers how awful these remarks are. Tell programmers who publish remarks like this that they may lose their jobs. That terrible.

12) Skipping tests

I'll simplify. If you develop code without tests because you think you're an excellent programmer, you're a rookie.

If you're not writing tests in code, you're probably testing manually. Every few lines of code in a web application will be refreshed and interacted with. Also. Manual code testing is fine. To learn how to automatically test your code, manually test it. After testing your application, return to your code editor and write code to automatically perform the same interaction the next time you add code.

Human. After each code update, you will forget to test all successful validations. Automate it!

Before writing code to fulfill validations, guess or design them. TDD is real. It improves your feature design thinking.

If you can use TDD, even partially, do so.

13) Making the assumption that if something is working, it must be right.

See this sumOddValues function. Is it flawed?

const sumOddValues = (array) => {
  return array.reduce((accumulator, currentNumber) => {
    if (currentNumber % 2 === 1) { 
      return accumulator + currentNumber;
    }
    return accumulator;
  });
};
 
 
console.assert(
  sumOddValues([1, 2, 3, 4, 5]) === 9
);

Verified. Good life. Correct?

Code above is incomplete. It handles some scenarios correctly, including the assumption used, but it has many other issues. I'll list some:

#1: No empty input handling. What happens when the function is called without arguments? That results in an error revealing the function's implementation:

TypeError: Cannot read property 'reduce' of undefined.

Two main factors indicate faulty code.

• Your function's users shouldn't come across implementation-related information.

• The user cannot benefit from the error. Simply said, they were unable to use your function. They would be aware that they misused the function if the error was more obvious about the usage issue. You might decide to make the function throw a custom exception, for instance:

TypeError: Cannot execute function for empty list.

Instead of returning an error, your method should disregard empty input and return a sum of 0. This case requires action.

Problem #2: No input validation. What happens if the function is invoked with a text, integer, or object instead of an array?

The function now throws:

sumOddValues(42);
TypeError: array.reduce is not a function

Unfortunately, array. cut's a function!

The function labels anything you call it with (42 in the example above) as array because we named the argument array. The error says 42.reduce is not a function.

See how that error confuses? An mistake like:

TypeError: 42 is not an array, dude.

Edge-cases are #1 and #2. These edge-cases are typical, but you should also consider less obvious ones. Negative numbers—what happens?

sumOddValues([1, 2, 3, 4, 5, -13]) // => still 9

-13's unusual. Is this the desired function behavior? Error? Should it sum negative numbers? Should it keep ignoring negative numbers? You may notice the function should have been titled sumPositiveOddNumbers.

This decision is simple. The more essential point is that if you don't write a test case to document your decision, future function maintainers won't know if you ignored negative values intentionally or accidentally.

It’s not a bug. It’s a feature. — Someone who forgot a test case

#3: Valid cases are not tested. Forget edge-cases, this function mishandles a straightforward case:

sumOddValues([2, 1, 3, 4, 5]) // => 11

The 2 above was wrongly included in sum.

The solution is simple: reduce accepts a second input to initialize the accumulator. Reduce will use the first value in the collection as the accumulator if that argument is not provided, like in the code above. The sum included the test case's first even value.

This test case should have been included in the tests along with many others, such as all-even numbers, a list with 0 in it, and an empty list.

Newbie code also has rudimentary tests that disregard edge-cases.

14) Adhering to Current Law

Unless you're a lone supercoder, you'll encounter stupid code. Beginners don't identify it and assume it's decent code because it works and has been in the codebase for a while.

Worse, if the terrible code uses bad practices, the newbie may be enticed to use them elsewhere in the codebase since they learnt them from good code.

A unique condition may have pushed the developer to write faulty code. This is a nice spot for a thorough note that informs newbies about that condition and why the code is written that way.

Beginners should presume that undocumented code they don't understand is bad. Ask. Enquire. Blame it!

If the code's author is dead or can't remember it, research and understand it. Only after understanding the code can you judge its quality. Before that, presume nothing.

15) Being fixated on best practices

Best practices damage. It suggests no further research. Best practice ever. No doubts!

No best practices. Today's programming language may have good practices.

Programming best practices are now considered bad practices.

Time will reveal better methods. Focus on your strengths, not best practices.

Do not do anything because you read a quote, saw someone else do it, or heard it is a recommended practice. This contains all my article advice! Ask questions, challenge theories, know your options, and make informed decisions.

16) Being preoccupied with performance

Premature optimization is the root of all evil (or at least most of it) in programming — Donald Knuth (1974)

I think Donald Knuth's advice is still relevant today, even though programming has changed.

Do not optimize code if you cannot measure the suspected performance problem.

Optimizing before code execution is likely premature. You may possibly be wasting time optimizing.

There are obvious optimizations to consider when writing new code. You must not flood the event loop or block the call stack in Node.js. Remember this early optimization. Will this code block the call stack?

Avoid non-obvious code optimization without measurements. If done, your performance boost may cause new issues.

Stop optimizing unmeasured performance issues.

17) Missing the End-User Experience as a Goal

How can an app add a feature easily? Look at it from your perspective or in the existing User Interface. Right? Add it to the form if the feature captures user input. Add it to your nested menu of links if it adds a link to a page.

Avoid that developer. Be a professional who empathizes with customers. They imagine this feature's consumers' needs and behavior. They focus on making the feature easy to find and use, not just adding it to the software.

18) Choosing the incorrect tool for the task

Every programmer has their preferred tools. Most tools are good for one thing and bad for others.

The worst tool for screwing in a screw is a hammer. Do not use your favorite hammer on a screw. Don't use Amazon's most popular hammer on a screw.

A true beginner relies on tool popularity rather than problem fit.

You may not know the best tools for a project. You may know the best tool. However, it wouldn't rank high. You must learn your tools and be open to new ones.

Some coders shun new tools. They like their tools and don't want to learn new ones. I can relate, but it's wrong.

You can build a house slowly with basic tools or rapidly with superior tools. You must learn and use new tools.

19) Failing to recognize that data issues are caused by code issues

Programs commonly manage data. The software will add, delete, and change records.

Even the simplest programming errors can make data unpredictable. Especially if the same defective application validates all data.

Code-data relationships may be confusing for beginners. They may employ broken code in production since feature X is not critical. Buggy coding may cause hidden data integrity issues.

Worse, deploying code that corrected flaws without fixing minor data problems caused by these defects will only collect more data problems that take the situation into the unrecoverable-level category.

How do you avoid these issues? Simply employ numerous data integrity validation levels. Use several interfaces. Front-end, back-end, network, and database validations. If not, apply database constraints.

Use all database constraints when adding columns and tables:

• If a column has a NOT NULL constraint, null values will be rejected for that column. If your application expects that field has a value, your database should designate its source as not null.

• If a column has a UNIQUE constraint, the entire table cannot include duplicate values for that column. This is ideal for a username or email field on a Users table, for instance.

• For the data to be accepted, a CHECK constraint, or custom expression, must evaluate to true. For instance, you can apply a check constraint to ensure that the values of a normal % column must fall within the range of 0 and 100.

• With a PRIMARY KEY constraint, the values of the columns must be both distinct and not null. This one is presumably what you're utilizing. To distinguish the records in each table, the database needs have a primary key.

• A FOREIGN KEY constraint requires that the values in one database column, typically a primary key, match those in another table column.

Transaction apathy is another data integrity issue for newbies. If numerous actions affect the same data source and depend on each other, they must be wrapped in a transaction that can be rolled back if one fails.

20) Reinventing the Wheel

Tricky. Some programming wheels need reinvention. Programming is undefined. New requirements and changes happen faster than any team can handle.

Instead of modifying the wheel we all adore, maybe we should rethink it if you need a wheel that spins at varied speeds depending on the time of day. If you don't require a non-standard wheel, don't reinvent it. Use the darn wheel.

Wheel brands can be hard to choose from. Research and test before buying! Most software wheels are free and transparent. Internal design quality lets you evaluate coding wheels. Try open-source wheels. Debug and fix open-source software simply. They're easily replaceable. In-house support is also easy.

If you need a wheel, don't buy a new automobile and put your maintained car on top. Do not include a library to use a few functions. Lodash in JavaScript is the finest example. Import shuffle to shuffle an array. Don't import lodash.

21) Adopting the incorrect perspective on code reviews

Beginners often see code reviews as criticism. Dislike them. Not appreciated. Even fear them.

Incorrect. If so, modify your mindset immediately. Learn from every code review. Salute them. Observe. Most crucial, thank reviewers who teach you.

Always learning code. Accept it. Most code reviews teach something new. Use these for learning.

You may need to correct the reviewer. If your code didn't make that evident, it may need to be changed. If you must teach your reviewer, remember that teaching is one of the most enjoyable things a programmer can do.

22) Not Using Source Control

Newbies often underestimate Git's capabilities.

Source control is more than sharing your modifications. It's much bigger. Clear history is source control. The history of coding will assist address complex problems. Commit messages matter. They are another way to communicate your implementations, and utilizing them with modest commits helps future maintainers understand how the code got where it is.

Commit early and often with present-tense verbs. Summarize your messages but be detailed. If you need more than a few lines, your commit is too long. Rebase!

Avoid needless commit messages. Commit summaries should not list new, changed, or deleted files. Git commands can display that list from the commit object. The summary message would be noise. I think a big commit has many summaries per file altered.

Source control involves discoverability. You can discover the commit that introduced a function and see its context if you doubt its need or design. Commits can even pinpoint which code caused a bug. Git has a binary search within commits (bisect) to find the bug-causing commit.

Source control can be used before commits to great effect. Staging changes, patching selectively, resetting, stashing, editing, applying, diffing, reversing, and others enrich your coding flow. Know, use, and enjoy them.

I consider a Git rookie someone who knows less functionalities.

23) Excessive Use of Shared State

Again, this is not about functional programming vs. other paradigms. That's another article.

Shared state is problematic and should be avoided if feasible. If not, use shared state as little as possible.

As a new programmer, I didn't know that all variables represent shared states. All variables in the same scope can change its data. Global scope reduces shared state span. Keep new states in limited scopes and avoid upward leakage.

When numerous resources modify common state in the same event loop tick, the situation becomes severe (in event-loop-based environments). Races happen.

This shared state race condition problem may encourage a rookie to utilize a timer, especially if they have a data lock issue. Red flag. No. Never accept it.

24) Adopting the Wrong Mentality Toward Errors

Errors are good. Progress. They indicate a simple way to improve.

Expert programmers enjoy errors. Newbies detest them.

If these lovely red error warnings irritate you, modify your mindset. Consider them helpers. Handle them. Use them to advance.

Some errors need exceptions. Plan for user-defined exceptions. Ignore some mistakes. Crash and exit the app.

25) Ignoring rest periods

Humans require mental breaks. Take breaks. In the zone, you'll forget breaks. Another symptom of beginners. No compromises. Make breaks mandatory in your process. Take frequent pauses. Take a little walk to plan your next move. Reread the code.

This has been a long post. You deserve a break.

If I could tell my younger self a few rules, they would boost my career.

1. Treat design like medicine; don't get attached.

If it doesn't help, you won't be angry, but you'll try to improve it. Designers blame others if they don't like the design, but the rule is the same: we solve users' problems. You're not your design, and neither are they. Be humble with your work because your assumptions will often be wrong and users will behave differently.

2. Consider your design flawed.

Disagree with yourself, then defend your ideas. Most designers forget to dig deeper into a pattern, screen, button, or copywriting. If someone asked, "Have you considered alternatives? How does this design stack up? Here's a functional UX checklist to help you make design decisions.

3. Codeable solutions.

If your design requires more developer time, consider whether it's worth spending more money to code something with a small UX impact. Overthinking problems and designing abstract patterns is easy. Sometimes you see something on dribbble or bechance and try to recreate it, but it's not worth it. Here's my article on it.

4. Communication changes careers

Designers often talk with users, clients, companies, developers, and other designers. How you talk and present yourself can land you a job. Like driving or swimming, practice it. Success requires being outgoing and friendly. If I hadn't said "hello" to a few people, I wouldn't be where I am now.

5. Ignorance of the law is not an excuse.

Copyright, taxation How often have you used an icon without checking its license? If you use someone else's work in your project, the owner can cause you a lot of problems — paying a lot of money isn't worth it. Spend a few hours reading about copyrights, client agreements, and taxes.

6. Always test your design

If nobody has seen or used my design, it's not finished. Ask friends about prototypes. Testing reveals how wrong your assumptions were. Steve Krug, one of the authorities on this topic will tell you more about how to do testing.

7. Run workshops

A UX designer's job involves talking to people and figuring out what they need, which is difficult because they usually don't know. Organizing teamwork sessions is a powerful skill, but you must also be a good listener. Your job is to help a quiet, introverted developer express his solution and control the group. AJ Smart has more on workshops here.

Cryptoland rollercoaster

I've lost money in crypto.

Unimportant.

The real issue: I didn’t understand how.

I'm surrounded with winners. To learn more, I created my own NFTs, currency, and DAO.

Web3 is a hilltop castle. Everything is valuable, decentralized, and on-chain.

The castle is Disneyland: beautiful in images, but chaotic with lengthy lines and kids spending too much money on dressed-up animals.

When the throng and businesses are gone, Disneyland still has enchantment.

The Real Story of Web3

NFTs

Scarcity. Scarce NFTs. That's their worth.

Skull. Rare-looking!

Nonsense.

Bored Ape Yacht Club vs. my NFTs?

Marketing.

BAYC is amazing, but not for the reasons people believe. Apecoin and Otherside's art, celebrity following, and innovation? Stunning.

No other endeavor captured the zeitgeist better. Yet how long did you think it took to actually mint the NFTs?

1 hour? Maybe a week for the website?

Minting NFTs is incredibly easy. Kid-friendly. Developers are rare. Think about that next time somebody posts “DevS dO SMt!?”

NFTs will remain popular. These projects are like our Van Goghs and Monets. Still, be wary. It still uses exclusivity and wash selling like the OG art market.

Not all NFTs are art-related.

Soulbound and anonymous NFTs could offer up new use cases. Property rights, privacy-focused ID, open-source project verification. Everything.

NFTs build online trust through ownership.

We just need to evolve from the apes first.

NFTs' superpower is marketing until then.

Crypto currency

What the hell is a token?

99% of people are clueless.

So I invested in both coins and tokens. Same same. Only that they are not.

Coins have their own blockchain and developer/validator community. It's hard.

Creating a token on top of a blockchain? Five minutes.

Most consumers don’t understand the difference, creating an arbitrage opportunity: pretend you’re a serious project without having developers on your payroll.

Few market sites help. Take a look. See any tokens?

There's a hint one click deeper.

Some tokens are legitimate. Some coins are bad investments.

Tokens are utilized for DAO governance and DApp payments. Still, know who's behind a token. They might be 12 years old.

Coins take time and money. The recent LUNA meltdown indicates that currency investing requires research.

DAOs

Decentralized Autonomous Organizations (DAOs) don't work as you assume.

Yes, members can vote.

A productive organization requires more.

I've observed two types of DAOs.

• Total decentralization total dysfunction

• Centralized just partially. Community-driven.

A core team executes the DAO's strategy and roadmap in successful DAOs. The community owns part of the organization, votes on decisions, and holds the team accountable.

DAOs are public companies.

Amazing.

A shareholder meeting's logistics are staggering. DAOs may hold anonymous, secure voting quickly. No need for intermediaries like banks to chase up every shareholder.

Successful DAOs aren't totally decentralized. Large-scale voting and collaboration have never been easier.

And that’s all that matters.

Scale, speed.

My Web3 learnings

Disneyland is enchanting. Web3 too.

In a few cycles, NFTs may be used to build trust, not clout. Not speculating with coins. DAOs run organizations, not themselves.

Finally, some final thoughts:

• NFTs will be a very helpful tool for building trust online. NFTs are successful now because of excellent marketing.

• Tokens are not the same as coins. Look into any project before making a purchase. Make sure it isn't run by three 9-year-olds piled on top of one another in a trench coat, at the very least.

• Not entirely decentralized, DAOs. We shall see a future where community ownership becomes the rule rather than the exception once we acknowledge this fact.

Crypto Disneyland is a rollercoaster with loops that make you sick.

Always buckle up.

Have fun!

Deep learning-based generative art is being researched. As usual, self-learning is better. Some models, like OpenAI's DALL-E 2, require registration and can only be used online, but others can be used locally, which is usually more enjoyable for curious users. I'll demonstrate the Stable Diffusion model's operation on a standard PC.

Let’s get started.

What It Does

Stable Diffusion uses numerous components:

• A generative model trained to produce images is called a diffusion model. The model is incrementally improving the starting data, which is only random noise. The model has an image, and while it is being trained, the reversed process is being used to add noise to the image. Being able to reverse this procedure and create images from noise is where the true magic is (more details and samples can be found in the paper).

• An internal compressed representation of a latent diffusion model, which may be altered to produce the desired images, is used (more details can be found in the paper). The capacity to fine-tune the generation process is essential because producing pictures at random is not very attractive (as we can see, for instance, in Generative Adversarial Networks).

• A neural network model called CLIP (Contrastive Language-Image Pre-training) is used to translate natural language prompts into vector representations. This model, which was trained on 400,000,000 image-text pairs, enables the transformation of a text prompt into a latent space for the diffusion model in the scenario of stable diffusion (more details in that paper).

This figure shows all data flow:

The weights file size for Stable Diffusion model v1 is 4 GB and v2 is 5 GB, making the model quite huge. The v1 model was trained on 256x256 and 512x512 LAION-5B pictures on a 4,000 GPU cluster using over 150.000 NVIDIA A100 GPU hours. The open-source pre-trained model is helpful for us. And we will.

Install

Before utilizing the Python sources for Stable Diffusion v1 on GitHub, we must install Miniconda (assuming Git and Python are already installed):

wget https://repo.anaconda.com/miniconda/Miniconda3-py39_4.12.0-Linux-x86_64.sh
chmod +x Miniconda3-py39_4.12.0-Linux-x86_64.sh
./Miniconda3-py39_4.12.0-Linux-x86_64.sh
conda update -n base -c defaults conda

Install the source and prepare the environment:

git clone https://github.com/CompVis/stable-diffusion
cd stable-diffusion
conda env create -f environment.yaml
conda activate ldm
pip3 install transformers --upgrade

Download the pre-trained model weights next. HiggingFace has the newest checkpoint sd-v14.ckpt (a download is free but registration is required). Put the file in the project folder and have fun:

python3 scripts/txt2img.py --prompt "hello world" --plms --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

Almost. The installation is complete for happy users of current GPUs with 12 GB or more VRAM. RuntimeError: CUDA out of memory will occur otherwise. Two solutions exist.

Running the optimized version

Try optimizing first. After cloning the repository and enabling the environment (as previously), we can run the command:

python3 optimizedSD/optimized_txt2img.py --prompt "hello world" --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

Stable Diffusion worked on my visual card with 8 GB RAM (alas, I did not behave well enough to get NVIDIA A100 for Christmas, so 8 GB GPU is the maximum I have;).

Running Stable Diffusion without GPU

If the GPU does not have enough RAM or is not CUDA-compatible, running the code on a CPU will be 20x slower but better than nothing. This unauthorized CPU-only branch from GitHub is easiest to obtain. We may easily edit the source code to use the latest version. It's strange that a pull request for that was made six months ago and still hasn't been approved, as the changes are simple. Readers can finish in 5 minutes:

• Replace if attr.device!= torch.device(cuda) with if attr.device!= torch.device(cuda) and torch.cuda.is available at line 20 of ldm/models/diffusion/ddim.py ().

• Replace if attr.device!= torch.device(cuda) with if attr.device!= torch.device(cuda) and torch.cuda.is available in line 20 of ldm/models/diffusion/plms.py ().

• Replace device=cuda in lines 38, 55, 83, and 142 of ldm/modules/encoders/modules.py with device=cuda if torch.cuda.is available(), otherwise cpu.

• Replace model.cuda() in scripts/txt2img.py line 28 and scripts/img2img.py line 43 with if torch.cuda.is available(): model.cuda ().

Run the script again.

Testing

Test the model. Text-to-image is the first choice. Test the command line example again:

python3 scripts/txt2img.py --prompt "hello world" --plms --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

The slow generation takes 10 seconds on a GPU and 10 minutes on a CPU. Final image:

Hello world is dull and abstract. Try a brush-wielding hamster. Why? Because we can, and it's not as insane as Napoleon's cat. Another image:

Generating an image from a text prompt and another image is interesting. I made this picture in two minutes using the image editor (sorry, drawing wasn't my strong suit):

I can create an image from this drawing:

python3 scripts/img2img.py --prompt "A bird is sitting on a tree branch" --ckpt sd-v1-4.ckpt --init-img bird.png --strength 0.8

It was far better than my initial drawing:

I hope readers understand and experiment.

Stable Diffusion UI

Developers love the command line, but regular users may struggle. Stable Diffusion UI projects simplify image generation and installation. Simple usage:

• Unpack the ZIP after downloading it from https://github.com/cmdr2/stable-diffusion-ui/releases. Linux and Windows are compatible with Stable Diffusion UI (sorry for Mac users, but those machines are not well-suitable for heavy machine learning tasks anyway;).

• Start the script.

Done. The web browser UI makes configuring various Stable Diffusion features (upscaling, filtering, etc.) easy:

V2.1 of Stable Diffusion

I noticed the notification about releasing version 2.1 while writing this essay, and it was intriguing to test it. First, compare version 2 to version 1:

• alternative text encoding. The Contrastive LanguageImage Pre-training (CLIP) deep learning model, which was trained on a significant number of text-image pairs, is used in Stable Diffusion 1. The open-source CLIP implementation used in Stable Diffusion 2 is called OpenCLIP. It is difficult to determine whether there have been any technical advancements or if legal concerns were the main focus. However, because the training datasets for the two text encoders were different, the output results from V1 and V2 will differ for the identical text prompts.

• a new depth model that may be used to the output of image-to-image generation.

• a revolutionary upscaling technique that can quadruple the resolution of an image.

• Generally higher resolution Stable Diffusion 2 has the ability to produce both 512x512 and 768x768 pictures.

The Hugging Face website offers a free online demo of Stable Diffusion 2.1 for code testing. The process is the same as for version 1.4. Download a fresh version and activate the environment:

conda deactivate  
conda env remove -n ldm  # Use this if version 1 was previously installed
git clone https://github.com/Stability-AI/stablediffusion
cd stablediffusion
conda env create -f environment.yaml
conda activate ldm

Hugging Face offers a new weights ckpt file.

The Out of memory error prevented me from running this version on my 8 GB GPU. Version 2.1 fails on CPUs with the slow conv2d cpu not implemented for Half error (according to this GitHub issue, the CPU support for this algorithm and data type will not be added). The model can be modified from half to full precision (float16 instead of float32), however it doesn't make sense since v1 runs up to 10 minutes on the CPU and v2.1 should be much slower. The online demo results are visible. The same hamster painting with a brush prompt yielded this result:

It looks different from v1, but it functions and has a higher resolution.

The superresolution.py script can run the 4x Stable Diffusion upscaler locally (the x4-upscaler-ema.ckpt weights file should be in the same folder):

python3 scripts/gradio/superresolution.py configs/stable-diffusion/x4-upscaling.yaml x4-upscaler-ema.ckpt

This code allows the web browser UI to select the image to upscale:

The copy-paste strategy may explain why the upscaler needs a text prompt (and the Hugging Face code snippet does not have any text input as well). I got a GPU out of memory error again, although CUDA can be disabled like v1. However, processing an image for more than two hours is unlikely:

Stable Diffusion Limitations

When we use the model, it's fun to see what it can and can't do. Generative models produce abstract visuals but not photorealistic ones. This fundamentally limits The generative neural network was trained on text and image pairs, but humans have a lot of background knowledge about the world. The neural network model knows nothing. If someone asks me to draw a Chinese text, I can draw something that looks like Chinese but is actually gibberish because I never learnt it. Generative AI does too! Humans can learn new languages, but the Stable Diffusion AI model includes only language and image decoder brain components. For instance, the Stable Diffusion model will pull NO WAR banner-bearers like this:

V1:

V2.1:

The shot shows text, although the model never learned to read or write. The model's string tokenizer automatically converts letters to lowercase before generating the image, so typing NO WAR banner or no war banner is the same.

I can also ask the model to draw a gorgeous woman:

V1:

V2.1:

The first image is gorgeous but physically incorrect. A second one is better, although it has an Uncanny valley feel. BTW, v2 has a lifehack to add a negative prompt and define what we don't want on the image. Readers might try adding horrible anatomy to the gorgeous woman request.

If we ask for a cartoon attractive woman, the results are nice, but accuracy doesn't matter:

V1:

V2.1:

Another example: I ordered a model to sketch a mouse, which looks beautiful but has too many legs, ears, and fingers:

V1:

V2.1: improved but not perfect.

V1 produces a fun cartoon flying mouse if I want something more abstract:

I tried multiple times with V2.1 but only received this:

The image is OK, but the first version is closer to the request.

Stable Diffusion struggles to draw letters, fingers, etc. However, abstract images yield interesting outcomes. A rural landscape with a modern metropolis in the background turned out well:

V1:

V2.1:

Generative models help make paintings too (at least, abstract ones). I searched Google Image Search for modern art painting to see works by real artists, and this was the first image:

I typed "abstract oil painting of people dancing" and got this:

V1:

V2.1:

It's a different style, but I don't think the AI-generated graphics are worse than the human-drawn ones.

The AI model cannot think like humans. It thinks nothing. A stable diffusion model is a billion-parameter matrix trained on millions of text-image pairs. I input "robot is creating a picture with a pen" to create an image for this post. Humans understand requests immediately. I tried Stable Diffusion multiple times and got this:

This great artwork has a pen, robot, and sketch, however it was not asked. Maybe it was because the tokenizer deleted is and a words from a statement, but I tried other requests such robot painting picture with pen without success. It's harder to prompt a model than a person.

I hope Stable Diffusion's general effects are evident. Despite its limitations, it can produce beautiful photographs in some settings. Readers who want to use Stable Diffusion results should be warned. Source code examination demonstrates that Stable Diffusion images feature a concealed watermark (text StableDiffusionV1 and SDV2) encoded using the invisible-watermark Python package. It's not a secret, because the official Stable Diffusion repository's test watermark.py file contains a decoding snippet. The put watermark line in the txt2img.py source code can be removed if desired. I didn't discover this watermark on photographs made by the online Hugging Face demo. Maybe I did something incorrectly (but maybe they are just not using the txt2img script on their backend at all).

Conclusion

The Stable Diffusion model was fascinating. As I mentioned before, trying something yourself is always better than taking someone else's word, so I encourage readers to do the same (including this article as well;).

Is Generative AI a game-changer? My humble experience tells me:

• I think that place has a lot of potential. For designers and artists, generative AI can be a truly useful and innovative tool. Unfortunately, it can also pose a threat to some of them since if users can enter a text field to obtain a picture or a website logo in a matter of clicks, why would they pay more to a different party? Is it possible right now? unquestionably not yet. Images still have a very poor quality and are erroneous in minute details. And after viewing the image of the stunning woman above, models and fashion photographers may also unwind because it is highly unlikely that AI will replace them in the upcoming years.

• Today, generative AI is still in its infancy. Even 768x768 images are considered to be of a high resolution when using neural networks, which are computationally highly expensive. There isn't an AI model that can generate high-resolution photographs natively without upscaling or other methods, at least not as of the time this article was written, but it will happen eventually.

• It is still a challenge to accurately represent knowledge in neural networks (information like how many legs a cat has or the year Napoleon was born). Consequently, AI models struggle to create photorealistic photos, at least where little details are important (on the other side, when I searched Google for modern art paintings, the results are often even worse;).

• When compared to the carefully chosen images from official web pages or YouTube reviews, the average output quality of a Stable Diffusion generation process is actually less attractive because to its high degree of randomness. When using the same technique on their own, consumers will theoretically only view those images as 1% of the results.

Anyway, it's exciting to witness this area's advancement, especially because the project is open source. Google's Imagen and DALL-E 2 can also produce remarkable findings. It will be interesting to see how they progress.