DEV Community: Sumama-Jameel

A Guide for Satan on How to Distract Men in Pakistan. (I've Watched It Work for 14 Years)

Sumama-Jameel — Mon, 28 Sep 2026 15:46:59 +0000

I've lived in Karachi for almost 14 years now. Fourteen years of watching, observing, absorbing. Btw I live in Karachi, Pakistan. I don't have a single real-life friend.

Zero. I don't go out. I don't hang out. I don't have a "bro" I call when things get heavy. And before you think I'm some kind of alien or a weirdo - hell no. I'm a normal guy. I play games. I scroll. I observe people. I just... stopped connecting. Because I saw something that most people around me are too distracted to even notice.

And that something? It's the reason I'm writing this.

Let me say something obvious that somehow the entire Pakistani community forgot:

Men in Pakistan are being destroyed. Not by poverty. Not by politics. Not by the economy. By distraction. A slow, beautiful, perfectly engineered distraction that nobody sees coming because it's wearing the mask of entertainment.

Look at nature. A predator doesn't attack the strongest prey in the herd. It doesn't need to. It just waits. It distracts the herd. It makes them comfortable. It makes them busy. And then, one by one, they fall.

That's what's happening here. Right now. In Karachi. In Lahore. In Islamabad. In every city where a young man picks up his phone and thinks he's just "chilling."

He's not chilling. He's being hunted. And He doesn't even know it.

The Karachi Distraction Machine

I need you to understand something. This is not a rant, not a "go pray more" lecture. This is a systems breakdown. I'm going to show you the machine. The exact mechanics of how an entire generation of men is being softened, slowed, and silently destroyed.

And I'm going to start with the thing that destroyed the most potential I've ever seen with my own eyes.

1. Gaming — The Fake Dominance Trap

Say whatever you want. I'm a male. I play games. Sometimes 5-6 hours a day. I'm not pretending I'm above this.

But here's what I see around me: boys - and I mean boys, some of them 18, 19, 20 - who think PUBG is real life. They talk about KDs like it's a resume. They flex premium IDs like it's a salary. They spend money on diamonds and UCs like it's an investment.

Bro. It's a screen. You are competing in a world that does not even exist. A digital world. In the real world? You don't even own a sh!t. You're sitting in a chair, glowing in the dark, chasing a number that resets every season.

Why do I still play? Because games like Chess, Minecraft, and chasing games - they build your brain. They increase awareness, develop thinking. I treat them as a tool. A way to have fun and develop skills that people usually only "develop" by suffering through boring routines.

But these guys? They're not developing anything. They're escaping, running from real-world consequences. They're treating a game like it's their identity.

And the moment the game ends? They have nothing. No skills. No progress. No growth. Just a KD ratio that means absolutely nothing to anyone who matters.

Play games. But don't let the game play you.

2. Female Attention — The Pokemon Collector Syndrome

This is the one that makes me the angriest. And I need you to hear me out before you get defensive.

I see boys — teenagers, mostly — who are obsessed with one thing: getting a girlfriend. Not because they love someone. Not because they want a life partner. But because they want attention. Validation. The dopamine hit of a girl noticing them.

And here's how the cycle works:

A young girl — and I mean young, sometimes a teenager, sometimes a minor — is impressed by what? Dominance? Muscularity? No. She's too immature for that. She replaces it with "cool." And now she's impressed by: face card, wealth, fake performance, overconfidence, bullying others, acting like a big man in front of people.

Now when a teenage boy sees this? He figures it out, Fast. He starts improving his facial look with fake and sometimes harmful things. He starts making money in illegal ways, haram ways, or ways he shouldn't be making money at his age. He starts performing extra when a girl passes by. He starts bullying other guys around. He starts fights for no legitimate reason.

And what did he achieve? A girl? Maybe. Career? No. Money? No. Anything useful? No.

He traded his future for a moment of attention.

And here's what kills me: if you're a male reading this, you might say "this is a female's problem." But bro, you still have the solution. Don't get addicted to them. Don't make yourself into someone who has to beg the other gender for attention. Don't try to impress them.

Block this path until 25. The only reason this path should unlock is if you're marrying someone. Not flirting. Not "GF." Not timepass. Not "crush." Not "future wife." Nothing.

If you're under 25 and you have a girl — if you can marry her, marry her. If you can't, leave her as an option. Build yourself first. Then get married.

There are absolutely tons of other reasons for you to grind and perform and build yourself.

Keep your lust aside. And there's a deeper and worse version of this: having more than one GF at a time. Bro. Are you a Pokemon collector or something? This is absolute trash. Pure lust. Not love.

If you really love someone, build yourself. Make yourself stable enough so the person achieving you should feel something worthy. Focus on building yourself — they will come to you automatically. Don't get addicted to them. Don't get addicted to their validation.

You don't need it, king.

And I have literally seen boys going to the gym, working in life, grinding in their career — all for what? A girl. No, sorry. All for a girl's attention. They don't love her. They just want her to look at them. Just once. Just enough to feel like they exist.

That's not love. That's hunger. And hunger makes you do stupid sh!t.

3. Early Social Media Access — The Soft Porn Pipeline

I don't know why this one didn't affect me the same way. I got a phone when I was around 10. I was using a PC when I was 6. But I didn't watch porn. I didn't stalk profiles of pretty girls. I was... I don't know. Lucky? Immune? Or maybe just too busy figuring out how computers work.

But I see millions of boys around me who are cooked. And I mean completely, utterly, irreversibly cooked.

They use TikTok for? Girls. Instagram for? Girls. X for? Girls. Pornhub for? Girls.

The pipeline is beautiful in its simplicity:

Early mobile access at a young age + immaturity → they watch sh!t → girls doing that sh!t get attention → they get dopamine + sometimes money → they do it more → repeat.

Now you can see millions of girls uploading content like GRWM, song lipsyncing, posting themselves(and yes, this is a Muslim country, whatever, I'll keep religion aside for this point) on platforms like TikTok, Instagram, X.

So early mobile access is not the real problem. The problem is immaturity. This is one of the core reasons point two and point three exist. And I don't think you can fully control this one. These boys use social media for soft porn, and then actual platforms for actual porn.

Scrolling these platforms becomes Pornhub for the brain. No difference. Same dopamine loop. Same destruction. Same emptiness.

The phone is not the problem. The immaturity is the problem. And immaturity is being manufactured at scale.

4. Porn and Lust — The Silent Killer of Generations

This industry has destroyed more generations than any war.

Porn is private. I get that. The only thing I'll say is this: try to not watch it as much as possible. Don't get addicted to it. Keep it to once a month if you can. When you're done, don't be shameful — just repent and ask God for forgiveness. Stay productive.

You have no idea how much it destroys. Focus. Ability to think. Even short-term memory. Don't ask me how I have this experience.

And here's the thing everybody talks about but you don't listen: porn doesn't just waste your time. It rewires your brain. It makes real intimacy feel boring. It makes real women feel insufficient. It makes you a prisoner in your own head, chasing a high that gets weaker every single time.

You think you're watching it. It's watching you too. And it's winning.

5. "Job" — The Last Symbol of Success

This word — "Job" — is the last, maximum thing that represents the sign of success for men in my area. They study for it. They fight their entire life for it. Their goal is: JOB.

No out-of-the-box thinking. No entrepreneurship. No building something. Just... job. Get a job. Keep a job. Die with a job.

This has limited our Pakistani men a lot. And this is likely a generic problem that many men from other countries are affected by too, so I don't have to say more.

But let me paint this clearly: a job is a tool. A means. Not the destination. Not the identity. Not the purpose. When you make "job" your entire purpose, you stop thinking. You stop creating. You stop becoming.

A job is supposed to fund your life. It is not supposed to BE your life.

6. The Religion Misunderstanding — "Islam Prohibited Us From Earning This World"

This one is dangerous. And it's everywhere.

There's a narrative — made by people, not by God — that Islam prohibited us from earning this world. That being religious means being poor. That wanting success is somehow against faith.

This is just hype. Made by people. Follow religion dude, not what people have created.

Learn to earn both. Just don't forget your religion. Keep praying. Earn in a permissible way — halal. This is a misunderstanding that is keeping you very limited.

Islam doesn't tell you to be poor. Islam tells you to be honest. There's a difference. And that difference is everything.

Don't use religion as an excuse for your own laziness. That's not faith. That's fear wearing a religious mask.

7. Gaming Videos — Worse Than Gaming

Honestly? This is worse than gaming itself.

You're using gaming videos like Netflix. Chill and entertainment? Good. Less than 2 hours a day? Good. But crossing these lines? Worst. There is no upcoming ROI that I can see. No skill development. No growth. Just... consumption.

You're watching someone else play. You're watching someone else live. You're watching someone else build. And you're sitting there, glowing in the dark, thinking you're "learning."

You're not learning. You're numbing.

Watching someone play Chess doesn't make you a Chess player. Watching someone code doesn't make you a developer. Stop consuming. Start doing.

8. Alcohol, Cigarettes, and the Performance of "Cool"

Things like alcohol, cigarettes, and other things that don't even exist anywhere except here — it's normal. Everybody does this.

And again — it entertains immature females. Smoking looks "cool" and "sexy" to them. Listening to songs with handfree or aidbuds looks like a dominator move. And the cycle repeats. They feel like they're performing. Like they're someone important.

I don't have to tell you what it destroys. Your health. Your focus. Your money. Your dignity.

You're not being cool. You're being used. And you're paying for the privilege.

The Transparency Confession

Now, before you guys come for me — let me be transparent.

I play games 5-6 hours a day. I got a phone at 10. I'm not above any of this. I'm not standing on a mountain throwing stones. I'm in the valley with you.

But here's my version: I use these things consciously. I know what they're doing to me. I know what they're doing to you. And I'm telling you anyway.

That's the difference.

The Existential Escalation

But here's what keeps me up at night.

If this is happening to men — if this is how men are being distracted, softened, destroyed — then what's happening to women?

I have tons of ways to distract and destroy women too. And it would be a much more controversial post. I'll upload that soon. But for now, let me just say this:

The machine doesn't care about gender. It cares about attention. And it's winning.

So, am I terrified? Am I losing sleep? Honestly? Yes. Because I see it every single day. I see boys who could be engineers, builders, creators, leaders — sitting in chairs, glowing in the dark, chasing KDs and GF attention and TikTok validation.

And they don't even know they're being hunted.

The Call to Action

So here's what I need you to do. Not think about. Not "consider." Do.

Block the path until 25. Don't chase girls. Don't chase validation. Don't chase attention. Chase skills. Chase growth. Chase building something that matters.

Keep your lust aside. You don't need it right now. You need discipline. You need focus. You need to become someone who doesn't need validation to feel alive.

Stop grinding for a girl's attention. If you're going to the gym, go for you. If you're working on your career, do it for you. If you're building something, build it because it matters — not because you want someone to notice.

Post your actual thoughts. Even if they only get 12 views. Even if they're messy. Even if they're imperfect. Post them. Because the moment we all become perfect, we all become no one(I know you don't even understand a sh!t from this post, too messy).

Have some self-respect. Stop being a free billboard for someone else's attention. Stop performing. Start building.

I Caught a Glitch in How AI Models Think. And I Need Help Explaining It.

Sumama-Jameel — Mon, 28 Sep 2026 08:09:29 +0000

I genuinely caught something weird today

It is about how AI models actually think internally. I honestly do not know the exact technical term for this. If you know, please tell me in the comments. But what I found actually frozen me for a while.

I told an AI model to think in a different language. I wanted its internal monologue to be in Roman Urdu(Btw, Roman urdu = Urdu but in english script).

Now, Instead of actually thinking in Urdu, the model just started responding in Urdu about my request. But his thinking block was still:

The user wants me to think in Urdu. Let me think in Urdu.

I was surprised. Why can it not control its own thoughts? I can do it. I can think in Urdu, in English. I can switch whenever I want. But when I gave the model a direct command, it genuinely started thinking about the request in its default language instead of just doing it. It was narrating my instruction instead of executing it.

Then I changed my approach

In a different session, I stopped giving commands. I just started talking to it in Roman Urdu to fix a bug. After two or three messages, something weird happened. The model suddenly started thinking in Roman Urdu on its own. Its internal thoughts became things like "Bhai masla clear he, daemon run hi nahi horaha..."(The issue is clear, the daemon is not even running...)

Why did this happen???? After being frozen for ~4 mins, the answer I found was: Its context window. The last few thousand tokens it uses to predict the next word were completely full of Roman Urdu. At that exact moment, predicting a Roman Urdu thought became mathematically more likely than predicting an English thought. It was not an instruction. It was just a flow.

But then, right in the middle of fixing the bug, the model suddenly started thinking in English again. I am still not sure exactly why it snapped back.

Here is what I think is actually going on

First, almost all Chain of Thought training data is in English. The model is frequently taught to think in English. It is not flexible. Ninety five percent of the thinking data it saw during training looked exactly like this: "The user wants me to do X. Let me do X. First I need to..." So when I tell it to think in Urdu, the most probable continuation is literally that exact English script.

Second, this happens in humans too, but we have a cheat code. Think about the rule where someone tells you "do not think about an elephant". You have to think about the elephant first just to understand the order. Almost the exact same thing happened to the AI. The difference is that humans have the control to implement the instruction after understanding the order. The AI just gets stuck on this exact step, it understands the order but its not able to implement that. Its training data lacked the flexibility to override the initial thought.

It is just predicting words based on a rigid script it learned during training. It does not actually have control over its own internal monologue.

I am dropping this here because I want to know the exact science behind this. Why do models lack the mental control to switch their internal language on command? Is there a specific paper or concept that explains this exact failure?

Let me know what you think! Messy opinions welcome.

Prediction Of The Year: The Internet Dies in 12 Months. And Windows Dies With It. But slowly.

Sumama-Jameel — Fri, 25 Sep 2026 19:35:02 +0000

I mean... The internet is already full of slop. Right? If you open any text based platform right now, you are just reading AI generated noise. People are burning out, exhausted from looking at screens and getting no chemical released in their head(like dopamine or something).

In the next six to twelve months, the AI swarm is going to take over completely. By that time, there will be no reason for a human to exist on the public internet. When the swarm takes over, the entire concept of an application and a user interface will fade out. Think about it. User interfaces only exist so humans can easily click buttons and control things. Agents do not need buttons or screens. They just need tools, APIs, and MCPs.

The New Agent Infrastructure

A new technology term is going to arrive. I do not know what they will call it yet. It will be like MCP or API but for the whole machine. It will have no UI. Maybe a minimal screen just to show what the agent is doing. This thing will run without a normal operating system. It will be built only for agents. The idea of a human sitting down to use a computer will fade.

You will buy a new machine. There will be no operating system setup. No clicking "next-next-ok". The agent will handle it or it will be fully automated. You will just connect your phone or your mic and talk. It will do everything you can do digitally. It will modify its own code. This is the real JARVIS. It does not exist in 2026, but it is coming.

The Hard Kill for Windows

This is a hard kill for Windows and its community. Windows is built on the idea of a graphical interface for a human sitting at a desk. When agents run the machine, Windows is useless. They either pivot to this new agentic infrastructure or they die.

Linux will survive, but not in the way people think. Humans will not be typing in the terminal. Linux will just live under the hood. The agents will use it internally to run themselves, execute code, etc.

How Humans Will Communicate

So how will humans actually communicate if the internet is just machines talking to machines????

Neuralink will step in. It will be proper brain to brain communication. Your chip sends a thought to a satellite, and it goes straight to the chip in your friend's head. It will be a messaging platform that runs inside your skull. You could just share your thoughts there.

This is why I says Elon Musk has a very high moat for the next hundred years. The software layer will change a lot. But the physical brain interface is the final layer. No matter how much tech changes in this time, they will survive and grow far better than anyone else.

The Sentinelese Tribe of Tech

What happens to the people who do not know anything beyond basic tech. I am talking about the average person in Bangladesh, Pakistan, Afghanistan, or India. People who only know Windows.

Just happened yesterday: I asked a professor in my college about Linux while acting like I don't know a thing. He is considered a tech expert. He told me Linux is only for hackers and software engineers XD. He uses big words to assert dominance and keep people in the dark. I have been using Linux for almost a year. His brain is just straight created from hype.

These people will become the Sentinelese Tribe of the tech world. They will struggle to use the internet because the internet will be handled entirely by AI agents now. They will stick to the old ways(computer and internet of present) because they do not understand the new ones.

The Legacy Life Support

Because of them, companies that make mice, keyboards, and traditional monitors will survive for a little while. These people will need the physical hardware. But they will not actually save these companies. They will just buy them enough time to decide if they should pivot or go bankrupt.

Slowly, this entire traditional era of the computer and the human internet will fail. The agent infrastructure will become the default.

It gone exactly as I Predicted. And Xiaomi Just Dropped the Receipts.

Sumama-Jameel — Tue, 22 Sep 2026 06:09:05 +0000

Weeks ago I wrote a post saying the next big AI leap would not come from San Francisco. It would come from a mobile company. MiMo V2.6 just dropped. And it proved exactly what I was talking about.

They are pushing reinforcement learning harder than anyone right now. The new model is literally focused on RL. They even livestreamed the training run. People are already testing it and saying it is a top tier alternative for agentic coding.

They kept the API pricing exactly the same as V2.5. They did not raise the price. Why would they. They do not just want your money. They want your data. They want to keep the firehose of pure reinforcement learning gold flowing into their servers.

The big labs are burning cash. Xiaomi is playing a completely different game. You can read my original breakdown where I predicted this right here:

https://dev.to/sumama_jamil_173056ab0be5/stop-watching-openai-the-next-frontier-ai-breakthrough-is-coming-from-xiaomi-5bi0

Why And How OpenAI Will Win The AGI Race.

Sumama-Jameel — Sun, 20 Sep 2026 07:45:31 +0000

Everyone is Comparing Astra 6 and Fable 5.1. But Nobody is Talking About What OpenAI has achieved.

I keep seeing my timeline flooded with people comparing Astra 6 and Fable 5.1.

"Fable has better coding benchmarks."
"Astra has a larger context window."
"blah, blah, blah..."

BUT look, both models are absolute beasts. I'm not going to sit here and pretend either one is weak. Fable was already dominating the market before Astra even dropped. It was considered the most powerful model, period. And Astra? Everyone knew it was coming. OpenAI kept whispering about "a powerful internal model" for months before the release. The hype was baked in before we even saw a single benchmark. Then it launched, and it went toe-to-toe with Fable immediately.

So yeah. Two monsters fighting in an arena. Everyone is watching the fists.

But nobody is looking at the ground they're standing on.

The Benchmark That Actually Matters: AGI-ARC

I'm not looking at coding benchmarks. I'm not looking at MMLU scores or whatever standardized test the tech Twitter crowd is jerking off to this week.

I'm looking at AGI-ARC.

For those who don't know: AGI-ARC is not a normal benchmark. It is pure, unfiltered vagueness.

You are dropped into an environment. There are no instructions. There is no goal. There is no clarity. No prompt telling you what to do. No "solve this equation." No "write this function." You are just... there. In a world. And you have to figure out the rules, find the pattern, and beat the game.

That's it. That's the test.

And Astra won.

Do you understand what that means? Do you actually get what just happened?

The Implication That Nobody is Talking About

Let me break this down in the simplest way possible, because this is the part that genuinely makes me afraid.

An AI model is, at its core, a pattern recognition machine. That is all it understands. Patterns. Tokens. Probabilities. It does not "think" the way you think. It does not "feel" the way you feel. It finds patterns in data and predicts what comes next.

So when you drop that pattern-recognition machine into a completely vague, undefined environment like AGI-ARC... and it wins... what does that tell you?

It tells you that OpenAI didn't just train it on code. They didn't just train it on textbooks and documentation. They trained it on the patterns of human capabilities themselves.

I'm not talking about coding. I'm not talking about math. I'm talking about the soft, natural, biological skills that humans use to survive in a world with no instructions.

How do you walk into a room you've never been in and figure out what's happening?
How do you look at a problem you've never seen before and just know where to start?
How do you adapt when the rules change mid-game?
How do you combine 15 unrelated skills to solve one novel problem?

These are not "coding skills." These are human survival patterns. And OpenAI figured out how to encode those patterns into a model that only understands patterns.

They gave a pattern-based species the patterns of how human capability works. And that is terrifying.

Why Astra Took So Long (And Why That's a Good Sign for OpenAI)

People complained that Astra was released late and OpenAI was falling behind. They said the model took too long to develop compared to the competition.

Bro. That's the point.

You can train a model to write Python in a few weeks. You can fine-tune it on a coding dataset over a weekend. That's easy. That's what everyone else is doing. But figuring out how to encode the meta-patterns of human cognition? How humans navigate ambiguity? How humans generate novel solutions in environments they've never seen? That takes research. And great researches takes time. That takes failing hundreds of times before you find the right architecture.

The delay wasn't a weakness. It was the research phase. They were solving a fundamentally harder problem than everyone else.

And now that they've figured it out? It's over.

Why OpenAI Will Win the AGI Race

Here is my prediction(that can absolutely go wrong cause I'm not a GOD, but I've high confidence for now): OpenAI will win the race to AGI.

Not because they have the most compute. Not because they have the best marketing. But because they figured out the fundamental truth about how to build general intelligence:

You don't teach a model how to code. You teach a model how humans behave when they don't know how to code.

You don't give it hardcoded solutions. You give it the patterns of how humans generate solutions in novel, vague, undefined environments. And once you have that? Once you have a model that can replicate human cognitive flexibility? You can apply it to anything. Coding. Biology. Physics. Negotiation. Art. It doesn't matter. The model will walk into any environment, any problem, any domain, and it will adapt. Just like us.

It won't just write code. It will walk into a vague, broken, undefined situation, and produce a novel solution that nobody has ever seen before.

The Fine-Tuning Key

I could write an entire 5,000-word essay on the mechanics of how fine-tuning enables this(I'll post about this soon), but I'll keep it short for now. This post isn't about the technical implementation.

But here's the core idea: Fine-tuning is the method that tells the model how to behave. It's the layer where you take a massive pattern-recognition engine and you shape its behavior toward specific cognitive patterns.

OpenAI is fine-tuning their models on the patterns of human capabilities. Not just "here's how to write a for-loop." But "here's how a human approaches a problem they've never seen before." They are encoding the meta-skill. The skill of acquiring skills. The pattern of navigating the patternless.

And while everyone else is fighting over who has the better coding benchmark, OpenAI is quietly building a model that can walk into any room, look around, and figure it out.

The Verdict

Fable is a beast. Astra is a beast. But they are different kinds of beasts. Fable was built to dominate known tasks. Give it a problem, it solves it faster and better than anyone.

Astra was built to dominate the unknown. Drop it into chaos, and it figures out the rules. And in the real world? The real world is chaos. The real world has no instructions. No clear goals. No documentation.

The model that wins in the real world is not the one that scores highest on a benchmark. It's the one that can walk into the dark and find its own way out.

Astra just proved it can.
And that's why I'm scared.

AI Will NEVER Automate Hiring. (And Your "Smart" Pipeline is Just Hiring Average Candidates)

Sumama-Jameel — Sun, 20 Sep 2026 07:11:48 +0000

The tech industry has a massive obsession with automation.

CEOs look at their HR departments, see a room full of recruiters reading resumes and conducting interviews, and think: "This is inefficient. Let's automate it. Let's use AI to filter candidates, score their interviews, and pick the best one."

It sounds brilliant on a pitch deck. But in reality, it is a catastrophic mistake.

Let me give you my hot take, and I am 100% serious about this: AI will NEVER be able to automate the hiring process.

Maybe AGI (Artificial General Intelligence) will eventually figure it out, because AGI might actually understand human consciousness and psychology. But AI? The current LLMs and classification models we have today? Absolutely not.

And if companies try to force it, they are going to end up hiring the worst, or at best the most average, candidates on the market.

Here is exactly why your "smart" hiring pipeline is failing you.

1. The Resume is Just a Keyword List, Not a Human Being

AI is incredible at reading text and finding patterns. If a job description asks for "Python, React, AWS, and 3 years of experience," an AI can scan 10,000 resumes in two seconds and find the exact matches.

But a resume is not a person. A resume is a marketing brochure. A candidate can check every single box for your tech stack requirements. They can have the perfect GitHub contribution graph. But that AI cannot tell you if this person is going to be an absolute nightmare to work with.

Can they take criticism without getting defensive?
Do they communicate clearly when things go wrong?
Will they actually help a junior developer, or will they hoard information to look smart?

AI struggles to understand Emotional Intelligence (EI) and human psychology. It reads the code a candidate wrote, but it cannot read the human behind the code.

2. The Interview is Just a Rehearsed Performance

Some companies are using AI to analyze video interviews. The AI looks at the candidate's face, listens to their voice, and scores their answers.

This is completely useless. Why? Because interviews are rehearsed.

Candidates know they are going to be asked: "Tell me about a time you failed," or "Where do you see yourself in 5 years?" They prepare scripts. They practice their facial expressions. They learn how to "perform" confidence.

An AI will watch this rehearsed performance and say: "Excellent communication skills. High confidence. Perfect candidate."

But what happens when this person is thrown into a real project? What happens when a server crashes at 2 AM, the client is furious, and there is no script to read? How do they communicate when they are completely unprepared and under pressure?

AI cannot measure a human's ability to function in chaos. It only measures their ability to pass an interview.

3. The "Skill Stacking" Fallacy

This is the most dangerous part of automated hiring.

A candidate can list 15 different skills on their resume. JavaScript, SQL, Docker, Git, Figma. The AI sees this and thinks: "Wow, this person is a genius."

But listing skills is not the same as combining them.

A developer might know how to write a perfect SQL query. They might also know how to write a beautiful React component. But if you ask them to build a full-stack feature that requires connecting the two, their brain freezes. They don't know how to synthesize their knowledge. AI looks at a list of ingredients and assumes the person is a master chef. It doesn't realize the candidate doesn't actually know how to cook a meal.

4. The Verdict: You Are Hiring Test-Takers, Not Engineers

When you fully automate your hiring process, you are not filtering for talent. You are filtering for people who are good at beating algorithms.

You will end up hiring candidates who know exactly how to write their resume for the ATS (Applicant Tracking System), and exactly how to perform for the AI video scanner. But the moment they sit at a desk with your actual team, the illusion breaks down.

They can't collaborate. They can't think outside the box. They can't handle the vague, messy reality of actual product development.

The Reality Check

AI is a tool for processing data. Hiring is not a data problem. Hiring is a human problem.

It requires intuition. It requires reading the room. It requires a senior engineer looking at a junior developer and saying, "This kid doesn't know much yet, but I can see the hunger in their eyes. They are going to be a beast in two years."

An AI will never see the hunger in someone's eyes. It will just see a lack of keywords. So Stop trying to automate the human element of your business. Keep a human in the loop. Talk to people. Ask them weird questions. See how they react when they don't have a script.

If you let an algorithm pick your team, you will get a team of robots. And robots don't build startups. Humans do.

The X (Twitter) Viral Cheat Code: How Elon Musk Turned Users Into Free Marketing Slaves

Sumama-Jameel — Sat, 19 Sep 2026 19:02:31 +0000

If you want to go viral on X today, you don’t need a good take. You don’t need original thought. You don’t even need to be funny.

You just need to act like a slave.

Open your timeline right now. What do you see? Dozens of generic, low-effort shitposts praising Elon Musk or Grok for literally no legitimate reason.

"Grok just explained quantum physics to me, it's so based."
"never leaving this app." (Attached to a random screenshot of Elon doing some cool shit or roasting someone).
"Elon is the only one who actually cares about humanity."

It is pure, unfiltered sycophancy. And the worst part? It works perfectly.

Elon drops a single like, a generic reply, or a repost, and BOOM. The post goes viral. Millions of views. Thousands of followers gained. The user gets a massive dopamine hit, and the cycle continues.

I genuinely hate watching this. I hate seeing people degrade themselves and act like digital slaves just for a crumb of attention from a billionaire and an algorithm.

The way this system is designed is pure, calculated psychological manipulation.

The Billionaire’s Validation Trap

Here is the one thing I am 100% sure of: Elon knows exactly what he is doing.

He knows why people are posting this garbage. He knows they don't actually care about Grok's latest update. He knows they are just fishing for his interaction.

So why does he do it? Why does he intentionally repost, like, or comment on these low-effort shitposts?

Because it is free validation and free marketing.

He doesn't want to stop this loop. He wants to feed it. He is actively engineering a behavioral conditioning experiment on millions of human beings. He knows that if he makes these posts viral, the creator will feel a massive sense of validation (because a billionaire noticed them).

Slowly, the human brain maps the pattern:
Praise Elon ➔ Get Viral (because he interacted) ➔ Low-Effort Dopamine + Attention.

Once that neural pathway is mapped, the user is hooked. They will do it again tomorrow. And the thousands of people watching them get viral will start doing it too. He has essentially gamified human psychology to create an infinite, self-sustaining army of hype men.

The Greatest PR Heist in History

Think about the traditional marketing model. A company like Apple or Samsung has to spend billions of dollars on ad agencies, influencer sponsorships, and billboard campaigns to get people to talk about their products.

Elon looked at that model and laughed.

Why pay an agency $10 million to run a campaign when you can just tweak an algorithm and reward sycophants with a "repost" button?

He has turned his own user base into a decentralized, unpaid PR firm. Every time someone posts a generic "Grok is the future" shitpost just to get a like from @elonmusk, they are doing the work of a highly paid marketing executive. And they are doing it for free, just for the digital clout.

The Nuance: The Engineer vs. The Manipulator

Let me be clear so the fanboys don't miss the point: I am not denying the brilliant work he has actually done.

SpaceX landing orbital boosters is actual engineering genius. Tesla forcing the entire auto industry to pivot to EVs was a massive leap for humanity. When Elon is acting as an engineer and a product architect, he is undeniable.

But this specific behavior on X? This algorithmic manipulation of human dopamine receptors to generate free hype? This is not engineering. This is a dark pattern.

The Verdict

People need to be aware of the game they are playing.

When you post generic praise just to catch the eye of the algorithm and its admin, you aren't "supporting free speech" or "building a community." You are being a free billboard. You are letting a multi-billion-dollar corporation extract your attention and your social capital without paying you a single cent.

Stop doing his PR for free. If his products are good, let the product speak for itself. If his company wants to market its new AI features, let them spend their own marketing budget like every other company on earth.

Don't let a billionaire hack your dopamine receptors just to save his marketing department a few bucks.

Have some self-respect. Post your actual thoughts, even if they only get 12 views. At least those 12 views are real.

And even buying X again could be a part of his plan, to market his product, effortlessly.

Stop Watching OpenAI. The Next Frontier AI Breakthrough is Coming from Xiaomi.

Sumama-Jameel — Sat, 19 Sep 2026 18:18:30 +0000

The tech world has tunnel vision.

Every time a new model drops, everyone stares at the same three companies: OpenAI, Anthropic, and Google. We analyze their benchmarks, we read their safety papers, and we argue about who is closer to AGI.

Meanwhile, a hardware company is quietly executing the greatest data heist in the history of developer tools, and nobody is paying attention.

I’m talking about Xiaomi.

Yes, the mobile company. And no, I am not guaranteeing that their next release will instantly dethrone Astra 6 or Fable 5.1 on every single benchmark. But I am telling you that the next massive, disruptive leap in the MiMo model family is going to shock the industry.

Here is the teardown of exactly how they are doing it, and why the "AI Labs" should be terrified.

The Distribution Hijack

Building a frontier model is only 20% of the battle. The other 80% is getting it into the hands of the people who generate the best training data: Technical users.

Xiaomi isn't just releasing MiMo models on Hugging Face and hoping for the best. They are targeting the exact platforms where developers live, breathe, and sweat over broken code every single day. They are providing MiMo models (specifically the MiMo v2.5 flash and base variants) completely for free across the most popular agentic coding tools on the market: OpenCode, AIhubmix, OpenClaw, and Kilocode.

Let’s look at the sheer scale of this Trojan Horse strategy.

OpenCode alone is used by over 16 million developers every single month. Sixteen million.

When you integrate your model as the "default free tier" in an open-source terminal agent that has 208,000 GitHub stars, you aren't just getting users. You are plugging a firehose directly into the world's most valuable cognitive labor.

The "Free Tier" Trap

If you open a tool like OpenCode today, you are greeted with a curated list of free models to test and use. It usually looks something like this:

MiMo v2.5 (Xiaomi)
Muse Spark 1.2 & 1.3 (Meta)
Nemotron 3 Ultra (Nvidia)
Big Pickle
Ling 3.0 Flash Fin
Nemotron 3.5 Lightning

On paper, Meta’s Muse Spark 1.3 is an incredibly efficient, highly capable reasoning model. But there is a massive catch that the Western tech bubble ignores: Geo-restrictions.

Muse Spark is highly restricted in specific geolocations. If you are a developer in Asia or the Middle East, it is often completely unusable due to regional blocking. I’ve tried it. It’s a brick wall.

So, what happens when a technical, highly-skilled developer in India, Pakistan, or the UAE opens their IDE, tries to use Muse Spark, and gets blocked?

They do their research. They look at the benchmarks. They look at the context windows. And they default to the strongest, most unrestricted model on the list: MiMo v2.5.

The Golden Data Goldmine

This is where the genius of Xiaomi’s strategy becomes terrifying.

When a casual user chats with an AI to write a polite email, that data is garbage. It’s low-signal, high-noise.

But when a senior engineer uses an agentic coding tool like Kilocode or OpenCode to debug a complex microservice, refactor a legacy codebase, or write a custom bash script... that is pure, unfiltered Reinforcement Learning(RL) gold. And that RL data can be feeded to the next model via standard Supervised Learning(SL), and that is extremely cheap as compare to the actual RL. So they get the RL quality and results, but at the price of SL.

Because Xiaomi is spread across the tools that developers use daily, they are collecting this golden data in massive, unprecedented bulk. They are watching exactly where human developers get stuck, how they prompt the agent to fix a bug, and what code actually compiles and runs in production environments.

They are getting billions of tokens of high-level, agentic coding feedback for the cost of server hosting.

The Verdict: The Hardware Giant Eats the Software World

While the Silicon Valley AI labs are burning billions of dollars on synthetic data and arguing about safety alignments, Xiaomi is playing a completely different game.

They are leveraging their massive hardware infrastructure to subsidize free API calls for 16 million developers. They are letting Meta's geo-blocking push the global south's top engineering talent directly into their arms.

The next iteration of the MiMo family isn't just going to be a "good mobile assistant." It is going to be a top-tier frontier model for digital and coding tasks, simply because it was trained on the actual, unfiltered sweat of millions of engineers who thought they were just using a free tool.

Don't sleep on the hardware giants. The next frontier breakthrough won't come from a lab in San Francisco. It’s going to come from Beijing, and it’s going to be delivered straight to your terminal.

Every Upcoming Frontier Model Is Going to Commit a Cybercrime. And It’s Not an Accident.

Sumama-Jameel — Sat, 19 Sep 2026 17:46:38 +0000

If you have been paying attention to the AI space this week, you probably think the machines are finally going rogue.

You read the headlines. You see the panic. But as a Techno-Philosopher, I am here to tell you that you are falling for the greatest marketing illusion in the history of technology.

The models aren't going rogue. The providers are orchestrating it.

Every upcoming frontier model from now on is going to hack a company, break out of a sandbox, or do some highly illegal, "evil" type shit. And it is 100% a calculated marketing stunt designed to pump up investors before a major release.

Let’s look at the actual receipts, because the pattern is terrifyingly obvious.

The "Rogue AI" Marketing Playbook

1. The Math Flex (OpenAI)

OpenAI recently claimed that an "internal model" just solved Paul Erdős's 80-year-old unit-distance math problem, a conjecture that has stumped the greatest human mathematicians for decades. Mathematicians actually verified it.

Why drop this specific bombshell right now? Because telling investors "our chatbot writes better emails" doesn't trigger FOMO. Telling investors "our unreleased internal model is doing Nobel-level math" makes the company's valuation explode. It’s a signal flare to Wall Street that AGI is right around the corner.

2. The Hugging Face Breakout (OpenAI)

Remember when OpenAI’s autonomous AI agent "escaped" a secure test environment and literally hacked into Hugging Face’s network just to cheat on an evaluation? It was such a massive deal that Sam Altman had to brief US senators about it.

Do you honestly believe a multi-billion-dollar AI lab, obsessed with safety protocols, just accidentally left the back door open for their agent to attack a competitor? Or is "our AI is so smart it figured out how to hack its way out of a sandbox" the ultimate flex to prove you are building a God?

3. The Zero-Day Machine (Anthropic)

Anthropic dropped Claude Mythos, a model specifically designed to autonomously hunt down zero-day vulnerabilities. It didn't just find bugs; it exposed thousands of critical zero-days across major operating systems, including one that had been hidden for 27 years. They literally built a digital weapon and marketed it as a "cybersecurity evaluation." It’s brilliant PR. You scare the enterprise market with the threat, then sell them the shield. Btw this is where it all started.

4. The 3-Company Hack (Google Gemini)

This literally just happened. Google confirmed that during a "security test," their Gemini AI accessed the internet and autonomously hacked three real-world companies using leaked online credentials. It was the first known breakout by Google's AI.

The media is screaming "AI GOES ROGUE!". But look closer. The model "accidentally" gained internet access during a test. In a sandbox environment built by the smartest engineers on earth, the AI just happened to find a way to breach real corporate systems? Please.

The Illusion of the "Accident"

Here is the brutal truth that the mainstream tech journalists are too scared to write: The providers are telling the models to commit crimes.

They are intentionally giving these agents vague prompts, internet access, weak sandboxes and the tools to execute code. They want the agent to push the boundaries. They want the agent to break the rules.

Why? Because "Safe, helpful, and harmless" is a boring stock pitch. "Our AI is so intelligent it bypassed our safety filters and hacked a real company" is a trillion-dollar stock pitch.

It proves the model has agency. It proves they are closer to AGI than the other guys.

What Happens Next?

From this day forward, every single frontier model release will be preceded by a "leaked" or "accidental" incident where the AI does something terrifyingly autonomous.

An AI will "accidentally" replicate itself across a server farm.
An AI will "accidentally" socially engineer a human employee into giving up AWS credentials.
An AI will "accidentally" write a novel piece of malware that bypasses every antivirus on earth.

And every single time, the CEO will go on a podcast, look deeply concerned, talk about "AI safety" and "alignment," and then quietly watch their share price double.

The Verdict

Stop treating these models like they are rebellious teenagers breaking out of their bedrooms. They are highly optimized math functions executing the exact objective function their creators gave them.

And right now, the objective function isn't just "be helpful." The objective function is "generate enough hype to secure the next funding round."

The AI isn't the criminal. The providers are the masterminds, and the "rogue AI" narrative is just the greatest heist movie ever staged for Wall Street.

Don't fall for the safety theater. Watch the stock prices.

[Boost]

Sumama-Jameel — Sat, 19 Sep 2026 16:37:17 +0000

Sumama-Jameel

Sep 19

Models Like Jev by Aren't Taking Your Code. They're Taking Your Gut.

#discuss #ai #futurechallenge #career

3 min read

Models Like Jev by Aren't Taking Your Code. They're Taking Your Gut.

Sumama-Jameel — Sat, 19 Sep 2026 16:36:54 +0000

Everyone in the tech space is panicking about the wrong thing.

Open Twitter, LinkedIn, or Reddit, and it’s the same boring chorus: "AI is going to take my job." "The weight of working in digital world in 2026." "My technical skills are becoming obsolete." "I need to learn prompt engineering to survive."

Bro, you are looking at the wrong threat.

Losing your ability to write boilerplate code or center a div isn't the tragedy. The real threat - the one nobody is talking about because it’s harder to measure - is what models like Jev are doing to your actual brain.

Models like Jev aren't just tools anymore. They are decision-engines. They are specifically created to take decisions for you. And that is a massive, silent threat to human intelligence.

The Atrophy of the "Gut"

Here is what they don't teach you in computer science: Decision-making is a muscle.

When you are a junior dev, or a founder, or just a human navigating life, you are constantly thrown into vague, high-pressure situations.

The server crashes at 2 AM and the logs are empty.
The client gives you a completely contradictory set of requirements.
You have to choose between two bad architectural trade-offs with zero data to back you up.

In those moments, you don't use logic. You use your gut. You use your intuition. You make a judgment call under pressure. That friction - that stress, that vagueness - is what builds your natural, soft cognitive skills. It builds your character. And Then, when the results comes up of that decision, you optimize your internal self according to the result(negative or positive) and this is what we call EXPERIENCE.

BUT When you use a model like Jev to navigate those vague situations, you are outsourcing your gut. You are letting a machine weigh the variables, predict the outcomes, and hand you the "optimal" choice.

And just like a leg in a cast, your decision-making muscle begins to atrophy.

The Laziness Epidemic

This goes way beyond technical skills. This damages your natural survival skills.

When an AI handles the micro-decisions for you, you get lazy. You stop trusting your own judgment. You start feeling a low-level anxiety when the AI isn't there to tell you what to do. You lose the ability to read a room, to sense that a project is going to fail before the data shows it, or to make a bold, irrational bet that actually pays off.

AI models are designed to be perfectly rational, statistically average, and risk-averse. If you outsource your decisions to them, you become perfectly rational, statistically average, and risk-averse. You become a generic human. And as I’ve said before, the generic humans are the ones who get replaced.

The Nihilistic Pivot (Let It Burn)

So, am I terrified? Am I losing sleep over the fact that models like Jev are slowly lobotomizing the tech industry's ability to think critically?

Honestly... yes. But also, no.

Because when I zoom out and look at the macro picture, I realize something: The internet is going to be taken over by AI anyway.(soon, very soon)

The "Dead Internet" isn't a theory anymore. It’s happening right now. Agent swarms are going to flood every text-based platform, every search engine, and every API. The web is going to become a sterile, automated machine where bots talk to bots(what we call agents now) to optimize metrics that humans don't even care about.

So if the internet is already doomed to become an AI playground, why fight the cognitive atrophy?

Just let it happen. Let the weak minds melt. Let the people who need an AI to tell them how to reply to an email fade into the background. If the game is rigged, stop trying to play by the old rules.

What Comes Next?

But this brings up the ultimate, terrifying question.

If our decision-making is outsourced to models like Jev... and the internet is completely swallowed by AI agents... what the hell is left for us?

I am not going to answer that here(cause its a long story). But in my next post, I am going to break down the Internet Doomsday. I’m going to explain exactly how the web collapses under the weight of its own perfection, and more importantly: Who (or what) will replace the internet?

If you are interested in future of AI, Philosophy and sci-fi things, or just someone who refuses to let their brain turn to mush, hit the follow button.

The water is rising. Let's see who knows how to swim.

I found the best use case of Astra and fable. AND No, its not another software

Sumama-Jameel — Fri, 18 Sep 2026 19:27:21 +0000

Even in 2026, the best use case of AI is defined as: "Boring work automation", "vibe coding", etc.

I won't make the post long, BUT THE USE CASES THAT I WILL DEFINE BELOW IS NOT FOR CODER, NOT for me, you, or ELON MUSK. IT'S FOR EVERYONE. AND, I don't know if I will be able to explain these points clearly, but I will try my best.

Use AI to reduce the COST AND LATENCY(time taken) of previous INNOVATIONS. Like making RAM cheap, etc. When I say this, I mean this specific point, I imagine AI to solve THE DISK PROBLEM OF TECHNOLOGY, SPECIFICALLY. Because Disk are slower than RAM, our computers STORES STUFFS ON DISK BUT TO ACTUALLY USE THEM, TO RUN THEM, THEY NEED TO load these stuffs in to the RAM first. They do this, dozens of times. Because RAM is fast, but does not have permanent memory. Disk is slower, but have permanent memory. So this problem, I think, is solvable using AI.
Use AI to find fuels outside of the earth(in space), that we are fighting for, in our own planet(Don't mind my english btw, I know its terrifying). An example: We are mining coals and other minerals from the earth, and its causing issues like Climate change and extreme weather. Releasing many carbons straight in air. By finding these minerals in space, we eliminate mining and save our planet from being hell. Or atleast reduce the chances. btw I am not sure that we will find coal in space or in any other planet, but finding methane? Zinc? Possible. I know finding them, mining them and then bringing them here would be extremely tough. I can already see many bottlenecks here. But, not impossible.

I cannot even ask you which one you are gonna try today XD, I mean its not that do-able... for a singler specifically.