Dima T.

Airdrops 💎 | Trading 📊 | Spot Investments 🚀 | CEO of my spot portfolio | Strategy | Emotions | On the way to the first $100K 💰

[Dima T. @DimasikUSDT]

5 Secret Claude Codes Everyone Is Talking About

Five codes that change how Claude responds.

/ghost before your prompt and you get an answer that sounds like a real human and passes AI detectors.

/UDA at the end and Claude breaks down your situation from every angle and finds the real problem.

/OODA before your prompt and you get a concrete step-by-step action plan right now.

/L99 at the end and Claude responds at the highest expert level possible with no simplification.

/godmode before your prompt and you get the most complete and detailed answer it’s capable of.

⚠�� Note: These are not official Claude features. They’re prompt techniques that may influence the style and structure of responses rather than activate hidden modes.

[rewind @rewind02]

Andrej Karpathy spent 4 minutes in an interview explaining a single idea

about how most people haven’t even started learning how to use AI

and everyone paying $20/month for a subscription.. that's not really using Claude at all

his point is that the real skill gap is the ability to build with AI

he identified 4 behaviors that break Claude Code and put them all into one file

a developer expanded it into 21 rules and published it - 82,000 stars and #1 on GitHub Trending

coding accuracy jumped from 65% to 94%

here's what these 21 rules actually are and why most developers using Claude every day have never configured them

the full breakdown is covered in the article below 👇

[Dima T. @DimasikUSDT]

7. Analyze What's Actually Working

Use this prompt:

Below are the titles and performance metrics from my last 10 TikTok videos.Analyze the data and tell me:
What patterns do my best-performing videos have in common?
Which topics generate the highest interest?
Which content formats perform the best (storytelling, lists, tutorials, trends, Q&A, etc.)?
Do you notice any patterns in posting times or publishing schedule?
What factors seem to drive the most views, comments, shares, and watch time?Finally, give me 3 clear, data-driven recommendations for improving my next videos.Base your conclusions only on the performance data I provide, and focus on practical, actionable insights rather than generic advice.

[Dima T. @DimasikUSDT]

Most people think you need 100,000 followers to make money on TikTok.

You don't.

What you actually need is:

A proven strategy + the right Claude prompts.

In less than 20 minutes, Claude can:

Build a complete content strategy
Generate viral video ideas
Analyze your target audience
Create a monetization plan

These are the 7 prompts we use. 👇

[Pubity @pubity]

Claude models are winning the battle against every other AI company right now, with paid subscriptions for Claude going up by 75% since January 2026.

Anthropic is widely regarded as the fastest growing software company in history.

[Dima T. @DimasikUSDT]

6. Generate a 30-Day TikTok Content Calendar

Use this prompt:

Create a 30-day TikTok content calendar for my niche:[Insert your niche]Structure the content using the following mix:
70% Educational content
20% Entertaining or engagement-driven content
10% Promotional or sales-focused contentFor each day, include:
The video topic
The content format (list, story, Q&A, trend, comparison, tutorial, etc.)
The primary goal (views, followers, trust, or conversions)
A high-retention hook for the first 3 secondsMake the calendar optimized for TikTok's algorithm, with a balance of evergreen content, trending formats, and audience engagement strategies that support consistent long-term growth.

[Dima T. @DimasikUSDT]

Notion + Claude = the second brain I always wanted. Here's how I built it and why it changed everything 👇 → Step 1: created a Notion database called "Brain" — every idea, project, client note goes in → Step 2: every entry gets a tag: #idea #project #client #learning #prompt → Step 3: each week I paste the week's entries into Claude: "Find patterns, gaps, and opportunities in this data" → Step 4: Claude surfaces connections I missed — "you mentioned X three times, here's a potential product" → Step 5: every prompt that worked gets saved in a "Prompt Library" page — searchable, tagged by use case → Step 6: client meeting notes → Claude → structured action items in 2 minutes → Step 7: monthly review — Claude reads my whole month and writes a personal retrospective This system has never let me forget an idea. It's built 3 products and closed 5 clients. Want the Notion template? Comment "BRAIN" ⬇️ #Notion #Claude #AIproductivity #PKM #SecondBrain #AItools #Freelance #Solopreneur #ProductivityHacks #WorkSmart

[Dima T. @DimasikUSDT]

@DITOGAMESch This is one of the few browser games that actually looks fun.

[Dima T. @DimasikUSDT]

@RoundtableSpace What's the craziest optimization that made this fit into just 1MB?

[Dima T. @DimasikUSDT]

@AshCrypto Most people notice the rotation after it's already happened😅

[Dima T. @DimasikUSDT]

@yourcryptodj solana:5pYB12kEhfhSFXJjZ7JtyqDpt6uUqhsF6iu6Ee9spump @catwifhatsolana

[Dima T. @DimasikUSDT]

In the Age of Drones and AI, Someone Still Has to Walk

The quiet human job that powers one of the world's most-used apps

The year is 2026. Autonomous vehicles drive themselves across continents. AI generates entire cities from a text prompt. Drones deliver parcels to your balcony. And somewhere on a humid street in Kuala Lumpur, a person is walking with a giant blue camera bolted to their back — manually scanning the world, one step at a time.

Meet the Google Street View Trekker. Same job as 2007. Slightly better shoes.

The Machine That Didn't Get Automated

Google Street View launched nearly two decades ago and has since mapped over 220 countries and territories, with more than 250 billion images stitched into its seamless panoramic layer. The cars came first — those unmistakable rooftop rigs crawling down highways and suburban streets. Then boats for rivers and canals. Then snowmobiles. Then trolleys pushed through museums and shopping malls.

And then the problem nobody fully solved: everywhere a car can't go.

Narrow alleys. Markets. Pedestrian bridges. Hiking trails. Ancient medinas. The insides of airports. Beach promenades. Favela staircases. The kind of places that make up a significant chunk of the planet's actual lived geography — and that remain stubbornly inaccessible to anything with wheels and a motor.

The solution, still in active use today, is a 20-kilogram backpack called the Trekker, topped with a spherical camera array that captures 360-degree imagery every two seconds. You strap it on. You walk. That's it.

Why Drones Haven't Fixed This

The obvious question is: why not just fly a drone over it?

The short answer is that flying over something and walking through it produce fundamentally different results. Street View is built around the human vantage point — eye level, at walking pace, from the middle of a path. A drone image taken from 30 meters up is useful for mapping rooftops and checking parking lot capacity. It tells you almost nothing about what it feels like to navigate a place on foot.

There are also regulatory constraints. Urban airspace in most countries is tightly controlled, and flying a drone through a busy pedestrian market in Kuala Lumpur, Bangkok, or Marrakech isn't something you can do without permits, risk assessments, and a reliable guarantee that it won't clip someone's ear. Walking through the same market with a backpack requires considerably less paperwork.

Then there's image quality. Consumer and prosumer drones are getting better, but the Trekker's camera array — purpose-built for stitching seamless spherical panoramas at street level — still produces the kind of imagery that makes Street View actually usable. Close enough to read the sign above the door. Close enough to see the step at the entrance. Close enough, arguably, to count the tables at a restaurant.

The Scope of the Problem

Here's what makes this genuinely surprising: the scale of what hasn't been mapped yet.

Google's cars have covered an estimated 16 million kilometers of road. That sounds like a lot until you consider that the global road network spans roughly 64 million kilometers — and that's just roads. Add footpaths, market lanes, pedestrian zones, hiking trails, campus pathways, and transit hubs, and the unmapped territory grows enormously.

For every city center with meticulous Street View coverage, there are dozens of neighborhoods where the blue line on the map simply stops. For every well-documented tourist destination, there are hundreds of ordinary streets, ordinary markets, and ordinary residential areas where someone's elderly relative can't virtually "walk" the neighborhood they grew up in before visiting, because nobody has walked it yet.

The Trekker program — and the network of partner organizations and volunteers who borrow the equipment from Google — is the primary tool for closing that gap.

The Irony Is the Point

There's something worth sitting with here. We live in an era of genuinely staggering automation. Large language models write code, legal briefs, and marketing copy. Image generators produce photorealistic scenes from text prompts. Robots sort warehouse shelves and weld car frames. The breathless narrative of technological displacement is everywhere.

And yet: one of the most data-intensive, globally distributed, continuously updated mapping projects in human history still relies, in significant part, on someone putting one foot in front of the other.

Not because Google lacks the engineering talent to imagine alternatives. Not because the problem hasn't been studied. But because some environments are genuinely hostile to automation in ways that are expensive to solve, and a human with a backpack is often the most practical, most flexible, most cost-effective sensor platform available.

The Trekker weighs 20 kilograms. A drone powerful enough to replicate its output, fly legally in dense urban environments, avoid pedestrians, navigate covered markets, and operate reliably in rain costs considerably more and requires considerably more support infrastructure. The human just... walks.

What the Walker Actually Does

It's not a mindless job, despite how it might appear from the outside. Trekker operators have to maintain a consistent pace — too fast and images blur, too slow and the stitching gets uneven. They need to position themselves in the center of paths rather than hugging walls. They stop and restart at doorways to avoid capturing interiors without permission. They navigate crowds without bumping equipment. They manage battery swaps. They log coverage gaps.

In complex environments — a multi-level market, a transit hub with mezzanines, a university campus with buildings connected by covered walkways — the route planning alone requires real judgment. An algorithm can suggest a path. Knowing when to deviate from it is something else.

Google has also developed indoor mapping tools that work similarly, extending the principle inward: cameras on carts, on backpacks, on poles, operated by humans moving through spaces that are simultaneously too structured for drones and too irregular for fixed installations.

A Different Kind of Infrastructure Worker

The Street View walker occupies an odd position in the landscape of modern work. They're not the romantic figure of the artisan, doing something irreplaceable because it requires human touch and soul. They're closer to an infrastructure worker — someone whose job is to fill in the gaps of a system that is otherwise largely automated, operating in the exact spaces where the automation runs out.

There are more of these jobs than people tend to notice. The person who manually reviews flagged content that AI moderation couldn't classify. The technician who physically installs the fiber optic cable that enables remote everything. The warehouse picker who handles irregularly shaped items that confuse robotic arms. The Trekker operator walking a covered market in Malaysia so that someone on the other side of the world can virtually stand inside it.

These aren't jobs that automation forgot. They're jobs that automation, for now, can't reach.

The Footage

The TikTok video that circulated in August 2025, posted by The Sun Malaysia, showed exactly this: a Trekker operator walking through a busy urban street in Kuala Lumpur, the blue spherical camera tower rising above the crowd. The caption read "Walking the World, One Step at a Time." It went mildly viral, because people found it surprising — charming, even — that this was still how it worked.

That surprise is itself informative. We have collectively absorbed the idea that everything gets automated, and we're genuinely startled when something hasn't been. A person, a backpack, a camera, a city street. The world's most sophisticated mapping platform, being extended one step at a time.

The drone era is here. AI is rewriting entire industries. And someone just put on a 20-kilogram backpack and walked into a market so that you can look it up on your phone.

Some problems don't get solved by adding more technology. Sometimes the answer is a good pair of shoes.

[Alerta Mundial @TuiteroSismico]

🚐🇨🇳Un Furgón en china se pasea por las noches con un láser verde hacia el cielo ?

Es un LIDAR móvil: dispara luz verde al cielo, las partículas la devuelven y mide contaminación en tiempo real (altura y cantidad) luego se toman medidas de acuerdo a esto

[Dima T. @DimasikUSDT]

@bnb_girl1 may be near:native

[Dima T. @DimasikUSDT]

@solana

[Dima T. @DimasikUSDT]

How to position AI as your unfair advantage to clients → "I deliver the same quality in half the time" → "I use AI to catch errors humans miss" → "My process scales without adding cost to you" → "Every deliverable goes through 3 AI quality checks" → This is a feature. Sell it as one. #AItools #Freelance #ClientWork #Solopreneur #AIproductivity #ChatGPT #Claude #MakeMoneyOnline #OnlineBusiness #FutureOfWork

[Dima T. @DimasikUSDT]

@AiwithYasir Yeah, I get the idea — and I agree that compounding context is powerful. I think the real disagreement is just where the complexity lives. For me, if the system requires too much setup or maintenance, it breaks the ‘flow’ it’s supposed to create.

[Dima T. @DimasikUSDT]

@0x_Anni Wild how fast we moved from ‘AI tools’ to full-on AI production systems. At this point the edge is no longer in the model, but in how you build workflows around it.

[Dima T. @DimasikUSDT]

This is actually a really interesting direction. The shift from ‘checking outputs’ in a terminal to having a real-time physical interface is underrated — it turns agents from something you use into something you coexist with.
Feels like we’re slowly moving from software tools → embodied systems for AI workflows.