About
I’m building stock.ex.gold to modernise how jewellery businesses manage stock, sales…
Services
Activity
3K followers
-
Omer Arshad shared thisThere's a line often attributed to Warren Buffett: make money while you sleep, or work until you die. I think the same now holds for AI agents. If you don't have agents working for you while you sleep, you're starting every morning a step behind the people who do, or soon will be. One caution. Invested money earns without you. What an agent leaves you today, for most kinds of work, is work: a draft, a report, a pull request you still have to read. I don't think that stays true for long, but for now the gain is a head start, not income. Nobody loses their job over a night. What you lose is the groundwork: the first pass, the triage, the draft that could have been waiting for you. The smallest job to try tonight: one that ends on its own, can be thrown away, can be checked in a short read, and sends, pays, publishes or deploys nothing. A briefing on yesterday's commits, say, or a digest of what changed in the repositories you follow. Write the brief before you close the laptop, and give it the narrowest access that lets it succeed. In the morning: did it finish? Did it stay inside the brief? Does it hold up when you run the tests or open the sources? Was it small enough to read in full? If it looks horrible, work out why before you run it again. Give it a try. #AIAgents #SoftwareDevelopment #Productivity
-
Omer Arshad shared thisIf you've ever visited rural Pakistan, you'll have seen them. Big, castle-like houses, sometimes more than you'd think a village could hold, standing mostly empty among old and broken ones. Overseas Pakistanis built a lot of them. I used to wonder why, until I grew up and understood the psychology. A young man in those villages dreams of such a house. Luck takes him abroad, and by the time he can afford it nobody may need it. That dream still lives, so in his old age he goes back and builds it. The dream outlives the need. That story came back to me while I was scrolling past SaaS launches on YouTube and LinkedIn. For a while, SaaS was the billion-dollar dream, and not everyone got to play. Now LLMs and agents give people the resources to fulfil that old dream. I caught myself asking the same question. Why? The answer was the same human psychology, and after working on many SaaS products for different people, I felt that pull to build one myself. The SaaS world has moved on, though it isn't 100% dead, as we still need some buildings in those rural areas. I'm questioning the reason, not the product. If you're looking for the next big business idea, don't start from the idea. Start with your resources, constraints and what you're good at, then work backwards with your agents. I asked 'Why?' about the castles long after they were built. You can ask it before you start building. #AI #SaaS
-
Omer Arshad shared thisI'm building an AI-native company, and when I hire someone I won't be paying for the languages they know or the tools they've used. I'll pay for their context. I don't mean the model's context window. I mean a person's context in their own field: what we're building, why, and how to control and ship it. Tools still matter, just not as the thing I'm paying for. I've shipped software for most of my career, and the ironic thing is how many people got by with hardly any context, on both sides. Some managers and product owners didn't know what they were asking for. Nicer words and a good report carried them. Some developers survived on copying code they didn't understand. It worked because others made sense of the jargon before it reached the code. I did it myself. I had a job and a salary to protect. With an agent, I can't count on that buffer. Give it a vague brief and it often guesses instead of asking what you meant, then hands back something confident and almost right. You need real context to see what's wrong with it. So the people with the least context feel the shift first, and bigger words won't carry them. How you build it hasn't changed: reading, studying, experimenting and shipping small components first-hand. The only difference is that there used to be ways around it, and now it's going to be much harder than ever to survive without it. #SoftwareEngineering #AI
-
Omer Arshad shared thisI’ve been building AI workflows to make my day-to-day work more efficient. But one small task kept feeling cumbersome: typing up the context they need. At the end of the day, I want to give my AI pipeline my meeting notes, work observations, to-do lists, and instructions for the next few days. You name it. Speaking those notes felt much more natural than writing everything out. I’d seen tools that offered voice dictation, but the ones I came across meant another subscription. So I built DayScribe. It’s a small, free, open-source Mac app that turns voice notes into text: • Press a shortcut, speak, and press it again to finish. • Your note is transcribed locally and added to a daily Markdown file. • It works offline after the initial setup, with no account or API key needed. That gives me a simple way to capture the context I want to feed into my AI workflows, without typing it all out. I built it for my own needs, but it could be useful to anyone who prefers talking through their notes. If you’re on macOS 14 or later, give it a try: https://lnkd.in/e4_4JvrvGitHub - laldinsoft/dayscribeGitHub - laldinsoft/dayscribe
-
Omer Arshad reposted thisOmer Arshad reposted thisadidas please think about damage Control Israel has created the world largest population of child amputees. Where is Adidas Support for These Kids not the IDF Let’s start with the thousands of Kids without Limbs they need all the help your Money can buy #boycottadidas will not be over anytime soon on the contrary #Boycott #Adidas has just started
-
Omer Arshad shared thisRunning an AI-native company (ANC) has changed my focus: closing the loop. When I first started incorporating AI into my work, I focused on making individual tasks faster. Now I spend more time connecting those tasks into complete systems. If I still have to carry every output from one tool to the next, explain the context again, and trigger the next step, I am still the bottleneck. The diagram below shows one component I have built and am already using through my media director AI agent. A single prompt produces a production-ready video. It handles the visuals, music, narration, and final edit as one connected workflow. What I find most interesting is where this component sits. The input prompt comes from another manager AI agent. A marketing AI agent picks up the finished video. This whole production workflow becomes one component inside a larger operation. This is how I am building my company: multiple agents coordinating across models and clouds to run departmental workflows, with smaller systems becoming components of larger ones. It is far more efficient, gives me finer control to get the output exactly how I want it with rarely any retries, and costs much less than most of the overhyped AI video sites. How are you connecting the dots in your company, and where do you still find yourself carrying work from one AI tool to another?
-
Omer Arshad shared thisAI may have changed how you work. But has it changed how you earn? People are using AI to automate tasks, work faster, serve more customers, and increase the output of their existing jobs or businesses. That is valuable. But it is still mostly about doing more of what they were already doing. I am interested in a different question: Have you used AI to create a completely new source of income? Not simply to become faster at your existing job. Not only to increase revenue in your current business. I mean a new service, product, audience, or small business that was not part of how you earned before-and that AI made practical for you to try. AI gives us leverage. But leverage does not automatically become opportunity. Often, the time we save is quickly absorbed by more tasks, more meetings, or higher expectations. Our productivity increases, but our economic options stay the same. AI creates efficiency. Efficiency creates capacity. The interesting question is: who benefits from that capacity? And have you used any of it to build something of your own? This is not about producing a generic AI ebook over the weekend or chasing effortless passive income. A sustainable income source still needs a real problem, a real customer, and something genuinely useful. But AI may have lowered a barrier to entry, cost, or a missing capability - that once made an idea impractical. So I am curious: has AI helped you create a genuinely new income stream, rather than simply improve an existing one? If it has, what did you build? #AI #FutureOfWork #Productivity #Entrepreneurship
-
Omer Arshad shared thisCan't quite believe I'm writing this already: stock.ex.gold is hiring. We started out building the system of record for jewellers: stock, live rates, storefront, one source of truth. Jewellers started using it. Now we need more hands: • Customer Success Leads — Dubai, and multiple cities across India • Senior Frontend Engineer (Angular) — Dubai • Backend Engineer (Go) — remote • Mobile Engineer (Flutter) — Dubai Five roles, no nonsense, every application read by a human. → https://lnkd.in/e6GUjhbZ #hiring #startup #jewellery #fintechBuild the system of record for jewellers.Build the system of record for jewellers.
-
Omer Arshad liked thisOmer Arshad liked thisOpenClaw, Hermes, Grok Bot, Muse. Different names, same big shift. 2026 is the year AI started moving beyond chatbots you visit to agents that work for you. These agents can connect to your apps, use computers, execute tasks, and keep working with minimal human intervention. Here are 10 personal AI agents and platforms worth watching 👇 1️⃣ OpenClaw: The open-source agent that helped popularize the personal AI agent movement. 2️⃣ Meta Muse: Bringing AI experiences closer to everyday users through Meta’s ecosystem. 3️⃣ Hermes Agent (Nous Research): An open-source agent designed to learn from experience and create its own reusable skills. 4️⃣ Grok Bot: An emerging approach to managed agents with dedicated cloud computing environments. 5️⃣ OpenAI Dots: OpenAI’s vision for a proactive personal agent that can remember context and work on your behalf. 6️⃣ Claude (Anthropic): Powerful agentic capabilities for working across files, code, tools, and complex workflows. 7️⃣ Gemini (Google): Bringing AI assistance into the Google ecosystem, including Gmail and Calendar. 8️⃣ Nanobot: A lightweight approach to building personal AI agents. 9️⃣ ZeroClaw / NanoClaw: Lightweight agent frameworks focused on running agents with greater control over their environments. 🔟 Instinct: An emerging name in the race toward more proactive personal AI agents. The capabilities and availability of these tools vary, but the direction is clear: AI is becoming more autonomous. THE NEXT BIG OPPORTUNITY ISN'T JUST BUILDING AGENTS. IT'S MAKING THEM WORK IN THE REAL WORLD. Someone needs to deploy them, connect them to business systems, monitor their performance, and make sure they deliver results. That's where new roles are emerging: 🔧 AI Agent Engineer: Builds, deploys, and extends AI agents. ⚙️ AI Automation Engineer: Connects agents to CRMs, email, Slack, WhatsApp, and business workflows. 🤝 Forward Deployed / AI Solutions Engineer: Adapts AI agents to real customer problems. 📊 AgentOps Specialist: Monitors reliability, performance, security, and operating costs. 🎛️ AI Agent Operator: Configures, supervises, and manages agents, increasingly without writing code. Five years ago, every company needed a website. Tomorrow, every company may need an AI agent strategy. And the biggest opportunity may not be building the smartest model. It may be deploying the right agents to solve real business problems. Which agents are you using today, and which roles are you hiring for? 👇 #AIAgents #OpenClaw #AgenticAI #FutureOfWork #Automation #ReactiveSpace
-
Omer Arshad liked thisOmer Arshad liked thisPlease keep pixelating your passwords. People like me really appreciate it. Because someone open sourced a tool that reads pixelated text straight off a screenshot. It's called Depix. Here's the part nobody tells you. Pixelation isn't hiding your secret. It's just averaging it. Every little block is a math operation, and math can be run backwards. Depix does exactly that. → You give it the screenshot with the pixelated text → You tell it the font that was used → It renders every possible character, pixelates each one the same way, and matches them block by block → Out comes the original text The mosaic you trusted was never a wall. It was a thin curtain with the math printed on the back. A few things before you panic: → It cracks pixelation, the blocky mosaic, not blur → It works best on text from editors and screenshots, like a password someone censored in Notepad → It needs the font and a clean cutout of the blocks → Heavy compression on top can break it This isn't new either. That's what makes it worse. The repo hit 26,152 stars before the creator quietly pulled it down. The reason he built it in the first place: someone pixelated part of a password for an account with Domain Admin rights. He thought that little square of pixels was protecting it. It wasn't. So go ahead. Keep mosaicing your bank details, your API keys, your addresses before you post that screenshot. Or switch to a solid black box and stop handing people like me the answer key. CC-BY-4.0. Open source. Still runs.
Experience & Education
Licenses & Certifications
Other similar profiles
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More