Denis Gramm's Linkedin Analytics

48 hours to build a functional app for a $750/month client. LFG Got a new order. A new client wants to monitor how their listings perform against competitors on a major travel site. Claude 4 just dropped, so instead of handing this to a coder, I’ve decided to 100% vibe code it MYSELF this weekend. So I’ve set the price at $750/month, 3-month commitment. They said yes. Now I need to show a working prototype by Monday. Going with Rails 7 + Tailwind. Features: daily scraping. Data comparisons. CSV exports. A bunch of other things. Back in my coding days, this would’ve taken me ~4 days of grinding. Can I 100% vibe code this in time? We’ll see.

I didn’t think AI avatars for videos would work. SO WRONG. Thought it’d look too cringe. BUT I just made an onboarding video for Botster using one - and I’m blown away. Clean. Fast. SCALABLE. Zero setup, zero gear. I don't say this often, but yes - this changes everything.

Complexity is the new advantage. AI can copy anything: Your features? Cloned. Your landing page? Stolen. Your script? Rewritten in 10 seconds - and somehow better. So what can’t it copy YET? - Ecosystems - Cross-platform flows - Real-world integrations - Supply chain ops - Community + network effects - Hybrid physical + digital models - Built-in distribution channels Our goal? Trigger the “NAH, TOO HARD” reaction in that dude in a student dorm scrolling your site for business ideas to rip. You want protection? Build a SYSTEM - not a tool. Make it painful to reverse-engineer. Simple gets copied. Complex survives.

Curious what viral content on LinkedIn actually brings? Here are the numbers: I wrote two posts in a row that went viral (350,000+ views). What did it get me? - 100+ new followers / connection requests - 60 website visits (yes, SIXTY) - Maybe a couple of sales - hard to say now, but there was some activity THAT'S IT The content had broad appeal, so the conversion rate was basically 0.00001%. Is it worth it as a business strategy? For a SaaS - definitely not. It's a distraction. You can’t consistently hit 300K views if you’re writing with a sales goal in mind. For someone selling custom services - maybe yes. One client can pay well, even if the reach is smaller. Will I keep going? Yes - for myself. I will keep posting my thoughts out loud without overthinking each post and expecting every other post to blow up. Virality is nice, but the next day your reach is back to 1000 views. As long as you treat social media like a side quest - not your core strategy - there’s no disappointment. Because these numbers don’t even come close to what SEO or YouTube can do. Viral content is a dopamine hit. Not a business strategy.

Crazy, but I actually did it. A full-fledged web app, delivered and demoed in under 48 hours. Screenshots in comments. - Daily data scraping - CSV exports - Scraping calendar - Data comparison And yes - it looks great and is mobile-friendly. Casually vibe coded. 99% written by AI. 1% me steering the ship, suggesting high-level decisions, and telling Claude to slow down when needed. Scraping offloaded to Botster .io’s no-code scraper. Sends me a JSON when done and I parse it into objects. Easy. Feels surreal tbh. But I guess this is our life now.

AI DESTROYED every excuse we had as developers. Remember how you used to hide behind "aligning divs" for 8 hours straight? Pretending you were working, and then - once the divs conformed - going home with a sense of accomplishment while in reality you moved nothing forward? I remember. Those days are OVER. AI aligns your divs in 1 second. AI builds your homepage in 5 minutes. AI does all the busywork you used to hide behind. Now there's nowhere to run. No excuses. No fake productivity. No more lying to yourself about "design tweaks." The REAL work is staring you in the face. And most are TERRIFIED. Because now everyone can see what you're actually capable of.

AI can’t help you if you can’t be helped. I know a guy who’s always full of business ideas. Never acts on them - always has a “good” reason why “now’s not the time.” Tried learning to code. Gave up after a day. Tried building something with nocode. Gave up after a week. Now it’s: “With AI, I can create anything!” NO YOU CAN’T. Because you never did. Still not building. Still giving advice. People don’t get this but it is NEVER about the tools.

I’m on the couch editing a video. Wife’s on the other couch filming a reel. A modern family. Making content in 2025 is a chore. If you’re not doing it - you're ngmi.

I’m ending this month with 400,000+ post views - first time ever. No paid coaches. No LinkedIn courses. Just wrote what I was thinking and hit post. And my impression? meh. People who tell you “virality = leads” are usually selling you a course, an ebook, or their marketing services. Don’t buy into it. Virality is fleeting. Social media’s not a sales channel, it’s a dopamine channel. You want customers? Focus on channels that scale and work while you sleep. SEO. YouTube. Systems that bring real traffic, real buyers. Don’t get distracted chasing impressions. Chase impact.

Web scraping is an arms race. Webmasters build defenses. Scrapers build smarter bots. If you’re in this circus - you already know. If you’re not, here’s the warning: We’ve got it good right now. It will never be this easy again. Things don’t get easier. They stay the same - or get more locked down. Access shrinks. The open web is quietly closing. So if you’re planning to build something with data? Do it now. Later it’ll cost more, take longer - or just won’t be possible at all.

My SaaS revenue is down. But custom projects are on track to hit an annual record. What’s going on? You already know the answer. People with even a tiny bit of technical sense are now thinking: “Why pay for this? I’ll just build it myself with GPT.” They’re tinkering, prompting, hacking things together with a goal of avoiding a $25/month payment But the people who never heard of automation before? Now they hear about it EVERYWHERE. TikTok, YouTube, LinkedIn, podcasts, ads. And they all have the same thought: “I need to automate something. NOW” BUT They’re not Googling. They’re not trying it themselves. They’re coming straight to ME. The result? AI is killing off small-tier solutions. But it’s exploding demand for custom, “done-for-you” builds. While some are trying to save $29/month… Others are happily paying $5K just to not think about it.

A big chunk of my readers are freelancers. I freelanced for over 5 years -- made good money, friendships, traveled the world nomad-style, all while learning new tech. Here’s one piece of advice I wish I had back then: STOP selling one-off projects. Start PRODUCTIZING early. Sell packages of recurring services. Sell subscriptions. This isn’t just a cashflow fix. It’s your path to scaling without working more hours. Want freedom? You need predictable revenue. You need to stop chasing new clients every month. You need systems that sell while you sleep. Trust me - I’ve been there. This shift changes everything. If you need advice, my DMs are open.

One of my freelancers asked to double his pay. “Why?” “AI makes it twice as fast to write code now.” His logic: He’s working half as long, delivering the same result - so he wants double the money. My logic: Bro discovered a calculator and demanded a raise because he doesn’t have to count on his fingers anymore. He thinks I owe him for using a tool that’s accessible to anyone. I first offered a 10% pay increase because finding and onboarding someone new is a pain. But he insisted on double. So… I replaced him with someone who uses AI AND charges the original rate THE SAME DAY. Using ai doesn’t make you special anymore, everyone has access to it. It’s a commodity, just like a calculator.

How I JUST closed a $7000 dev project Took me 5 steps and less than 24 hours - On a call, client casually mentions their CEO MANUALLY checks competitor bookings. - I drop an idea: “What if you had a dashboard tracking that in real time?” - Client: “That’s actually... super interesting.” - Same night, I mock up a prototype with v0. Takes me 10 mins, one screenshot, sent it over. - Client: “How fast can you build this?” Me: “10 days.” Client: “Deal.” Speed wins every time. Pretty sure the client still doesn’t know what just happened.

There’s a growing wave of a new type of SEO software: “Track how your site ranks across ChatGPT, Perplexity,” etc. This HAS to be useful, right? Just like tracking 1,000 keywords on SEMrush back in the day, founders are now racing to be the first at TRACKING PROMPTS across dozens of LLMs. But there’s a TINY problem: AI doesn’t rank consistently across prompts or models. Slightest change in wording, context or user avatar? Whole new answer. You “rank” now, vanish in an hour, and still get ZERO traffic either way. OK, let’s say you “rank” 3rd for a prompt. Now what? You don’t know. The app maker doesn’t know. Nobody knows what it actually means - but everyone’s excited to track it. If you’ve got $150/month to burn to watch your site “rank” in hallucinations - be my guest, I’d rather buy lunch, at least I know what I’m getting. “But how do I dominate ChatGPT answers bro?” EASY Keep building content people search for. Own your niche. Create something worth mentioning - with or without a prompt.

The biggest threat to my SaaS? Basic computer literacy. Got this email: “I was planning on scraping phone numbers… but then I wrote the script myself.” TRANSLATION: I don’t need your tool. I’ve got a browser and a brain. That’s the real danger. At this rate selling tech alone won’t buy us Balenciagas anymore. HORROR. So what do we need to sell? Speed. Simplicity. Status. Solving complex problems faster than they can Google it. Because if all you offer is code, you’re one prompt away from being obsolete.

The thing that surprises me most about AI-assisted development: Setting up something complex like data scraping of a large website is a breeze. But fixing padding? Absolutely impossible. 10 minutes in, I’m still arguing with the AI about margins. Some things never change.

Trying to stop web scrapers is like locking your front door and leaving the windows wide open. You’re not stopping anything. You’re just wasting everyone’s time, including your own. Doesn’t matter how many CAPTCHAs, rate limits, or JavaScript puzzles you throw in - if it’s on the internet, it’s getting scraped. "Yeah but ..." NO. Make an API. Put your products in a CSV and link it in the footer. Offer to sell your data - and people WILL buy. Collaborate like a professional. Don't fight visibility. Profit from it.