Telegram relatively recently added a feature that lets bots interact with each other.
Cool idea, but I haven't seen a practical and effective implementation anywhere yet.
🐻 An example of a good implementation is integrating staking into a mobile neobank. The rest of the implementations I've come across so far are just API or someone's SDK integrations without any know-how.
🔢 Also, some AI-based chatbots have added the ability to download content from social media with transcription of audio or video content.
Fun fact: I've got relatives in England, and I actually speak with a British accent. Because of that, half the people in Europe don't understand me — I speak English too fast 🙈
I'm actually thinking about taking some extra lessons: brushing up on the German I studied at school, and the French I learned in lycée.
I feel like the bare minimum — speaking three foreign languages — is almost as essential these days as knowing your multiplication table.
Especially for a man.
It's like cooking, which, in my opinion, you should master with the same finesse as running a business. There's a good saying: how you do anything is how you do everything.
So, in my humble subjective view, whatever you're truly passionate about, you should do first-class.
People outside the VC industry usually get all starry-eyed about it after hearing tales of someone tossing in a couple of bucks and ending up with a billion-dollar company, making the investor unfathomably rich.
Let's set aside the microscopic odds of actually building a business like that and just look at how much money you have to pour into the system to get that kind of scale on the other end. Turns out — no magic.
Per the Crunchbase Unicorn Board as of July 2026, today's private unicorns have collectively raised $1.52T, with a combined valuation of $8.7T.
The math is simple: every dollar invested turned into roughly $5.7 of paper valuation. Not profit. Not cash in the investor's account. Not a guaranteed exit. Just paper value based on the latest round.
If you roughly spread that paper multiple across 7–10 years, you get about 19–28% annually.
That's a great result for large-scale capital. But it's no x1000.
🎁 The average unicorn isn't a "tossed in a couple of bucks" story. It's the story of dumping around $175–180M into a company's furnace and waiting 7+ years. Want a $10B valuation? Be ready to raise, on average, about $1.7B.
VC only looks like magic in a handful of legendary early-stage cases that get sold to the wider public. In bulk, it's just an expensive, slow, ultra-risky game with a solid but very down-to-earth return at the end.
What was shown at Y Combinator Demo Day (June 2026) 🔖
If you keep an eye on the IT market and AI startups in particular. But you don't have time to hop on crunchbase, peerlist or producthunt to check out the current trends. Then this roundup, with promising directions, is for you (save it):
— Adialante mobile MRI clinics for early cancer detection. A compact MRI machine that fits in a small truck.
— Sazabi AI for finding and fixing bugs in production. Analyzes logs, pinpoints the root cause of the failure and helps ship a fix right away.
— Ploy website marketing and growth automation. Generates landing pages, copy and campaigns, while AI agents continuously optimize the content.
— Lightsprint building features without writing code. The PM describes the task, AI puts together a solution, and the engineer just reviews and merges.
— Dispatch shipping cargo from space to Earth. Building reusable capsules for returning goods manufactured in space.
— Complir compliance for physical goods. AI handles regulations, risks, documentation and labeling for international shipments.
The main trend I'm seeing: AI is increasingly being sold not as a "helper," but as a tool that takes over a specific business process turnkey.
WhatsApp has nearly caught up with Telegram's decade-old feature set. You can already create channels and buy a Premium subscription. And now you can even add a username to your profile…