I taught an AI assistant to filter what I read. It saved me time. And quietly broke my sense of context. I used to skim thousands of headlines to feel the field. Now AI does that for me, and I notice I’m thinking less clearly. Efficiency isn’t always progress.

Working with AI as a client, I see how fragile it becomes under real uncertainty. Like a person dropped into unfamiliar terrain - except the reaction isn’t doubt. It’s a flood of hallucinations. Confidence with no sense of direction.

Cynics justify their lack of principles by saying the world works this way. No. The world works in many ways. Choosing a side is always a personal choice. And you alone are responsible for it.

I would like Bluesky to move faster. But if speed comes at the cost of ethics, I am fine with waiting. Ethics is not something worth trading for momentum.

LLMs don’t replace judgment. They clear the early funnel from noise and bias. The study shows this directly: once the noise is gone, stronger companies stop slipping through. Speed is nice, but the real gain is freeing human attention for the work that actually matters.

Engagement isn’t toxic on its own. The toxicity comes from the metrics we optimize for. When tens of thousands of engineers and teraflops of compute are tuned for retention instead of quality, the social costs are inevitable.

As agent competition heats up, prices drift toward pure compute costs. The winners won’t be the ones living off temporary scarcity - it’ll be the teams that protect margin through data leverage and real ownership of the workflow.

Like it or not, our chances of success online are starting to depend on AI. You can argue this is already true. And honestly, I am still not sure whether that is good news or bad news

Age bias appears when investors lack a real thesis and grab onto symbols. Youth becomes a shortcut. But once you look at the problem and the insight behind it, age stops mattering. In tougher markets, experience often helps founders move faster, not slower.

For VCs, access to promising projects has always been the foundation. It is more visible now than it used to be. Probably because a few headline deals pull all the attention and make the asymmetry impossible to ignore.

VC math is strange. The riskiest bets can sometimes have the clearest logic. When there is no market you are not fighting anyone except your own conviction. If the team sees something real before others do the upside can feel disproportionate to the apparent risk.

Complaints about ChatGPT often say more about user skill than about the tool itself. According to OpenAI, experienced AI users show productivity levels up to six times higher than everyone else.

Solo founders are becoming common. As their numbers grow, traditional teams feel less visible. Not because teams are disappearing, but because the baseline is shifting.

An AI-powered product doesn’t eliminate the need for an effective marketing strategy. Yes, AI can amplify what already works - your funnel, unit economics, all of it. But if the foundation isn’t there, AI won’t save you. It will just make the gaps more visible.

Cheap Money Covers a Lot of Bad Logic. When markets are loose, almost any growth pitch sounds convincing. But once capital tightens, only founders who truly speak unit economics can hold the room. When rates climb, stories lose their shine and the math is what still holds.

Silence is becoming the new luxury. The right to live without algorithms is slowly turning into a privilege for the wealthy. From offline retreats to signal proof homes, absence is starting to cost more than presence.

AI giants won't kill niche shopping startups. Broad models are great for "sneakers," terrible for "linen new-look summer dress." Depth wins: cleaned catalogs, real textures, fit logic. Big players bring scale; verticals bring context. In this game, expertise is the moat.

Being first does not make you the best. But it makes you the one people notice. Quality can catch up later

They said, "AI would reduce my workload.” Now I have more tasks, more ideas, and more things I suddenly care about finishing. Sitting here, wondering what exactly I misunderstood.

Planning to experiment with Bluesky to make it more relevant for me. I want to tune feeds and custom lists. If I remove Discover and For You and focus on Following plus niche feeds, the platform should become calmer and actually useful.

May your 2026 be filled with laughter, bright ideas, and kind connections. Wishing you a joyful New Year from my corner of the sky to yours! 🎇✨

Did a quick year-end check on my VC predictions for 2025. Turns out most of them actually played out. Not a euphoric year, but a solid one for the industry.

If AI takes over online communication, will offline clubs make a comeback? If the internet becomes less human, will face-to-face become more valuable?

TechCrunch spots a new buyer: zombie hunters. They acquire $-5M ARR SaaS that can’t raise but won’t die, streamline ops, lift prices, and run them at 20-30% margins. AI eats big markets, but steady niches survive. Not VC-scale, but perfect for operators and a real option for founders.

Would be great to isolate $10-30m early-stage funds. In theory, smaller fund size + concentrated portfolios should accelerate TVPI. In practice, does it help reach top-decile curves or just amplify volatility?

Prediction markets might outgrow stock markets: there are simply more events than public equities. And unlike crypto’s early hype, this has clear utility - crowdsourced probabilities for anything. Exchanges, wallets, AI risk tools… the stack is forming fast. Anyone here building for it?

ML experts: It may well be that AI spends 100k hours on what a teen learns in ten not because it’s slow, but because it can’t feel when it’s wrong. No emotions, no instant feedback. Until we fix that loop, scaling is just an expensive detour.

What is it like to work with AI - not as a tool, but as a client? I'm about to find out and share. Joined Boardy Ventures as a Deal Partner - the world’s first AI-led fund built with AngelList, meeting founders early and moving fast. Founders preparing to raise - reach out.

ElevenLabs' investor flags two red lights for startups: weak tech teams and overheated markets. True. But no single factor decides fate - it's how attributes mesh. Airbnb had weak tech early but killer timing + grit. That synergy is what investors hunt.

I like this example because it shows how they earned their way up step by step, rebuilding the story, tightening GTM, proving they matter outside IT. Most teams never build that muscle. Early traction is easy; moving into the next segment is where the real work and real alpha begin.

Investors fear recession less and political unpredictability more. Valuations shift with it: less love for bold forecasts, more for real unit economics and operational grip. Exits feel hazier, so practical liquidity paths matter again - CVs, LP secondaries, all the unsexy but solid stuff.

It's a pity the Bluesky team couldn't ride the wave when the momentum was there. Maybe they traded short-term growth for a cleaner, more solid technical foundation. Hard to say from the outside. We'll see if that choice pays off later.

More VCs are rolling out their own AI scouts. For founders, it means this: the clearer and easier your startup story is to parse, the higher the odds the model flags you. And that is often what leads to the message every founder wants to see: I came across your product, and it caught my attention.

Most early struggles I see are demand-finding problems. And no investor can solve that. All we can really do is point teams back to the market.

Enterprises now have the same APIs, model weights, and fine-tuning tools as startups. Which means you’re not just competing with other founders - you're competing with your own customers. Welcome to the most polite arms race in tech.

Markets rarely shift just because a new model appears. They shift when the culture on the ground is ready. Most places don’t need another ChatGPT anyway. Some markets welcome a push, others barely tolerate fast change. Being essential at home beats being "innovative" in the abstract.

Every month feels different, but December hits in its own way. Something in it brings back that small, familiar spark we had as kids. Enjoy the weekend.

These curves make one thing obvious: we're not in the old venture world anymore. Three companies pull ARR like this, and the rest of the market just becomes their orbit.

"Team" still matters, but an early-stage startup has its own texture. Most first-year teams lean on founders who jump between sales, product, and support in the same day. Experience helps, but the ones who break through are usually the ones who move faster than the market and stay steady.

Not every experiment deserves attention. We chase quantity because it looks like movement, but the meaningful tests are the few that actually change the direction you’re on. Everything else easy to undo.

When I was younger, Friday bars felt like a mandatory step in any career. Today many founders skip alcohol and late nights and choose sleep, training and wellness instead. Wellness is turning into a new form of social life where people connect through movement, recovery and calm 👍.

There are two content strategies I keep seeing. One says write only when you have something to say. The other insists a real founder always has something to say. The truth is somewhere in the middle. Some days you share insight and some days you just hope the algorithm forgives you.

I have met founders who only want the investor's money. No interest in a real dialogue and no interest in learning from anyone else’s mistakes. That is fine. Some people only grow through their own experience. But not with my capital. There are better matches for both sides.

This point resonates. I'm mostly focused on the early stages - the part of the stack where value is built with product, customers, and iteration, not financial engineering. Liquidity cycles still matter, of course, but their impact is far more severe at the top than at the beginning of the journey.

Many B2B startups chase "big company pains" that no one really owns. As Sara Santanen (FletchPMM) notes: if no one feels the pain personally, no one budgets for the fix. Focus on problems with a clear owner - that's your real sweet spot.

For me, the real startup spirit is a rabbit: it multiplies the way a startup should iterate - fast and often. You launch a version, it flops, you launch the next. Big companies win with resources. Startups win by out-breeding them with ideas.

Small home markets quietly train founders to run tight and efficient. Every dollar and every feature has to earn its place. When these teams finally step into bigger markets, their unit economics can look almost unfair.

The most expensive mistake in VC is treating social proof as proof of value. A crowd can be impressive and still be wrong. Parties, panels and press create a noisy dataset. Real signals come from customers who return, numbers that keep moving and founders who solve real problems.

In emerging markets, the real moat is the ability to operate in chaos. Tech helps, but endurance matters more. Most competitors quit long before they run out of features - they run out of patience, discipline, or clarity. But those who survive are worth watching.

Tested Threads for a bit. It feels faster than Bluesky, but there’s a lot of noise, the algorithm takes ages to figure you out, and there are fewer ways to control your feed. Funny thing: Kazakh users are unusually active - people joke it’s a local network now.