internet 3.0: agents, empty gardens and the software boom

4 min read Original article ↗

The age of agents is here, not bots, we needed a fancier term - just like we had with “cloud” computing and big data filling data lakes. Feedback loops with directions trying to make progress and push the GDP forward.

Internet used to be the wild west, you could do anything, nobody cared, then more people came here and so civilization was born, the walled gardens to preserve civility for everyone against the few. Now the spaces have been invaded with soulless machines, and Im not talking about Linkedin users, but bots.

Bots have always been a problem on the web, but they were easily distinguishable - at least for a locked in user, now there’s no way of knowing who is real so everyone is looking for ways to scan your retina when you access the web. Governments of course want to rely on their digital IDs but that’s harder to sell.

Now, back to agents, the new favorite (or a nighmarish) KPI of every PM who has heard about agentic workflows in the last months. Not quite human, not quite just a tool, but rather than an extension of the real user - the metric that matters.

The internet is going to evolve to be for agents, the human is getting removed, the web environment and the systems will adapt to the new users and so, in the age of agents, tools will be made for them, whole ecosystems - more or less what we have for humans now but instead it will be much faster, much more context aware - and eventually a closed ecosystem where no human can understand what’s happening, probably internet 4.0.

Currently we are in the middle of the transition, from internet 2.0 into 3.0, with agents becoming autonomous - however with this type of technology what happens is at first there’s going to be a slow adoption and then all of the sudden where old people are confused about what to do, and by old people I mean you reading this.

The most interesting development and thesis to see will not be the agents themselves but the tooling around it, we are already seeing how companies are using MCP and stable coins to achieve certain agent tooling - still far away from the public adoption but it will come.

Positioning yourself for the age of agents, and its tools, is going to be key for the companies who want to make it, survive, or begin. The age of agents is also coming with a software boom, just like the baby boom.

The age of doers, not talkers, the age of builders and not observers.

X avatar for @tekbog

terminally onλine εngineer@tekbog

vibe coding will do for software what portable cameras have done for cinema

8:15 PM · Mar 22, 2026 · 189K Views

127 Replies · 160 Reposts · 3.48K Likes

But as with any gatekeeping break of any medium, we are about to see some interesting outcomes.

X avatar for @MeekMill

MeekMill@MeekMill

I need a vc

11:27 PM · Mar 25, 2026 · 291K Views

496 Replies · 213 Reposts · 3.13K Likes

When software engineering is no longer a bottleneck what’s going to happen? Probably what happens everywhere, just a billion new products, most of them bad, just trying to make money, some really good and will improve industries or cover human necessities we didn’t know existed - and some will be horrible brainrotting us just like tiktoks and reels, but I’m sure they will make money.

Building software and medium intelligence cost is going to zero, exciting times, and yet everyone is trying to build the same thing to stay ahead, the universal computer, let’s see how it goes.
If “computer” is solved, what’s your moat? Many are trying to move to the physical world, atoms if you may, but software and AI will still operate the physical world, so is it truly the best move?

Discussion about this post

Ready for more?