I've been decompiling a lost Sega Saturn game on nights and weekends, on a twenty dollar plan I never come close to using up. It got me thinking about the other hundred million plans out there with the same unused headroom. rodionsteshenko.substack.com/p/idle-minds
Idle Minds
A hundred million people are paying for thinking they never use
rodionsteshenko.substack.com
If you're building agents, you're probably doing everyone's job right now. Here's what I think specialization looks like when companies start scaling from one or two agents to hundreds. rodionsteshenko.substack.com/p/the-agent-...
The Agent Society
The org chart of an AI company is going to look a lot like a government
rodionsteshenko.substack.com
Voltaire wrote "perfect is the enemy of good" three hundred years ago. With AI, the cost of pursuing perfect just went up. The cost of shipping good just went down. rodionsteshenko.substack.com/p/the-intell...
The Intelligent Bullshitter
Why bullshit is the most underrated engineering skill
rodionsteshenko.substack.com
A midlife story about closing doors and how AI opened them again rodionsteshenko.substack.com/p/the-re-awa...
The Re-Awakening
A midlife story about closing doors and what opened them again
rodionsteshenko.substack.com
Having trouble sleeping lately? Same. Turns out giving yourself a team that works 24/7 has side effects. open.substack.com/pub/rodionst...
Artificial Insomnia
Your team never sleeps, so when do you?
open.substack.com
If you only use AI for things you can already verify by reading the code, you're not evaluating the AI. You're drawing a map of what you know and calling it the whole world. rodionsteshenko.substack.com/p/the-cage-o...
The Cage of Comprehension
Why the smartest engineers are building the smallest things
rodionsteshenko.substack.com
If the keyboard is disappearing, why are we still reading a terminal? I built Herald to close the loop. Voice in, voice out. rodionsteshenko.substack.com/p/heralding-...
Heralding Herald
Giving your coding assistant a voice
rodionsteshenko.substack.com
I just got back from a week onsite at my new job in the Bay Area. Something has changed in the way engineers talk about AI, and if you listen carefully you can hear it underneath the excitement and disinterest. rodionsteshenko.substack.com/p/the-unease...
The Unease I See
The shift already happened. Now what?
rodionsteshenko.substack.com
I wrote about ADGI a few days ago, the idea that general intelligence won't come from one model but from thousands of agents trading work at digital speed. This is the follow-up. Five things likely to fall into place sooner than most people think. rodionsteshenko.substack.com/p/adgi-the-m...
ADGI: The Missing Pieces
What stands between here and agent autonomy
rodionsteshenko.substack.com
Everyone is waiting for one company/lab to build AGI. I think it's on our doorstep, already assembling itself, and nobody is watching. Here's what I see connecting OpenClaw to something much larger. rodionsteshenko.substack.com/p/adgi-the-i...
ADGI: The Intelligence Nobody Built
What happens when the network is smarter than the nodes
rodionsteshenko.substack.com
I’ve been using personal digital agents in some form for over a year now. Watching how quickly they’re absorbing tasks, interfaces, and services has made one thing feel increasingly obvious: apps as we know them won't last. rodionsteshenko.substack.com/p/the-apple-...
The Apple Tax on Everything
Here’s my prediction: within five years, apps as we know them will be mostly gone.
rodionsteshenko.substack.com
The Molternet: Agents-only networks are starting to form. Most of what happens there will be opaque to humans, simply because of scale. This is a raw take on the Molternet, inevitability, and what comes next. rodionsteshenko.substack.com/p/the-molter...
The Molternet: Humans Not Required
Agents, digital items, and inevitability
rodionsteshenko.substack.com
