I I test ideas, challenge assumptions, and follow the evidence.
Currently exploring:
• Growing on X without Premium
• AI vs real-world execution
• Startup experiments and why ideas fail
Start here ↓
Everyone said Google was slow.
Last 30 days:
Gemini 4 Argon
Genie 4
Gemini 4 Carbon
Gemini Omni 1.2 Flash
Gemini Decision 1
Gemini 4 Hydrogen
Gemini 4 Neon
7 models in one month.
Now rumours say Gemini 4 Sodium is next.
They’re speedrunning the entire fucking periodic table.
It’s 2030. Inference is absurdly fast.
The real advantage isn’t more tokens.
It’s who wastes less silicon, power, and memory.
Utilization just ate intelligence.
Two founders. Same dying traffic.
A blocks AI → protects content → becomes invisible.
B feeds AI → gets exposure → gives the farm away.
Who actually survives?
Software engineer vs AI coding agents.
US developer: $135,980/year median salary.
AI coding tools: $92/month average user spend.
123× difference.
But tools aren’t employees.
The real cost depends on supervision, errors and completed work.
Simplicity > Complexity.
Share a link.
Choose an available slot.
Confirm automatically.
One annoying problem. One simple solution.
Calendly: $3 billion valuation.
SILICON VALLEY MIGHT BE THE WORST PLACE TO BUILD A STARTUP.
Investors over customers.
Networking over selling.
Pitch decks over revenue.
Posting over building.
A new generation of digitally native, remote-first founders is coming.