Sabitlenmiş Tweet
I thought I was paying attention to AI.
Then I actually looked at the data from the last 3 months.
I wasn't paying attention. I was watching the surface while the ground shifted underneath.
Here's what's actually happening:
AI adoption hit 39% of US adults in two years.
For comparison:
→ PCs took 12 years to reach that level
→ The internet took 5 years
→ AI did it in 2
This is the fastest-spreading technology in recorded history. And most people are treating it like a new app to try.
The capabilities are moving faster than the headlines.
In one year:
→ AI's ability to solve real coding problems jumped from 4% to 72%
→ Inference costs dropped 280x
→ Model efficiency improved 142x (same performance, fraction of the size)
Sam Altman said AI costs are falling 10x per year. Faster than Moore's Law ever did.
The workforce data is where it gets uncomfortable.
26% of US workers now use AI on the job. But here's the part nobody talks about:
→ 45-49% of them are hiding it from their managers
→ 52% are reluctant to admit using it for important tasks
→ 78% bring their own AI tools to work without employer direction
Half the workforce is already adapting. Quietly. Without permission.
The job shifts are happening underneath people.
→ Klarna cut 40% of its workforce. AI handles 2.3 million customer chats per month.
→ Duolingo replaced translators with GPT-4 and declared itself "AI-first"
→ Entry-level job postings dropped 35% since 2023
→ Freelance writing gigs fell 28%. Graphic design fell 33%.
These aren't predictions. These are last quarter's numbers.
41% of companies plan to reduce their workforce by 2030 due to AI.
But 77% of those same companies plan to reskill existing workers.
The net projection: 170 million jobs created, 92 million displaced.
The math is positive. The distribution won't be.
I keep thinking about this stat:
ChatGPT reached 100 million users in 2 months.
→ TikTok took 9 months
→ Instagram took 2.5 years
By early 2025, ChatGPT had 700 million users.
Over 1.2 billion people used AI tools within 3 years of launch.
And most conversations I hear still start with "I should probably try that AI thing."
The gap isn't between people who use AI and people who don't.
It's between people who understand what's happening and people who think they have time to figure it out.
Anthropic's CEO thinks AI will be better than humans at "almost everything" within 2-3 years.
Google DeepMind's CEO moved his AGI timeline from "10 years" to "3-5 years" — in the span of 3 months.
These aren't random Twitter predictions. These are the people building the systems.
I'm not saying the sky is falling.
I'm saying the data doesn't match the vibe.
Half the workforce is already adapting, mostly in secret.
And the tools they're using today are the worst versions they'll ever use again.
That's not fear-mongering. That's just the math.
English
