The Reality of AI Productivity Gains: 2-3x Not 10x
Why 2-3x productivity gains from AI are more realistic than 10x claims - and why that's still transformational for businesses
I know the industry as a whole is enamored with '10x.' What I'm seeing personally right now is more like 2-3x.
Meaning: inside more mature systems with paying customers, adding effective AI workflows at all the right stages of the product development lifecycle can 2-3x throughput (the finished work that reaches customers).
If you're lucky + good with what you're shipping, that can mean 2-3x value created.
It's not 10x (which you can get at some stages), but 2-3x is still really good! Like meaningfully change your business good if you're set up to capitalize on it and aren't just moving the bottleneck for value creation + capture somewhere else.
Related insights
Almost $10K in Tokens in April
15 billion tokens, 302 Linear issues shipped to production, just me in Claude Code. Real 'headcount equivalent' territory—and a reminder to measure impact, not just output.
WIP Limits Still Matter in an AI-Powered World
My workflow makes it easy to spin up 10+ parallel agent threads, but 5 or 6 well-sized tasks is the sweet spot - for my sanity and to keep the final review pile manageable.
4+ Billion Tokens in February
252 PRs merged, $2700 at list price for my $200 max plan - continuing to find the edge of what's possible within sane guardrails
Get More Like This
Follow along as I build and share what I learn
Found this helpful? Share it with your network!