Scott Wu, Cognition
David Senra
Jun 28
Scott Wu, Cognition
Scott Wu, Cognition
David Senra
Jun 28
Scott Wu, co-founder and CEO of Cognition, discusses his intense competitive drive, tracing it back to a childhood math competition loss. He applies this strategic, game-like thinking to building Devin, an AI software engineer, aiming to create a generational company that redefines human-computer interaction.
Wu explains that his motivation stems more from the fear of losing than the thrill of winning, a trait he shares with his mother. He envisions Devin as the ultimate interface, enabling universal software creation and shifting humanity from survival to creative mode. The company's model-neutral strategy dynamically selects the best AI for each subtask, focusing on real ROI like completed projects rather than token spend. Wu emphasizes that startups win by focusing narrowly, caring more than larger competitors. He is committed to building an independent, generational business, driven by passion over money, and would only sell if it were the most ambitious move. He accepts potential failure but not regret from not trying hard enough.
00:02
00:02
He views life and building a company as a strategic game.
02:06
02:06
Aiming to become world champion
06:38
06:38
Math trophies replaced family photos.
08:35
08:35
Fear of losing is a stronger motivator than winning
11:50
11:50
AI shifts humanity from survival mode to creative mode
12:55
12:55
Devin enables teams to ship 10 times faster and do 10 times more.
18:16
18:16
AI will autonomously execute complex tasks within five years
18:45
18:45
People will forget life without AI
22:57
22:57
Why couldn't it extend to days or months?
29:00
29:00
Successful people find value in unexpected places
30:18
30:18
A bold idea in late 2023
31:12
31:12
I couldn't sleep, envisioning a future where all software could be built by simply telling an AI what to do.
39:34
39:34
Most Fortune 500 companies are now software companies.
42:41
42:41
Enterprise cycles still typically take 12 to 18 months.
48:34
48:34
Measure real outcomes, not token spend.
50:05
50:05
Dynamically selects the best model for each subtask.
56:22
56:22
The world is not a single end-state.
57:17
57:17
Founders are rationally optimistic
1:00:27
1:00:27
Money is not my motivation.
1:03:51
1:03:51
I could accept failure, but not regret.
