AI’s “Dog Years” Era: Why Progress Feels Faster Than Ever

amy 22/08/2025

Opening Hook:
“One day in AI today equals a year of progress a decade ago.” That’s how Zhou Hongyi (360 Founder) describes the breakneck pace of AI development—a Cambrian explosion of innovation where yesterday’s breakthroughs are today’s benchmarks. But can his new venture, Nami AI Search, replicate 360’s disruptive impact? Here’s why the game has changed.

1. Then vs. Now: Disruption in Different Eras

2008 (360’s Rise):

  • Low-Hanging Fruit: Antivirus software was “so bad that even average work stood out.”
  • First-Mover Edge: Free antivirus was a “bull in a china shop” move—unrefined but revolutionary.
  • Less Competition: Big tech dismissed startups as “toys.”

2024 (AI’s Reality):

  • No Blind Spots: Every tech giant (Google, OpenAI, DeepSeek) is all-in, spending billions.
  • “One-Week Lifespans”: AI products lead for days before being matched (e.g., new agent frameworks).
  • No Room for Error: Startups must out-innovate and out-execute—simultaneously.

Key Quote:
“Back then, disruption was about guts. Today, it’s about stamina.”

2. The “Dog Years” Dilemma

AI’s timeline is compressed:

  • Pre-2023: Slow burn (decades of research).
  • Post-LLMs: Exponential leaps. Daily breakthroughs now rival annual progress pre-2020.

Example:
Nami AI’s agent might lead today—but by next week, OpenAI/DeepMind could release a rival with identical capabilities.

Data Point:
Zhou estimates AI’s full impact will take 5–10 years, but the rate of change is unprecedented.

3. Why This “Golden Age” Is Different

  • No Underdogs: Unlike early internet days, no major player is asleep. Microsoft, Apple, and Tencent are hyper-aggressive.
  • Open-Source Arms Race: Tools like Meta’s Llama lower entry barriers—but also accelerate commoditization.
  • China’s Role: With AI now a geopolitical battleground, startups face pressure to align with national tech sovereignty goals.

Zhou’s Edge:
His experience with 360’s battles (e.g., vs. Tencent) taught him how to pivot fast—a survival skill in AI’s “iterate or die” landscape.

4. A Historian’s Perspective: Industrial Revolutions Revisited

  • First (Steam): ~150 years.
  • Second (Electricity): ~100 years.
  • Third (Digital): ~50 years (1960s–today).
  • Fourth (AI): Projected 10 years to maturity—but with 10x the yearly output.

Implication:
AI’s “early stage” is already more disruptive than the internet’s first decade.

Final Thought: The Anti-FOMO Mindset

Zhou’s reflection—“You can’t be purely opportunistic”—hints at a deeper truth:In AI’s “dog years,” sustainable impact requires:

  1. Long-Term Patience (e.g., Nami’s 10-year horizon).
  2. Short-Term Agility (e.g., weekly product updates).
  3. Ethical Guardrails (avoiding “move fast, break things” pitfalls).

Question to Readers:
Is AI’s speed a blessing or a curse for startups? Can anyone “outrun” the giants anymore?

Related articles