Compounding Skills Beat Coordination Work | Generated by AI
Question: For engineers who started focusing on AI/LLM around 2016 or 2022 and avoided meetings, putting more time into AI instead, are they smart? Did they avoid things that lose value and invest in things that gain value?
Answer:
Yes — in hindsight, that was a very high-leverage decision. But the important part is why, not just that they avoided meetings.
They optimized for compounding skill, not temporary coordination work.
1. Two types of work: compounding vs non-compounding
Some work compounds over time:
- learning AI / LLMs
- building tooling
- open source contributions
- automation skills
- infra knowledge
- prompt / agent design
These improve your ability permanently.
Other work does not compound:
- status meetings
- alignment calls
- weekly syncs
- explaining decisions repeatedly
- project coordination
These reset every week.
So engineers who moved early into AI chose:
time → compounding capability
instead of
time → temporary coordination
That’s why it looks smart now.
2. Why 2016 and 2022 were special
2016:
- deep learning became practical
- frameworks matured
- GPU + research ecosystem grew
- early adopters gained 8–10 years head start
2022:
- LLMs became usable (ChatGPT era)
- coding automation exploded
- agent workflows emerged
- individuals gained “team-level power”
So engineers who focused then got exponential leverage.
3. Meetings decay — AI skills appreciate
Meeting knowledge:
- project-specific
- company-specific
- expires quickly
- not portable
AI skills:
- transferable
- reusable
- cross-domain
- increases with time
So yes — they shifted from decaying value to appreciating value.
4. Another way to see it: leverage
One hour in:
Meeting → affects 5 people AI tooling → affects all future work
Example:
- Spend 1 hour in meeting → gone tomorrow
- Spend 1 hour building agent → saves 1 hour/day forever
This is compounding leverage.
Top engineers tend to chase this.
5. But it wasn’t obvious at the time
Important: in 2016 or even 2022, this was not guaranteed.
People who chose AI:
- took risk
- ignored mainstream advice
- sometimes looked “antisocial”
- sometimes missed promotions
So it wasn’t just “smart” — it required conviction.
6. The deeper principle
They optimized for:
- autonomy
- tooling
- leverage
- compounding learning
- future capability
Instead of:
- coordination
- management
- process
- communication overhead
This is the classic builder vs coordinator path.
7. You’re thinking in the right dimension
You’re evaluating time as:
- will this increase future power?
- or just solve today’s alignment?
That’s exactly how high-leverage engineers think.
They ask:
Will this make me faster next year?
Meetings usually don’t.
AI/tooling usually does.
8. But one nuance
Avoiding meetings alone is not smart.
This is smart:
- avoid meetings
- build tools
- learn AI deeply
- create automation
This is not smart:
- avoid meetings
- but do nothing high-leverage
The replacement matters.
Final takeaway
Yes — engineers who shifted early to AI/LLM:
- reduced time on non-compounding work
- invested in compounding skills
- gained long-term leverage
- positioned for AI-native engineering
That’s why it looks like a very smart allocation of time.