Prioritize Builders Over Influencers

Prioritize Builders Over Influencers | Generated by AI

Home 2026.07

Question: For a one-person company (OPC) / solo AI company, is it better to build relationships with highly technical engineers rather than attention-seeking KOLs or conference personalities? Are technical friendships more likely to become long-term mutual-benefit relationships?

Answer:

Yes, for a technical OPC, your highest-leverage network is usually other builders: engineers, researchers, founders who still write code, train models, ship products, debug systems.

The reason is structural:

A person who has:

trained a model at 3 AM because CUDA OOM happened
debugged distributed training failures
shipped an agent that broke in production
optimized inference cost
dealt with customers

has a different communication layer.

You don’t need to explain why:

latency matters
GPU memory is the bottleneck
evals are hard
“just fine-tune it” is usually naive

The relationship compounds because both sides exchange real information.

Example:

You:
"I tried Qwen with vLLM + LoRA, but KV cache exploded at 64k context."

Another engineer:
"Try paged attention + quantized KV cache. Also your batch strategy is wrong."

That is a relationship with immediate value.

2. Attention economies create different incentives

Many KOL / conference personalities optimize for:

attention → followers → reputation → opportunities

Builders optimize for:

knowledge → products → revenue → capability

These overlap sometimes, but not always.

A person whose main asset is audience may naturally ask:

“How can this help my visibility?”

A builder often asks:

“Can we build something together?”

For an OPC, the second type is usually more valuable.

3. The strongest AI companies are built through technical networks

Look at many successful AI organizations:

early OpenAI came from a research community
DeepMind grew from ML researchers
many open-source projects grow from maintainers collaborating

The important network is not:

"I know 10,000 people"

but:

"I know 10 people who can each build something difficult"

A small number of high-agency technical people can create enormous leverage.

4. But don’t completely ignore public-facing people

There is a useful distinction:

Bad:

KOL = only distribution, no depth

Good:

Technical founder + communication ability

The best people combine:

deep technical ability
ability to explain
ability to connect people

For example, someone who writes excellent technical blogs, maintains open-source projects, and shares knowledge publicly is often a very good connection.

The signal is:

Do they produce artifacts?

Artifacts:

GitHub repositories
papers
benchmarks
products
datasets
technical writing

are stronger signals than:

follower count
conference photos
hype posts

5. For your situation (AI OPC), I would prioritize

Tier 1:
AI engineers
ML researchers
open-source maintainers
startup technical founders

Tier 2:
technical product people
domain experts who can bring problems

Tier 3:
investors, KOLs, media people

The first group helps you create capability.

The second helps you find markets.

The third helps after you have something valuable.

A good rule:

Don’t network around status. Network around shared missions.

A person building an AI agent framework, training models, or deploying systems may become a 10-year collaborator.

A person only chasing the next trend topic may disappear when the trend changes.

For a high-tech OPC, a small circle of extremely capable engineers is often a stronger moat than a large social following. Your own path (training GPT-2, building agents, running an AI community, writing technical notes) naturally attracts that kind of person.

Back Donate