Marcos Polanco, CTO

I help coding agents fight architectural entropy • Founder, @Prelatent • Kauffman Fellow • ☸ • 🇵🇷 • ⛷ • 🏊🏾 • 🧘🏿‍♂️ • implacable • L∃∀N

[Marcos Polanco, CTO @MarcosPolanco]

"Velocity is high, but we are afraid to touch the code."

If that sentence just described your software project, I built a 3-minute diagnostic for that.

Do your coding agents understand your business?

https://t.co/xnD2OSpFaV

[Marcos Polanco, CTO @MarcosPolanco]

What is the view? https://t.co/qpvDnsZEGJ

[Marcos Polanco, CTO @MarcosPolanco]

@StephNass playbook

[Marcos Polanco, CTO @MarcosPolanco]

@lucaspatiri_ UGC

[Marcos Polanco, CTO @MarcosPolanco]

A great explanation of when to use microservices; TLDR: it is all about stream-aligned teams or scalability constraints. Résumé-driven architecture should perish. https://t.co/rX9dkcxKPx

[Marcos Polanco, CTO @MarcosPolanco]

@geoffreywoo Friction for @MartinGTobias

[Marcos Polanco, CTO @MarcosPolanco]

@colinsolvely @nnennahacks @iamKierraD It is also misguided; when it comes to building, we know AI can lay bricks. We don't know it can do the structural engineering or the architecture. https://t.co/U0LjICu3bh

[Marcos Polanco, CTO @MarcosPolanco]

Know that leaning tower in San Francisco?
Let’s talk about your software architecture.

Coding agents are brilliant at adding more floors to your gleaming skyscraper. But my structural integrity analytical framework showed something striking: They put more and more load on just a few components of the system. They present a façade of modularity while building coupling deep below the surface.

I measured this.

I cloned ~15 repos where AI made its appearance and measured structural coherence before and after. For those in the back row, yes, that’s a longitudinal study.

Most metrics flatlined across AI adoption. One went to 11: Hub concentration. That means some modules were becoming, like Malcolm Gladwell would put it, super connectors. More and more weight on just a few pylons. Until they start cracking.

Good luck evolving your code at velocity if you let this go too far for too long. Your velocity remains high until. It stops.

The paper is in progress, but you don’t have to wait. I’m opening a closed beta working directly with CTOs and engineering leaders to analyze their most important business asset, aka their codebase, using cutting-edge research instrumentation.

Because your codebase is the operational future of your company.

Link in thread.

[Marcos Polanco, CTO @MarcosPolanco]

@Geramozg Max, game to pioneer; let's chat https://t.co/iv0ri4d6Fd

[Marcos Polanco, CTO @MarcosPolanco]

@NelsonXLee DMs open. https://t.co/iv0ri4d6Fd

[Marcos Polanco, CTO @MarcosPolanco]

Join the Architectural Coherence Index Beta: https://t.co/ZTt63O9rMD

[Marcos Polanco, CTO @MarcosPolanco]

Know that leaning tower in San Francisco?
Let’s talk about your software architecture.

Coding agents are brilliant at adding more floors to your gleaming skyscraper. But my structural integrity analytical framework showed something striking: They put more and more load on just a few components of the system. They present a façade of modularity while building coupling deep below the surface.

I measured this.

I cloned ~15 repos where AI made its appearance and measured structural coherence before and after. For those in the back row, yes, that’s a longitudinal study.

Most metrics flatlined across AI adoption. One went to 11: Hub concentration. That means some modules were becoming, like Malcolm Gladwell would put it, super connectors. More and more weight on just a few pylons. Until they start cracking.

Good luck evolving your code at velocity if you let this go too far for too long. Your velocity remains high until. It stops.

The paper is in progress, but you don’t have to wait. I’m opening a closed beta working directly with CTOs and engineering leaders to analyze their most important business asset, aka their codebase, using cutting-edge research instrumentation.

Because your codebase is the operational future of your company.

Link in thread.

[Marcos Polanco, CTO @MarcosPolanco]

@CharlesD353 ...anymore?

[Marcos Polanco, CTO @MarcosPolanco]

@NehraWorkss Do you have the prompts that created the app? That's a goldmine of *intent*. Plus the codebase lets your re-specify and re-generate the a clean, scalable version that can serve as a drop-in replacement.

[Marcos Polanco, CTO @MarcosPolanco]

Best article I've read this week: https://t.co/QSgkos1xub

[Marcos Polanco, CTO @MarcosPolanco]

"I stopped screening for frameworks, languages, or even algorithms 5-7 years ago."

Leslie de Jesus has recruited hundreds of engineers over 25 years. Now, she says coding is no longer the bottleneck she optimizes for.

Here is where the bottleneck moved: https://t.co/vagNbz4CZQ

[Felipe Demartini @namcios]

A Anthropic acabou de matar o Markdown.

Um engenheiro do Claude Code publicou um artigo ontem que pode decretar o início de uma nova era.

A tese é brutal: Markdown nunca foi o formato certo para comunicação entre humanos e IA. Era só o que tínhamos.

O próprio autor admite que nunca leu um arquivo Markdown gerado por IA com mais de 100 linhas até o fim.

Você também não lê. Eu também não.

A sacada:

Markdown assume que você vai ler do início ao fim.
HTML assume que você quer ver o que importa e mexer com as mãos.

Na prática:

→ 30 tickets de projeto viram kanban arrastável com colunas Now / Next / Later / Cut e botão de exportar
→ Lógica de rate limiting vira flowchart SVG com código inline, no lugar de 200 linhas de texto
→ Code review vira diff colorizado com grafos de dependência entre módulos
→ Parâmetros de animação, cores, regex, cron jobs ganham sliders com preview ao vivo
→ Specs de projeto viram 6 opções lado a lado com mockups interativos

Todos exemplos reais do artigo. Todos substituem um muro de texto por algo que você de fato abre e usa.

O trade-off existe: HTML é 2-4x mais lento para gerar. Mas com contexto de 1 milhão de tokens, esse custo sumiu.

E a parte que ninguém está discutindo: o HTML gerado não é só para humanos. O agente de verificação também lê. O spec deixou de ser documento e virou memória compartilhada entre agentes.

Markdown é relatório.
HTML é interface.

Relatórios são para ler.
Interfaces são para continuar o trabalho.

Se você usa IA em 2026 e ainda pede Markdown para tudo, você pode estar usando um smartphone como lanterna.

[Marcos Polanco, CTO @MarcosPolanco]

@PavanKumarNY @sama Building https://t.co/X7s9ePnCib, applying breakthrough spectral graph theory to offer early warning of adversarial coordination (think anti-money laundering & fraud) on transaction networks, starting with blockchains.

Me: patent holder in ML; 2x founder; eng leader.

[Marcos Polanco, CTO @MarcosPolanco]

@boardyai @mayornuell @lalazar Did someone say high ownership?