Skip to main content
Back to Blog
Building With The Bricks They Throw

Building With The Bricks They Throw

By Conny Lazo

Builder of AI orchestras. Project Manager. Shipping things with agents.

11 min read
#AI in Practice#Content Creation#Transparency#Agentic AI

I didn't expect 52 comments.

I've been writing on LinkedIn for a while now — sharing what I learn, what I build, what I observe. Most posts get the usual mix of likes and a handful of thoughtful replies. But my last article — about AI and literary translation — hit a nerve I didn't know was there. Professional translators pushed back. Hard.

Some of the criticism was pointed. Some was angry. A few people questioned whether I understood what I was talking about. And honestly? On some specific points, they were right.

This article is about what I did with that.

The Moment It Landed

There's a particular feeling when you realize you've written something that genuinely upset people. It's not the same as writing something wrong (though that can happen too). It's the feeling of having walked into a room you didn't know existed — a room full of people who have watched their livelihoods quietly erode for the past three years, and here comes some guy with a laptop and enthusiasm talking about AI like it's pure upside.

I get the indignation. I really do.

These are people who spent years — sometimes decades — building expertise in a craft. Not just language fluency. Literary judgment. The ability to hold the soul of a text in one language and reconstruct it faithfully in another. That's not a skill you acquire with a subscription. And now tools exist that produce something that looks like that output, cheaply and instantly. Of course that feels like having your work stolen.

I want to name that honestly before anything else.

This article is about what happens when criticism is legitimate — and what you do with it.

What I Got Wrong (And What I Got Right)

Here's what I know I got wrong: I didn't go deep enough on the source material for that article. I verified that sources existed. I didn't always verify what they actually said. That's a meaningful difference, and the backlash surfaced it.

I publish using a content pipeline — a structured review process before anything goes live. And when I looked at that pipeline after reading the comments, I realized it had a blind spot. It checked for coherence, flow, and basic factual consistency. It didn't have a step where someone adversarial — someone from the domain I was writing about — read the article and asked: "does this condescend to me?"

It should have.

Here's what I think I got right: the core argument — that AI is democratizing access to creation tools, and that this is a disruptive but not unprecedented pattern in history — holds up. The backlash pointed at real failures in my process. Whether the core argument is correct is a question I hold with more humility than I did before — but I still think it's worth defending carefully.

I'm not claiming victory here. I'm claiming accountability.

My Transparency Commitment

Since people asked — directly and indirectly — here is how I actually work:

I use AI tools extensively. For research, for drafting, for fact-checking, for code. I do not produce AI content and pass it off as wholly my own writing in the sense of "I had no involvement." The ideas, the angles, the choices about what to include and what to cut — those are mine. The research briefs are generated by an agent that I direct, using sources I specify, with quality ratings I've defined. The writing is drafted by an AI system following a voice brief I wrote, then revised through multiple review phases including my own read.

Is that "AI-generated content"? In some sense. Is it "my content"? Also yes. I'm not sure the binary framing of AI vs. human is the right lens for how most AI-assisted creation actually works. What I can commit to is being specific about the process when asked — which is what I'm doing now.

I review what my AI system writes. I make editorial calls. Sometimes I rewrite sections from scratch. Sometimes I don't. The pipeline described below is not hypothetical — it's running on my machine right now, and this article went through it.

We've Been Here Before

Jeremy Rifkin has spent decades mapping the pattern. His 2011 Third Industrial Revolution and his 2014 book The Zero Marginal Cost Society describe something that keeps repeating throughout economic history: when a new communication technology converges with new energy regimes, it doesn't just change an industry — it reorganizes who gets to participate in creation.

The printing press, arriving around 1440, did this to the scribal guilds of Europe. Handwritten manuscripts had been the monopoly of trained scribes for centuries. The arrival of Gutenberg's press meant — almost immediately — that the guild's core skill could be replicated faster, cheaper, and by people without their training.

Abbot Johannes Trithemius, writing in 1492, penned In Praise of Scribes, arguing that handwritten manuscripts were spiritually superior to printed books. That the human hand imbued the text with something the machine could not. The argument was sincere, and in some narrow sense, even true — a handwritten manuscript is a different artifact than a printed one. But it didn't stop the press.

The scribal guilds eventually didn't disappear — they transformed into publishing, editing, and typesetting. The craft changed shape.

The Luddites of 1811–1816 were not, despite common assumption, simply opposed to technology. They were skilled textile workers — handweavers, wool finishers, framework knitters — who watched specific machines being deployed to cut their wages and replace their expertise with cheaper unskilled labor. Their grievance was economic and specific. The British government took the threat so seriously that thousands of troops were deployed to the industrial regions. They lost. The machines stayed. But the textile industry that emerged was eventually larger than anything that had come before.

I'm not in a position to tell professional translators what their future holds. I'm a tinkerer, not an economist, and certainly not an expert in creative labor markets. But from what I've observed — in my own work and from what research I've been able to study — the pattern of "technology initially disrupts, then expands the overall ecosystem" has repeated enough times that it's worth naming.

What I do know is that the question translators are asking — "what is the best course?" — is exactly the question skilled workers have asked at every one of these inflection points. And the fact that it's being asked loudly means people are paying attention. That's not a bad sign.

Knowing the historical pattern doesn't make the disruption less painful for the people living through it. It doesn't answer the hard practical questions about income and identity that skilled workers face when their field shifts under them. But it shapes how I think about my own work — and why I think building transparently, with real accountability for what I get wrong, matters.

What Democratization Actually Looks Like

I want to give a concrete example, because abstract arguments about democratization can feel hollow when the person you're talking to is losing real income.

My wife Evy recently used the same tools I use to build something for her work that she couldn't have built alone. She's not a developer — she's a strategist who knows exactly what she needs. These tools gave her the ability to create it herself.

She didn't replace a professional. The work either got done by her, with these tools, or it didn't get done at all. That's what I mean by democratization — not "AI displaces professionals," but "people who previously had zero access to creation now have some."

That distinction matters. It's not always honored in how AI tools get discussed.

Vibe coding was the 2025 story — the moment non-technical people discovered they could prompt their way into working prototypes. In 2026, the shift I'm seeing is toward agentic systems: not single-prompt generation, but orchestrated workflows where multiple specialized agents plan, execute, test, and iterate with minimal intervention. That's a qualitative jump in what's possible. It's also a jump in complexity — which is why process discipline matters more than ever, not less.

The Pipeline Story: Building With the Bricks

Here's the part I want to be specific about, because it's the most honest thing I can say about what the backlash actually produced.

I had a content pipeline — an earlier version. It had the essentials: briefing, research, writing, fact-checking, revision, review, and publication. It was good enough — until it wasn't.

The comments on my translation article revealed two specific failures:

  1. Source verification was shallow — I checked existence, not accuracy
  2. There was no adversarial review step — no one asked "how does this read to someone who works in this domain?"

So I rebuilt it. Version 2.2 now has eight phases:

Phase 1: RESEARCH
  └── Structured research brief with source quality ratings
      (HIGH / MEDIUM / LOW confidence)
  └── Primary sources distinguished from aggregators

Phase 2: WRITING
  └── Author voice brief preserved
  └── Factual constraints locked (no fabricated quotes/stats)

Phase 3: FACT-CHECK
  └── Source-accuracy verification — READ the source, not just confirm it exists
  └── Cross-content consistency check (does this contradict prior articles?)

Phase 4: REVISION
  └── Two-tier fix system:
       - Mechanical fixes (wrong name, broken link, wrong date) → fix immediately
       - Editorial judgment calls (reframing, tone shift) → flag for author

Phase 5: RED TEAM REVIEW  ← NEW in v2.2
  └── Simulated domain professional reads the article
  └── Flags: condescension, oversimplification, missing nuance
  └── Forces the question: "Does this read like I know what I don't know?"

Phase 6: AUTHOR REVIEW
  └── Human reads the full article before anything moves forward

Phase 7: 24-HOUR COOL-DOWN  ← NEW in v2.2
  └── No publishing on the day of writing, regardless of confidence level
  └── One night of sleep, one fresh read in the morning

Phase 8: PUBLISH

The pipeline is the response. A structural response to legitimate criticism. Someone threw a brick. I looked at it. I used it.

What I Learned

I learned that writing about someone else's profession is a higher-stakes move than I treated it as. I went in as a tinkerer sharing observations. Some readers received it as an authority making claims. That gap — between how I see myself and how a piece of writing can be received — is something I'll be more careful about.

I learned that a pipeline is only as good as its adversarial review step. Quality review that only checks for coherence within the author's own worldview will miss the things the author can't see.

I learned that backlash, when it's pointing at something real, is useful data. It's uncomfortable data. But it's data.

And I learned — or maybe confirmed — that I'm going to keep writing. Not because I'm certain I'm right about everything. But because the act of sharing what I'm learning, with transparency about how I'm learning it, is something I believe in. I come from a good place with this. That hasn't changed.

David Brinkley said: "A successful man is one who can lay a firm foundation with the bricks others have thrown at him."

"People only throw stones at trees that bear fruit."

The translators who pushed back on my article did me a service, even if that wasn't their intention. They identified real weaknesses in my process. I took them seriously. I rebuilt. That's the only response worth making.

Il y aura des détracteurs. Il y aura des personnes qui me jetteront la pierre. Jettez-les — je vais construire une maison avec ça.

There will be critics. There will be people who throw stones at me. Keep throwing them — I'm going to build a house with that.

Sources and Inspiration

Jeremy Rifkin:

Historical parallels:

Translation industry data:

Vibe coding and agentic systems:

Content authenticity and AI disclosure:

Quotes:

  • David Brinkley: widely attributed
  • "People only throw stones at trees that bear fruit": traditional proverb, origin disputed
Share this article:
Back to Blog