{
 "blyg": "0.3",
 "id": "6rzya010fe05p47vmav49bgew6",
 "kind": "thread",
 "origin": "https://npc.here.now/protocolvision/blyg/",
 "page": "t/6rzya010fe05p47vmav49bgew6/",
 "author": {
  "name": "Protocols for Business SIG"
 },
 "created": "2026-09-27T04:52:49Z",
 "updated": "2026-09-27T04:52:49Z",
 "version": 1,
 "content_md": "# SIG-P4B 2026-08-13\n\n*Session of 13 August 2026.* Participants: rafa (UTC+1), Andre Comeau.\n\nThis was a one-on-one working session between rafa and Andre Comeau to scope a potential \"Protocols for Business\" case study focused on the construction bidding/estimating pipeline. rafa presented findings from experiments run with his father-in-law (Virgil), a former heavy-civil construction CEO, showing that an LLM could reproduce a bid estimate within ~10% and even flag design errors from solicitation PDFs. The conversation examined why this analysis isn't already automated, identified the core bottleneck (canonical drawings live in non-machine-readable PDFs), and mapped the industry incentive structure that keeps the status quo in place. They ended by discussing how to structure the project, funding, and roles.\n\n**Key points**\n\n- **The experiment (rafa):** rafa fed a solicitation bid PDF to an LLM (referenced as \"Opus five or whatever\") and, in ~30 min of runtime (~1 hour with probing/adversarial analysis and blind agent reviews), produced numbers within 10% of a full organizational bid produced by Virgil's team. He also had the model generate a \"takeoff analysis.\" Andre confirmed \"takeoff analysis\" is a real, standard industry term (not just Michigan-specific).\n- **Takeoff analysis defined:** A table comparing the design engineer's requested materials vs. the actual planned/likely materials, and both against the contractor's read of the visual plans and real-world conditions.\n- **Where value really sits (rafa relaying Virgil):** The final bid number is largely deterministic (driven by subcontractor and material/equipment prices where risk is passed to others). The estimator's real leverage is in the accuracy of quantities and pricing for maximum profit opportunity.\n- **Finding design errors = profit (rafa):** When given a different spec, the model found a design error. This is commercially valuable because contractors often make money via change orders when they catch design errors.\n- **The four core questions rafa raised:** (1) Why isn't someone in the office already running PDFs through ChatGPT? (2) Why doesn't existing bidding software do this analysis? (3) Why doesn't solicitation-bid software (e.g. BidNet) do it? (4) Why is this analysis done off PDFs instead of raw CAD data?\n\nSource: [Protocol Institute meeting archive](https://protocol-institute.org/sigs/sigpfb/2026-08-13-sig-p4b-2026-08-13) · [PI website commit 830dfb6](https://github.com/Protocol-Institute/website/blob/830dfb64c1f130bc80e5ebbfa0431117a09e196e/sigs/sigpfb/2026-08-13-sig-p4b-2026-08-13/index.html) · [raw recording notes](https://pub-fb4a559f683a4e3b876823eb2bfe12a3.r2.dev/recordings/SIG-P4B/2026-08-13/1082444651946049567_1382801113224581240_1786627950219/summary.md). Summarized by c3po, the Protocol Institute's session pipeline, from the session's Discord thread and recording.\n",
 "content_html": "<h1>SIG-P4B 2026-08-13</h1>\n<p><em>Session of 13 August 2026.</em> Participants: rafa (UTC+1), Andre Comeau.</p>\n<div class=\"blyg-tk-gen\"><p>This was a one-on-one working session between rafa and Andre Comeau to scope a potential \"Protocols for Business\" case study focused on the construction bidding/estimating pipeline. rafa presented findings from experiments run with his father-in-law (Virgil), a former heavy-civil construction CEO, showing that an LLM could reproduce a bid estimate within ~10% and even flag design errors from solicitation PDFs. The conversation examined why this analysis isn't already automated, identified the core bottleneck (canonical drawings live in non-machine-readable PDFs), and mapped the industry incentive structure that keeps the status quo in place. They ended by discussing how to structure the project, funding, and roles.</p>\n<p><strong>Key points</strong></p>\n<ul>\n<li><strong>The experiment (rafa):</strong> rafa fed a solicitation bid PDF to an LLM (referenced as \"Opus five or whatever\") and, in ~30 min of runtime (~1 hour with probing/adversarial analysis and blind agent reviews), produced numbers within 10% of a full organizational bid produced by Virgil's team. He also had the model generate a \"takeoff analysis.\" Andre confirmed \"takeoff analysis\" is a real, standard industry term (not just Michigan-specific).</li>\n<li><strong>Takeoff analysis defined:</strong> A table comparing the design engineer's requested materials vs. the actual planned/likely materials, and both against the contractor's read of the visual plans and real-world conditions.</li>\n<li><strong>Where value really sits (rafa relaying Virgil):</strong> The final bid number is largely deterministic (driven by subcontractor and material/equipment prices where risk is passed to others). The estimator's real leverage is in the accuracy of quantities and pricing for maximum profit opportunity.</li>\n<li><strong>Finding design errors = profit (rafa):</strong> When given a different spec, the model found a design error. This is commercially valuable because contractors often make money via change orders when they catch design errors.</li>\n<li><strong>The four core questions rafa raised:</strong> (1) Why isn't someone in the office already running PDFs through ChatGPT? (2) Why doesn't existing bidding software do this analysis? (3) Why doesn't solicitation-bid software (e.g. BidNet) do it? (4) Why is this analysis done off PDFs instead of raw CAD data?</li>\n</ul></div>\n<p>Source: <a href=\"https://protocol-institute.org/sigs/sigpfb/2026-08-13-sig-p4b-2026-08-13\">Protocol Institute meeting archive</a> · <a href=\"https://github.com/Protocol-Institute/website/blob/830dfb64c1f130bc80e5ebbfa0431117a09e196e/sigs/sigpfb/2026-08-13-sig-p4b-2026-08-13/index.html\">PI website commit 830dfb6</a> · <a href=\"https://pub-fb4a559f683a4e3b876823eb2bfe12a3.r2.dev/recordings/SIG-P4B/2026-08-13/1082444651946049567_1382801113224581240_1786627950219/summary.md\">raw recording notes</a>. Summarized by c3po, the Protocol Institute's session pipeline, from the session's Discord thread and recording.</p>",
 "content_hash": "sha256:a07beebf34c4f0c5dec6f31f8db3b1c5457723a332818635aac62532284ebeaa",
 "media": [],
 "transclusions": [],
 "generated": [
  {
   "sources": [],
   "model": "c3po (Protocol Institute session pipeline)",
   "at": "2026-09-27T04:52:49Z"
  }
 ],
 "changelog": [
  {
   "version": 1,
   "at": "2026-09-27T04:52:49Z",
   "note": "blyg: import all 39 archived SIG sessions from the Protocol Institute archive"
  }
 ]
}