5D Library › One-Person Enterprise Automation, Avatar Livestream Funnels and Task-Exposure Arbitrage · 中文

Chapter 7. Bai Gui in the Records of the Grand Historian, the Division of Crafts in the Kaogongji, and Coase’s Nature of the Firm

Why read these sources here

The three questions of this book (how small a firm can be, how one person should trade and staff, how work is divided) each have a classical counterpart. Sima Qian’s chapter on the money-makers names Bai Gui as the founder of the art of managing wealth; the Kaogongji, the record of crafts appended to the Rites of Zhou, counts the specialised crafts of a state workshop; and Coase’s essay of 1937 gives the modern account of where a firm’s boundary lies. Two points come out of the check. First, “lean troops and simple administration” (jing bing jian zheng), often attached to Bai Gui in business writing, is not in the Records at all; it is a slogan of the 1940s. What the Records say is that he shared hardship with the servants who worked for him. Second, Bai Gui’s rule of buying what others discard works where prices swing back, and not otherwise.

Bai Gui: take what others discard

Source text

Sima Qian, Records of the Grand Historian, “Biographies of the Money-Makers”

白圭,周人也。当魏文侯时,李克务尽地力,而白圭乐观时变,故人弃我取,人取我与。

能薄饮食,忍嗜欲,节衣服,与用事僮仆同苦乐,趋时若猛兽挚鸟之发。

盖天下言治生祖白圭。

(司马迁 1959)

“Bai Gui was a man of Zhou. In the time of Marquis Wen of Wei, when Li Ke was making the most of the land, Bai Gui delighted in watching the changes of the season, so what others discarded he took, and what others took he gave out … He could make do with plain food, restrain his desires, economise on clothing, and share hardship and joy with the servants who did his work; when the moment came he moved like a beast of prey or a bird of prey striking … So all under heaven who speak of managing wealth take Bai Gui as their ancestor.” The rule is a contrarian inventory rule, and it can be tested.

Calculation

# Chapter 7. p13: "Take what others discard, give what others take."
# A trader holds at most 10 lots of a stored good. Monthly log price:
# x' = x + k (0 - x) + 0.08 e, price = 100 exp(x); k = 0.15 is a market
# that swings back (harvests come and go), k = 0 a random walk. Rules:
# contrarian (buy a lot when price is under 92, sell one when over 108),
# momentum (buy after a 5% monthly rise, sell after a 5% fall), and
# holding 10 lots throughout. Storage costs 0.3 a lot a month.
# Profit over 120 months, 5,000 paths each (seed 90).
import math
import random

def run(rule, k, rng, months=120):
    x, inv, cash = 0.0, (10 if rule == "hold" else 0), 0.0
    prev = 100.0
    if rule == "hold":
        cash -= 10 * 100.0
    for _ in range(months):
        x += k * (0 - x) + 0.08 * rng.gauss(0, 1)
        p = 100 * math.exp(x)
        if rule == "contrarian":
            if p < 92 and inv < 10:
                inv, cash = inv + 1, cash - p
            elif p > 108 and inv > 0:
                inv, cash = inv - 1, cash + p
        elif rule == "momentum":
            if p > prev * 1.05 and inv < 10:
                inv, cash = inv + 1, cash - p
            elif p < prev * 0.95 and inv > 0:
                inv, cash = inv - 1, cash + p
        cash -= inv * 0.3
        prev = p
    return cash + inv * p

if __name__ == "__main__":
    for k, label in ((0.15, "price swings back"), (0.0, "random walk")):
        print("   {} (k = {:.2f}):".format(label, k))
        for rule in ("contrarian", "momentum", "hold"):
            rng = random.Random(90)
            r = [run(rule, k, rng) for _ in range(5000)]
            loss = sum(1 for v in r if v < 0) / len(r)
            print("     {:10s}: mean profit {:7.1f}, losing paths {:.3f}"
                  .format(rule, sum(r) / len(r), loss))
#    price swings back (k = 0.15):
#      contrarian: mean profit   399.0, losing paths 0.029
#      momentum  : mean profit  -380.1, losing paths 0.997
#      hold      : mean profit  -349.1, losing paths 0.979
#    random walk (k = 0.00):
#      contrarian: mean profit   -48.9, losing paths 0.392
#      momentum  : mean profit    97.7, losing paths 0.721
#      hold      : mean profit   111.2, losing paths 0.637

Where the price swings back to a normal level, as grain prices did between good and bad harvests, the contrarian rule makes a mean profit of 399.0 over ten years and loses on only 0.029 of paths, while momentum and holding lose on almost every path. Where the price is a random walk, the same rule loses on average (-48.9) and on 0.392 of paths. Bai Gui’s skill, as Sima Qian puts it, was “watching the changes of the season”: knowing whether a market swings back is the precondition, and the rule is only its consequence.

The Kaogongji: one product gathers many crafts

Source text

Kaogongji (Rites of Zhou, “Offices of Winter”)

国有六职,百工与居一焉。

知者创物,巧者述之守之,世谓之工。

凡攻木之工七,攻金之工六,攻皮之工五,设色之工五,刮摩之工五,搏埴之工二。

故一器而工聚焉者,车为多。

(郑玄 and 贾公彦 1980)

“A state has six kinds of occupation, and the hundred crafts are one of them … The wise create things, the skilled transmit and preserve them, and these the world calls craftsmen … Of those who work wood there are seven crafts, metal six, leather five, colour five, scraping and polishing five, and clay two … So of the objects for which crafts are gathered, the chariot gathers the most.” A chariot needed at least the wheelwright, the box-maker and the shaft-maker, each working to fixed proportions that the text records. That is the point that matters here: crafts that work to a written standard need not negotiate with every other craft.

Calculation

# Chapter 7. p14: The Kaogongji's crafts and the cost of talking.
# The Kaogongji counts 7 woodworking, 6 metal, 5 leather, 5 colouring,
# 5 polishing and 2 clay crafts: 30 in all. A chariot alone gathers
# three wood crafts (wheel, box, shaft) and others besides. If every
# pair of crafts working on one product must agree on its interface,
# there are n(n-1)/2 channels; if each craft works to one written
# standard, n. Each channel independently goes wrong with probability
# 0.02. The same count applies to a team of AI agents.
CRAFTS = {"wood": 7, "metal": 6, "leather": 5, "colour": 5,
          "polish": 5, "clay": 2}

if __name__ == "__main__":
    n_all = sum(CRAFTS.values())
    print("   crafts in the Kaogongji: {}".format(n_all))
    print("    n   pairwise   standard   P(no slip) pairwise  standard")
    for n in (3, 5, 8, 12, n_all):
        pairs = n * (n - 1) // 2
        print("   {:2d}   {:8d}   {:8d}   {:15.5f}   {:8.5f}".format(
            n, pairs, n, 0.98 ** pairs, 0.98 ** n))
#    crafts in the Kaogongji: 30
#     n   pairwise   standard   P(no slip) pairwise  standard
#     3          3          3           0.94119    0.94119
#     5         10          5           0.81707    0.90392
#     8         28          8           0.56798    0.85076
#    12         66         12           0.26359    0.78472
#    30        435         30           0.00015    0.54548

The Kaogongji’s crafts number 30. If every pair working on a product must agree on an interface, 30 crafts have 435 channels, and with a 0.02 chance that any channel goes wrong the probability of no slip is 0.00015; if each works to one standard, there are 30 interfaces and the probability is 0.54548. For 8 agents the figures are 28 against 8 channels, and 0.56798 against 0.85076. Written specifications are to an agent team what the Kaogongji’s proportions were to a chariot workshop.

Coase: where the firm stops

Coase wrote that “a firm will tend to expand until the costs of organizing an extra transaction within the firm will become equal to the costs of carrying out the same transaction by means of an exchange on the open market” (Coase 1937). Chapter 1 applied this with a third option that Coase did not have: a transaction carried out by software that is neither an employee nor a market supplier. When that option is cheap, the boundary of the firm moves inward, toward the founder, until the founder’s own review time becomes the binding cost.

Hypothesis · not established

The author’s interpretation (not yet tested): a founder who writes a specification for each agent’s output (format, checks, acceptance test) spends less review time per deliverable than one who corrects outputs case by case, by at least a third after a month. What would show it wrong: in logged review time, written specifications reduce review per deliverable by less than a third.

Summary

Bai Gui’s “take what others discard” works where prices swing back and fails on a random walk; the frugality the Records praise is sharing hardship with workers, not having none. The Kaogongji’s 30 crafts would need 435 channels without standards, and 30 with them. Coase’s boundary moves toward one person when software becomes the cheapest way to carry out a transaction, and stops where the founder’s review time binds.

References

Coase, Ronald H. 1937. “The Nature of the Firm.” Economica 4 (16): 386 to 405.
司马迁. 1959. 史记. 中华书局.
郑玄, and 贾公彦. 1980. 周礼注疏. 十三经注疏(阮元校刻本). 中华书局.

Fan, Y. (2026). One-Person Enterprise Automation, Avatar Livestream Funnels and Task-Exposure Arbitrage: Coasean Transaction-Cost Collapse, Cyclic Script Retention and High-Margin Solopreneur Portfolios [一人公司自动化架构、数字人直播漏斗与任务级人工智能套利] (Version 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.24090235 · Licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0).