When Switching Stops Being a Choice为什么越来越难换一家為什麼越來越難換一家 · 03
Your Complaint Has Been Received
申诉已经收下,然后呢
申訴已經收下,然後呢
A complaint can enter a system without moving anything inside it. Effective challenge needs authority, independent understanding, and usable alternatives.
窗口还开着,问题也可能原地不动。除了能不能问,还要看谁能重新判断,谁有本事看懂,谁承受得起改变。
窗口還開著,問題也可能原地不動。除了能不能問,還要看誰能重新判斷,誰有本事看懂,誰承受得起改變。
Beacon's assistant stops working for three days. Only then does the city see how much has passed into its hands: building-permit screening, injury-claim processing, clinic bookings. Telephone lines fill. Old manual procedures have been removed and staffing reduced. A fault in one service travels through entrances belonging to many different institutions, each of which had made its own adoption decisions.
This is still the fictional Kingdom State. Keep the earlier benefits in view: permit screening had fallen from weeks to about one, and a backlog had been cleared. Genuine convenience can make a standby route look increasingly unnecessary. We do not need to make technology useless before recognizing how the gradual disappearance of alternatives changes the cost of interruption and challenge.
Beacon asks for resilience funding and a year's exemption from liability for certain outages, saying half the state would stop if it failed. Legislators hold a hearing. No decision is recorded by year-end. That does not prove assent or capture. The assertion of dependence still fails to provide sufficient grounds for the requested funding and exemption. Those require answers of their own.
There are two different difficulties here. A route for challenge can be closed outright. Or hearings, complaint forms, and courts can remain while those doing the questioning lack the conditions to make the arrangements move. The second deserves attention too. An inventory of institutional names cannot tell us whether people have effective ways of reopening what governs their lives.
Start with the rules inside the entrance. Beacon writes an access code for services its assistant can call, then revises it. A small clinic is removed. The same access-review team handles its appeal without reasons; access returns three months later. Restoration is one event. Why removal occurred, what review was available, and how the interval is handled are separate matters.
An internal review team is not automatically dependent. Examine its function: can the clinic initiate a usable review, can Beacon cancel it or remove its people, and can initiating it bring retaliation? What can the outcome change? The story does not specify another route. We cannot invent its absence, or treat the team's existence as evidence that everything necessary is already supplied.
Companies need operating rules. But an unavoidable position making rules for many people cannot retain sole, indefinite authority over the entire reach of its own jurisdiction. This does not require advance public approval for every update. It requires a bounded scope and a route of challenge the rule maker cannot unilaterally close when its exercise becomes inconvenient to defend.
Revision deserves the same scrutiny as initial rule-making. Retiring an old interface looks technical, yet the city's payment system needs costly changes and loses assistant-based access for three weeks. Whether ninety days' notice suffices requires examination. Having given notice does not establish that affected people have no remaining question. A version change can redistribute the costs of an arrangement.
Epic v. Apple supplies a concrete judicial example. In December 2025 the Ninth Circuit upheld civil-contempt findings concerning Apple's 27-percent commission and restrictions on purchases through external links, while reversing and remanding parts of the sanctions, including the unqualified commission ban. Both aspects matter. One cannot cite the affirmation and silently turn a qualified remedial ruling into a blanket ban on every fee.
As of November 20, 2026, the Supreme Court has granted review on the contempt question, with argument scheduled for January 2027. This is not a completed final story. We use the findings already made below to illuminate how revising a rule can burden a route previously ordered open, not to issue an independent judgment on everything the company does.
Rules may also cease to resemble a fixed table. Imagine a future drought: the council gives households priority and protects hospitals and schools, while a model recalculates supply every six hours. The East Market goes without water longer than the northern district. An explanation about aging pipes may be valid. Changing weights do not remove residents' need to question how each allocation is made.
A nonfixed rule does not mean nobody decides. Who set the process for changing it, within what bounds, and how can someone reach an office able to alter an outcome? Fewer than one percent of complaints going to a human would not independently establish failure; perhaps many are resolved earlier. Actual effects, not a human-transfer ratio alone, need to be examined.
The council sets objectives, the water authority delegates operation, and a company builds the system. Each has a part to explain. Saying the model decided cannot make all those addresses disappear. Nor can responsibility simply be loaded onto the employee sending the last reply. Trace actual decisions and actual control rather than stop at whichever human name is nearest the output.
Another dependence sits in the regulator's eyes rather than the pipes. The state model office exists and has engineers but no testing equipment of its own. Methods for spot checks come from the firms examined. Beacon's institutional assistant reads submissions, including Beacon's. A formal reviewer is present while part of its capacity to see is borrowed from the object of review.
A useful tool does not prove a mistaken assessment. Engineers arriving from industry do not prove divided loyalty. Ask instead whether the office can obtain evidence, understand submissions, and test claims without relying exclusively on the reviewed party. Capability dependence is something we can identify without announcing that the regulator has been bought or that a particular decision is corrupt.
Keep this distinct from dependence for a livelihood or career. Who controls a reviewer's income, tenure, or future employment is one question. Whether the reviewer could understand the material without a firm's people and equipment is another. A returning industry engineer may connect both. The two need separate examination, not one account used to declare the other settled in advance.
Later the state funds a laboratory with its own equipment and engineers, reporting directly to the office and legislature. One retest agrees with a firm's result, another does not, and the firm revises its explanation. This proves neither universal unreliability in self-testing nor the end of all dependence. It shows a practical way of building some independent capacity to understand.
An effective challenge need not reverse every conclusion. Reasons that can be checked, a pause, a correction, a revised rule, or accountability for a responsible office may all have force. Adequately maintaining an earlier decision can be legitimate too. Calling every unchanged outcome ineffective would quietly turn the right to question into an obligation to obey every objection.
Return to the outage. Legislators can summon Beacon without possessing a viable way to continue services independently. Clinics need appointments; residents need claims processed; public bodies themselves rely on the entrance. The difficulty lies not merely in size but in alternatives and continuation capacity becoming so thin that a change is extremely costly to contemplate, let alone carry out.
Being unable to afford failure is therefore a dependence to address, not a badge conferring immunity. Public support must identify whose needs it protects, how continuation could be arranged, where money goes, and what conditions accompany it. Securing a supply and preserving the provider in every existing form are different undertakings. One cannot be treated as proof of the other.
Institutions' own choices belong in the account. The city reduced its department to three staff; the labor office left no manual route. Beacon did not make those decisions for them. Adoption and withdrawal of alternatives have responsible decision makers. We should not first spend away standby capability and then attribute every ensuing vulnerability solely to a supplier's supposed character.
Late structural changes can genuinely be expensive. Systems, data, staff, and settlement arrangements grow together; migration brings risks. Courts considering remedies weigh concrete causation, proposed arrangements, and costs. A decision against a particular breakup proposal does not establish that no challenge has had effect. Nor does finding one proposal inadequate prove that every possible separation is technically impossible.
That makes advance preparation worth examining: interfaces that permit continuation, independently held records, testing capacity, and a defined handover in an outage. Such arrangements guarantee no loss-free future. They avoid postponing every choice until a sole supplier cannot be interrupted. Their costs still need scrutiny; a possible future danger cannot automatically approve every proposed budget or technical plan.
Influence may also cross borders. A state retains admission powers but may find interruption too costly to exercise them. Joint constraint is one possible response; the European Union offers a real institutional form, not a guarantee of effectiveness. Its decisions retain their particular laws and procedures. A larger reviewing institution does not automatically solve dependence, or turn a company into the equivalent of a state.
Look beyond the complaint hotline number posted on the wall. Are there people who can understand, offices that can change something, and ways to continue not entirely held by the party being questioned? Keeping formal routes is important. It can also reassure too soon. Received does not mean answered; an answer still needs to reach the particular matter someone placed before it.
灯塔助手停摆三天,城市才发现,原来已经有那么多事交在它手里。建筑许可的初审停了,工伤理赔积下数千件,诊所重新接电话,线路很快占满。人工流程先前撤了,留在岗位上的人也少了。一家公司的故障,沿着许多机构的入口传开。
这仍是王国州的虚构故事。也要记它此前做好的那面:审批从数周缩到约一周,积压曾被清空。正因为日常的方便是真的,备用的路才可能一点点被当作多余。我们不需要先把技术写得毫无用处,才能看见后来难在哪里。
灯塔提出一笔备用与韧性的支持,还请求一年内对某类停摆免责,理由是它一倒,半个州停。议会开了听证,议员也问,难道倒不起就得一直托着?到年底尚未表决。没有表决,不证明已经答应,更不证明议会被拿住了;但那句倒不起,仍不是支持与免责的充分根据。
这里有两种不同的困难。一种是问的路直接被关了。另一种更不容易看见:窗口、听证、法院都在,问的人却已经没有足够的条件,使问题对安排产生作用。我们得把第二种也读出来,否则只要机构名称齐全,就像一切已经有了保障。
先看门内的规则。灯塔写接入守则,决定哪些服务可以由助手调用,后来改了条款。小诊所被下架,申诉仍交给原来的接入审核组,不说明理由,三个月后恢复。恢复是一件事,为什么下架、能否复核、这三个月怎样处理,是另外几件。
审核组在公司里面,不自动证明它不独立。关键是功能:诊所能不能发动一条真正的复核路,灯塔能否独自取消、撤人或让发动者吃亏,复核能改动什么。故事没有交代外部可用的路,不能擅补成绝对没有,也不能因为有审核组就补成已经足够。
公司当然需要立经营规则。但一个别人绕不过去的位置,能管哪些事、管到哪里,不能把全部边界也永远留给自己决定。这不要求每次升级、每一项条款都先报批;它要求权限有范围,作出的规则有一条自己关不上的追问路。
改规也同样。停用旧接口,表面是版本更新,接在上面的市政缴费却要改造,花去预算,三周不能在助手里办。提前九十天通知够不够,需要具体核;通知过了,不等于别人已经无事可问。技术变化也会把一项安排的代价移到别人身上。
现实里,Epic诉苹果案显示过一扇门怎样在允许打开的同时又变难走。第九巡回上诉法院2025年12月维持了对苹果的民事藐视认定,认为站外购买27%的抽成及相关链接限制违反既有禁令;同时改判并发回部分制裁,不维持全面禁止任何抽成的原样安排。两件要一起写。
截至2026年11月20日,最高法院已就其中的藐视问题受理审查,辩论排在2027年1月。这不是已经走完所有程序的终局故事。这里照下级法院已作的认定说明:改一条规,可以改变原来被允许的路线;不由此替法院判断一家公司的全部行为。
规则还可能不在一张固定表上。设想未来旱年,市议会规定家庭优先、医院学校不停水,模型每六小时重排一次供水。东市停得比北城久,运维方说旧管网损失大。它可以有道理,但市民不能因为规则会变,就失去问每次按什么切水的地方。
权重不固定,不等于没有人在决定。要问谁定了让它这样变,变化范围是什么,受影响的人怎样把一件事送到能改变结果的位置。申诉转人工不到百分之一,也不独自证明无效;也许许多问题已答完,需要核实际结果,而不是靠人工比例替所有过程作证。
市议会给出目标,水务局委托运维,公司设计系统,这些位置各有一份账。不能把它们合写成模型决定了,便再找不到人。也不能把全部责任塞给最后回信的工作人员。安排怎样分工、谁真正控制哪一段,应当沿实际决定的链条查。
还有一种依赖,不发生在供水管里,而发生在审查者的眼睛里。州模型事务处设了,有五位工程师,却没有自己的测试设备;抽查要请被查的一方提供方法。它用灯塔的机构版读申报,灯塔自己的申报也由灯塔的工具读。
工具好用,不证明读错了。工程师从企业来,也不证明只为旧东家说话。问题是,离开被审的一方,事务处还能不能取到信息,看懂材料,自己测一遍?我们读得出能力依赖,不能因此直接宣布审查者已经被收买或控制。
这一种与前途上的依赖还不同。一个审查者的收入、任期或下一份工作受谁影响,是一问;没有企业的人和设备,他还能不能判断,是另一问。同一位来回任职的工程师可能同时牵着两条线,需要分别核,不拿一种风险替另一种结账。
后来州在学院建了检验室,有自己的工程师与设备,报告直接交事务处和议会。两项复测,一项相符,一项不符,对方改了说明。这不是证明企业自测一概不可信,也不是八个人已经解决所有依赖;它只让我们看见,独立理解可以从实际条件里一点点建立。
问得动,不必每次推翻。补交可核的理由、暂停、修改、重定规则,乃至让一个失职的位置承担后果,都可以是效力。有根据地维持原决定也可能成立。若把每次没有改判都叫无效,就把听取异议悄悄改成了必须服从异议。
再回到停摆的三天。议会能叫公司来,不等于手里已经有能接上的路。诊所要排号,市民要办事,公共机构自己也依赖那个入口。这里的困难不只在公司大,而在替代与继续做事的能力已经薄到改变很难承受。
因此倒不起是一项要处理的依赖,不是公司可以免责的奖章。若公共资金要介入,得说明保护哪些人的什么需要,是否有别的接续,钱给谁、换来什么条件。保护供给,与无条件保存这家公司的全部样子,不能混为一件事。
机构自己的选择也要入账。苹果城把科室缩到三人,劳动厅没有保留人工流程,这些不是灯塔做的。采用一种方便的工具、撤掉其他路径,都有决定者。不能先把备用能力花掉,再将全部脆弱只记在供应商的性格上。
事后要改,往往确实很贵。训练系统、数据、人员和结算长在一起,迁移会有风险。法院选择救济时,也会权衡具体的因果、可行方案和代价。没有下令拆分,不证明问题没人管;对某个方案认定不足,也不证明任何分开都做不到。
这使事前准备更值得问:接口能否接续,记录是否另存,谁还能独立检验,停摆时哪一项由谁接手。保留这些,不保证将来没有损失;它至少不把每一个选择都拖到唯一的供应者已经无法停下之后。准备的成本也要算,不能拿未来的危险替任何预算自动盖章。
有时影响跨过国境,一国握着准入权,却承受不起不用它。几国合作约束是一种可能的回应,欧盟提供了现实中的一种制度形态。它的具体决定仍按各自法律与程序读,不能仅凭规模更大,就保证协作一定有效,或把一家公司等同于一个国家。
所以,真正要追的不是墙上有没有申诉电话,而是电话后面有没有看得懂的人、改得动的位置,以及不全押在被问一方手里的继续办法。形式留下来很重要,却也可能让人太早放心。收到,不等于答完;答复,也还得看它有没有落到那件事上。
燈塔助手停擺三天,城市才發現,原來已經有那麼多事交在它手裡。建築許可的初審停了,工傷理賠積下數千件,診所重新接電話,線路很快佔滿。人工流程先前撤了,留在崗位上的人也少了。一家公司的故障,沿著許多機構的入口傳開。
這仍是王國州的虛構故事。也要記它此前做好的那面:審批從數周縮到約一週,積壓曾被清空。正因為日常的方便是真的,備用的路才可能一點點被當作多餘。我們不需要先把技術寫得毫無用處,才能看見後來難在哪裡。
燈塔提出一筆備用與韌性的支持,還請求一年內對某類停擺免責,理由是它一倒,半個州停。議會開了聽證,議員也問,難道倒不起就得一直托著?到年底尚未表決。沒有表決,不證明已經答應,更不證明議會被拿住了;但那句倒不起,仍不是支持與免責的充分根據。
這裡有兩種不同的困難。一種是問的路直接被關了。另一種更不容易看見:窗口、聽證、法院都在,問的人卻已經沒有足夠的條件,使問題對安排產生作用。我們得把第二種也讀出來,否則只要機構名稱齊全,就像一切已經有了保障。
先看門內的規則。燈塔寫接入守則,決定哪些服務可以由助手調用,後來改了條款。小診所被下架,申訴仍交給原來的接入審核組,不說明理由,三個月後恢復。恢復是一件事,為什麼下架、能否複核、這三個月怎樣處理,是另外幾件。
審核組在公司裡面,不自動證明它不獨立。關鍵是功能:診所能不能發動一條真正的複核路,燈塔能否獨自取消、撤人或讓發動者吃虧,複核能改動什麼。故事沒有交代外部可用的路,不能擅補成絕對沒有,也不能因為有審核組就補成已經足夠。
公司當然需要立經營規則。但一個別人繞不過去的位置,能管哪些事、管到哪裡,不能把全部邊界也永遠留給自己決定。這不要求每次升級、每一項條款都先報批;它要求權限有範圍,作出的規則有一條自己關不上的追問路。
改規也同樣。停用舊接口,表面是版本更新,接在上面的市政繳費卻要改造,花去預算,三周不能在助手裡辦。提前九十天通知夠不夠,需要具體核;通知過了,不等於別人已經無事可問。技術變化也會把一項安排的代價移到別人身上。
現實裡,Epic訴蘋果案顯示過一扇門怎樣在允許打開的同時又變難走。第九巡回上訴法院2025年12月維持了對蘋果的民事藐視認定,認為站外購買27%的抽成及相關連結限制違反既有禁令;同時改判並發回部分制裁,不維持全面禁止任何抽成的原樣安排。兩件要一起寫。
截至2026年11月20日,最高法院已就其中的藐視問題受理審查,辯論排在2027年1月。這不是已經走完所有程序的終局故事。這裡照下級法院已作的認定說明:改一條規,可以改變原來被允許的路線;不由此替法院判斷一家公司的全部行為。
規則還可能不在一張固定表上。設想未來旱年,市議會規定家庭優先、醫院學校不停水,模型每六小時重排一次供水。東市停得比北城久,運維方說舊管網損失大。它可以有道理,但市民不能因為規則會變,就失去問每次按什麼切水的地方。
權重不固定,不等於沒有人在決定。要問誰定了讓它這樣變,變化範圍是什麼,受影響的人怎樣把一件事送到能改變結果的位置。申訴轉人工不到百分之一,也不獨自證明無效;也許許多問題已答完,需要核實際結果,而不是靠人工比例替所有過程作證。
市議會給出目標,水務局委託運維,公司設計系統,這些位置各有一份帳。不能把它們合寫成模型決定了,便再找不到人。也不能把全部責任塞給最後回信的工作人員。安排怎樣分工、誰真正控制哪一段,應當沿實際決定的鏈條查。
還有一種依賴,不發生在供水管裡,而發生在審查者的眼睛裡。州模型事務處設了,有五位工程師,卻沒有自己的測試設備;抽查要請被查的一方提供方法。它用燈塔的機構版讀申報,燈塔自己的申報也由燈塔的工具讀。
工具好用,不證明讀錯了。工程師從企業來,也不證明只為舊東家說話。問題是,離開被審的一方,事務處還能不能取到資訊,看懂材料,自己測一遍?我們讀得出能力依賴,不能因此直接宣佈審查者已經被收買或控制。
這一種與前途上的依賴還不同。一個審查者的收入、任期或下一份工作受誰影響,是一問;沒有企業的人和設備,他還能不能判斷,是另一問。同一位來回任職的工程師可能同時牽著兩條線,需要分別核,不拿一種風險替另一種結帳。
後來州在學院建了檢驗室,有自己的工程師與設備,報告直接交事務處和議會。兩項復測,一項相符,一項不符,對方改了說明。這不是證明企業自測一概不可信,也不是八個人已經解決所有依賴;它只讓我們看見,獨立理解可以從實際條件裡一點點建立。
問得動,不必每次推翻。補交可核的理由、暫停、修改、重定規則,乃至讓一個失職的位置承擔後果,都可以是效力。有根據地維持原決定也可能成立。若把每次沒有改判都叫無效,就把聽取異議悄悄改成了必須服從異議。
再回到停擺的三天。議會能叫公司來,不等於手裡已經有能接上的路。診所要排號,市民要辦事,公共機構自己也依賴那個入口。這裡的困難不只在公司大,而在替代與繼續做事的能力已經薄到改變很難承受。
因此倒不起是一項要處理的依賴,不是公司可以免責的獎章。若公共資金要介入,得說明保護哪些人的什麼需要,是否有別的接續,錢給誰、換來什麼條件。保護供給,與無條件保存這家公司的全部樣子,不能混為一件事。
機構自己的選擇也要入帳。蘋果城把科室縮到三人,勞動廳沒有保留人工流程,這些不是燈塔做的。採用一種方便的工具、撤掉其他路徑,都有決定者。不能先把備用能力花掉,再將全部脆弱只記在供應商的性格上。
事後要改,往往確實很貴。訓練系統、資料、人員和結算長在一起,遷移會有風險。法院選擇救濟時,也會權衡具體的因果、可行方案和代價。沒有下令拆分,不證明問題沒人管;對某個方案認定不足,也不證明任何分開都做不到。
這使事前準備更值得問:接口能否接續,記錄是否另存,誰還能獨立檢驗,停擺時哪一項由誰接手。保留這些,不保證將來沒有損失;它至少不把每一個選擇都拖到唯一的供應者已經無法停下之後。準備的成本也要算,不能拿未來的危險替任何預算自動蓋章。
有時影響跨過國境,一國握著準入權,卻承受不起不用它。幾國合作約束是一種可能的回應,歐盟提供了現實中的一種制度形態。它的具體決定仍按各自法律與程序讀,不能僅憑規模更大,就保證協作一定有效,或把一家公司等同於一個國家。
所以,真正要追的不是牆上有沒有申訴電話,而是電話後面有沒有看得懂的人、改得動的位置,以及不全押在被問一方手裡的繼續辦法。形式留下來很重要,卻也可能讓人太早放心。收到,不等於答完;答覆,也還得看它有沒有落到那件事上。
Source & reading note原论文与阅读说明原論文與閱讀說明
Nine essays draw on the Monopoly prequel, seven papers and dialogue with earlier thinkers. Kingdom State, Crossways, Beacon and future water provision are imagined, not disguised allegations or forecasts. Linked cases distinguish investigations, allegations, jurisdiction, liability and remedies within their stated institutional and dated scope. Size, success, defaults and investment do not alone establish a violation; benefits do not answer for every exercise of control. Separation, interoperability and standby capacity are discussed as proposals with costs, not proven universal solutions. This is not legal or investment advice. English is independently written.
九篇据《垄断论》前置篇、七篇正文与传统对话篇改写。王国州、通衢、灯塔及未来供水等场景为设想,不是现实企业的化名或对未来的预测。真实案件与制度材料另列来源,区分调查、指控、管辖判断、责任判决与救济;按相应机构和材料日期说明,不以一项法律结论代答全部规范问题。规模、成功、默认设置或投资关系本身不直接证明越线;有益的一面也不替具体控制作答。关于拆分、互通与备用能力的建议不是已经验证的万能方案,不提供法律或投资建议。简繁中文同文,英文独立重写。
九篇據《壟斷論》前置篇、七篇正文與傳統對話篇改寫。王國州、通衢、燈塔及未來供水等場景為設想,不是現實企業的化名或對未來的預測。真實案件與制度材料另列來源,區分調查、指控、管轄判斷、責任判決與救濟;按相應機構和材料日期說明,不以一項法律結論代答全部規範問題。規模、成功、默認設置或投資關係本身不直接證明越線;有益的一面也不替具體控制作答。關於拆分、互通與備用能力的建議不是已經驗證的萬能方案,不提供法律或投資建議。簡繁中文同文,英文獨立重寫。