Jane Ruffino 将 TAT-8 的退役界定为网际网路的一段物质史：工程师在 2025 年开始拆除第一条跨洋光纤电缆，这项基础设施由 AT&T、British Telecom 和 France Telecom 于 1988 年 12 月 14 日启用。文章否定了流传已久的鲨鱼叙事与近年的破坏焦虑，强调目前近 600 条海底电缆透过仅数千名专业人员构成的劳动体系，承载了几乎全部洲际流量。文中也把 TAT-8 放回电信史脉络：光纤概念在 1960s 成熟、陆上部署于 1970s 开始，而在美国监管压力将新电缆核准与相对铜缆的重大技术进步绑定后，Bell Labs 与 STC 于 1978 承诺建设约 6,000 km 的海底链路。

核心证据来自 Subsea Environmental Services 回收船 MV Maasvliet 的船上报导：在 Leixoes 外海，14 人船员在风暴 Dexter 与 Erin 造成天候延误后，卸下 1,012 km 回收的 TAT-8。这条电缆在 2002 年因故障修复成本不再合算而退役；它在 18 个月内就达到规划容量，且所属系统到 2001 年已扩展至 TAT-14。回收作业在技术上精密、在体力上高度手动：中继器（TAT-8 中超过 100 个）每个约长 2 m、重约 400 kg、耐压至 8,000 m 深度，并由 3 人小组操作；盘缆工采 8-hour 班次与 30-minute 开/停循环；抓钩航次约以 1 knot（约 1.85 km/h）推进，当涌浪达 4 m 可暂停。文中也重访历史上的鲨鱼说法，透过 Optican-1（1985）分流故障调查、水族箱测试与颚部模拟器试验，显示即使「fish bite protection」钢层因抗磨耗而成为标准，仍未出现可靠的咬痕模式。

其影响聚焦于基础设施治理、环境取舍与循环材料经济：目前约 2 million km 的退役电缆多数仍留在海床上，但选择性回收可清理已验证路径，并限制对新海床区域的扰动。文中引用 UK National Oceanography Centre 工作的初步发现指出，主要冲击更多来自船舶与抓钩作业，而非电缆本体，且敏感栖地常被保留不受触及。回收电缆会进入工业回收流程（钢、高品位铜与聚乙烯；光纤本身较难再利用），而这一点之所以重要，是因 International Energy Agency 警告约 10 years 内可能出现铜短缺。文章也强调劳动风险与接班风险：关键专业知识高度依赖个人、正在老化且难以替代，Elaine Stafford 与 87-year-old Jean Devos 等资深从业者即为例证；同时，低轨卫星被定位为互补方案，但在长距离容量、可修复性与生命周期耐久性（例如约 5 years 的更换周期）方面仍弱于海底光纤。

Jane Ruffino frames the retirement of TAT-8 as a material history of the internet: engineers began removing the first transoceanic fiber-optic cable in 2025, dismantling infrastructure launched on December 14, 1988 by AT&T, British Telecom, and France Telecom. The article rejects the long-running shark narrative and recent sabotage anxieties, emphasizing that nearly 600 subsea cables now carry almost all intercontinental traffic through a labor system of only a few thousand specialists. It also situates TAT-8 in telecom history: fiber concepts matured in the 1960s, terrestrial deployment started in the 1970s, and Bell Labs plus STC committed in 1978 to a roughly 6,000 km submarine link after US regulatory pressure tied new cable approvals to major technical advances over copper.

The core evidence comes from onboard reporting on Subsea Environmental Services’ recovery vessel MV Maasvliet, where a 14-person crew offloads 1,012 km of recovered TAT-8 near Leixoes after weather delays from storms Dexter and Erin. The cable, retired in 2002 after a fault became uneconomic to fix, had filled its planned capacity within 18 months and belonged to a system that had grown to TAT-14 by 2001. Recovery operations are technically precise and physically manual: repeaters (more than 100 in TAT-8) are about 2 m long, weigh around 400 kg each, are pressure-rated to 8,000 m depth, and are handled by teams of 3; coilers work 8-hour shifts in 30-minute on/off cycles; grapnel runs proceed at about 1 knot (about 1.85 km/h) and can be suspended when swells reach 4 m. Historical shark claims are revisited through Optican-1 (1985) shunt-fault investigations, aquarium tests, and jaw-simulator trials, showing no robust bite pattern even though “fish bite protection” steel layers became standard for abrasion resilience.

Implications center on infrastructure governance, environmental tradeoffs, and circular materials economics: most of roughly 2 million km of retired cable remains on seabeds, but selective recovery clears proven routes and limits disturbance to new seafloor areas. Early findings cited from UK National Oceanography Centre work suggest major impacts come more from vessels and grapnel activity than from cables themselves, and sensitive habitats are often left untouched. Recovered cable streams into industrial recycling (steel, high-grade copper, and polyethylene; fiber itself is less reusable), which matters as the International Energy Agency warns of possible copper shortfalls within about 10 years. The piece also highlights labor risk and succession risk: key expertise is embodied, aging, and hard to replace, illustrated by veterans such as Elaine Stafford and 87-year-old Jean Devos, while low-Earth-orbit satellites are presented as complementary but still weaker than subsea fiber in long-haul capacity, repairability, and lifecycle durability (for example, replacement cycles around 5 years).