硅谷夢幻產品需要接地氣
Imagine a place where driverless cars roam the streets, intelligent electricity grids regulate the power and everything is built from the ground up to the specifications of the ideal “smart” city.
想象一下這個地方，無人駕駛汽車行駛在街上，智能電網調節電力，一切都按照理想的“智能”城市的規格從一張白紙開始建設。
This idealised place has been getting a lot of attention in Silicon Valley of late. Y Combinator, the region’s best-known tech incubator, is the latest to dream, announcing a research project aimed at stimulating ideas and designs that it hopes to use in a real-world location.
這種理想之地最近在硅谷獲得極大關注。硅谷最著名的科技孵化器Y Combinator是最新的夢想家，宣佈實施一個研究項目，旨在激勵有望在真實世界用得上的創意和設計。
It is far from alone. From Google co-founder Larry Pageto libertarian venture capitalist Peter Thiel, ideas like this have had plenty of currency. Y Combinator says it wants to to speed up the adoption of technologies for ordinary people, not just build “crazy libertarian utopias for techies”.
Y Combinator絕非唯一一家有此類想法的公司。從谷歌(Google)聯合創始人拉里•佩奇(Larry Page)到信奉自由意志主義的風險投資家彼得•塞爾(Peter Thiel)，此類想法頗有市場。Y Combinator表示，它希望加快普及面向普通人的技術，而非僅僅是“爲技術迷建造瘋狂的自由意志主義烏托邦”。
This hankering is prompted by more than a passing bout of California dreaming. It reflects a deeper issue that the tech industry faces as it moves into a new phase — and one that will have real-world implications for investors.
推動這種言論的不僅僅是轉瞬即逝的加州夢想。它反映出科技行業進入新階段面臨的更深層次的問題，這個問題將對投資者產生切實的影響。
Companies like Google and Facebook — where Mr Thiel was the first outside investor — made their fortunes at a time when the digital and physical worlds were largely separate. But technologies like robotics and artificial intelligence are breaking down that barrier, pushing algorithms deeper into the functioning of the everyday world.
谷歌和Facebook這類公司——塞爾是Facebook第一個外部投資者——是在數字和真實世界基本上分離的時候異軍突起的。但機器人和人工智能等技術打破了數字和真實世界之間的壁壘，推動算法深入日常世界的運行。
Hence the dream of an idealised place that is both open to — and safe for — experimentation. But it is little help to present-day tech companies, which will have to make their products function in today’s world.
因此人們夢想有這樣一個理想之地，它對前沿實驗既是開放的，也是安全的。但這對當前的技術公司沒有什麼幫助，這些公司將不得不讓它們的產品在當今世界運行。
Last week’s news of a fatal accident involving Tesla Motors’ autonomous driving technology is a case in point. The first death of a person at the wheel of a car that was operating under its own control raises difficult ethical and regulatory issues.
最近涉及特斯拉汽車(Tesla Motors)自動駕駛技術的致命車禍是一個很好的例子。第一例自動駕駛死亡事故引發了棘手的道德和監管問題。
From the purely rationalistic position taken by Tesla boss Elon Musk, one death after the 130m miles of autonomous driving his vehicles have clocked up is better than the average fatality rate on US roads. Even before this, “the better-than-human threshold had been crossed”, Tesla declared in a defensive blog post on Wednesday.
從特斯拉老闆埃隆•馬斯克(Elon Musk)所採取的純理性的角度看，自動駕駛行駛超過1.3億英里以來才發生一例車禍死亡，在安全記錄上優於美國道路交通的平均致命事故率。特斯拉發佈的一篇採取守勢的博文稱，甚至在該事故發生之前，“自動駕駛就已跨過了比人工駕駛更安全的門檻”。
On the other hand, if it is found that the technology resulted in a death that an attentive driver would readily have avoided, it would be a setback. It also would raise the question of whether the messy human world, where streets were not designed for robots and other drivers make mistakes, can ever be made safe for driverless vehicles.
另一方面，如果人們發現，自動駕駛技術導致了一起專心駕駛的司機本可輕鬆避免的死亡事故，那將是一個挫折。它也將令人質疑，雜亂的人類世界——街道並非爲機器人設計，其他駕駛者會犯錯——究竟能否被改造得對無人駕駛汽車安全。
This puts US regulators in an unenviable position as they consider how to respond. And what goes for cars may one day go for many other everyday situations, as algorithms are embedded in more objects.
美國監管者正在考慮如何做出迴應，他們的處境不令人羨慕。隨着算法被植入更多物體，汽車發生的情況，有朝一日也可能發生在其他許多日常場合。
With its threat of punitive damages through the court system, the US can be a risky place for product experimentation. Mr Musk said on Wednesday that the crash had not yet led to any product liability suits. But regulators may still be cautious. Driverless-cars development is already threatened by fragmented state-level regulation, and the Tesla crash certainly won’t help.
由於可能被法庭判決支付懲罰性損害賠償金，在美國進行產品試驗風險很大。馬斯克最近表示，此次事故還沒有導致任何產品責任訴訟。但監管者仍可能採取謹慎態度。無人駕駛汽車的發展已經受到各自爲政的各州監管的威脅，特斯拉的撞車事故當然不會有幫助。
The danger for tech is that it ends up like the pharmaceuticals industry, which has complained for years about a supposedly risk-averse regulatory system that adds huge costs and delays to new drugs that might offer benefits. Yet drug development has not been driven offshore, despite repeated warnings. The US is still the world’s richest drugs market, so it makes sense to persevere.
高技術的風險在於，它最終就像製藥行業一樣，後者多年來抱怨：厭惡風險的監管體系增加了鉅額成本，並拖延了可能造福於人類的新藥上市。然而藥物研發並未被擠出境外，儘管有人再三警告。美國仍然是全球最富裕的藥物市場，因此堅持是明智的。
The same might not be true of technologies like driverless cars and drones. China is racing to capitalise on the technology and Baidu — whose expertise in deep learning stems partly from Silicon Valley — has a goal of getting to mass production in five years.
無人駕駛汽車和無人機等技術可能不會這樣。中國正在搶着利用這種技術，百度(Baidu)制定了在5年內大規模生產的目標——該公司在深度學習上的專長部分源於硅谷。
There are at least two things Silicon Valley could do to help itself. One is to pay more attention to designing the new interactions between humans and “intelligent” machines. Tesla warns its drivers to keep their hands on the wheel at all times, but it also brags that its technology is better than a human driver — the kind of mixed message that makes it harder for the humans still in the loop to know exactly when and how they may be called on to override the technology.
硅谷至少可以做兩件事來自助。一是更加致力於設計人類與“智能”機器之間新的互動。特斯拉警告，駕駛者必須時刻將手放在方向盤上，但它也誇口稱，其技術比人類駕駛者更加優越，這種混雜信息讓仍然“在環”的人類更難知道自己何時以及如何被要求接手，取代無人駕駛技術。
The second is to do more to educate users about both the benefits and the risks of robotics and AI. That would help to ensure an appropriate response to the shortcomings of the technology — especially when, inevitably, a human dies due to the acts or omissions of a robot.
二是採取更多舉措來教育用戶，讓他們明白機器人和人工智能帶來的益處和風險。這將有助於確保對技術缺陷做出正確迴應，尤其是在人類不可避免因機器人的行爲或遺漏而死亡的時候。
