In the tech industry, 2023 was a year of transformation.

2023年是科技行业的转型之年。

Spurred by the success of last year’s breakout tech star, ChatGPT, Silicon Valley’s giants rushed to turn themselves into artificial intelligence companies, jamming generative A.I. features into their products and racing to build their own, more powerful A.I. models. They did so while navigating an uncertain tech economy, with layoffs and pivots galore, and while trying to keep their aging business models aloft.

在2022年横空出世的科技业明星ChatGPT的成功鼓舞下，硅谷巨头们迅速转型为人工智能企业，将生成式人工智能功能融入到产品之中，并竞相打造属于自己的、更强大的人工智能模型。与此同时，科技行业经济环境动荡，裁员和业务调整比比皆是，而这些公司还要努力维持陈旧商业模式的运转。

Not everything went smoothly. There were misbehaving chatbots, crypto foibles and bank failures. And then in November, ChatGPT’s maker, OpenAI, melted down (and quickly reconstituted itself) over a failed boardroom coup, proving once and for all that there’s no such thing in tech as resting on your laurels.

Every December in my Good Tech Awards column, I try to neutralize my own negativity bias by highlighting a few lesser-known tech projects that struck me as beneficial. This year, as you’ll see, many of the awards involve artificial intelligence, but my goal was to sidestep the polarized debates about whether A.I. will destroy the world or save it and instead focus on the here and now. What is A.I. good for today? Whom is it helping? What kinds of important breakthroughs are already being made with A.I. as a catalyst?

每年12月，我都会在“优秀科技大赏”栏目中着重介绍几个在我看来颇有益处但不太为人所知的科技项目，以此来中和自己的负面偏见。正如你们将看到的，今年许多奖项都与人工智能有关，但我希望回避人工智能到底是要毁灭还是拯救世界的极端争论，而是专注于当下。人工智能对今天的世界有何益处？它正在对谁起到帮助？人工智能已经催生了哪些重要突破？

As always, my award criteria are vague and subjective, and no actual trophies or prizes are involved. These are just small, personal blurbs of appreciation for a few tech projects I thought had real, obvious value to humanity in 2023.

和以往一样，我的评奖标准模糊且主观，也不会颁发真的奖杯或奖赏。本文只是我个人对一些我认为在2023年对人类产生了切实且明确价值的科技项目的小小赞美。

Accessibility — the term for making tech products more usable by people with disabilities — has been an underappreciated area of improvement this year. Several recent advances in artificial intelligence — such as multimodal A.I. models that can interpret images and turn text into speech — have made it possible for tech companies to build new features for disabled users. This is, I’d argue, an unambiguously good use of A.I., and an area where people’s lives are already improving in meaningful ways.

无障碍是让残障人士更方便地使用科技产品的专用语，而这一领域在今年取得的技术进步是被忽视的。人工智能近来取得的进展——比如能够理解图像并将其转换为语音的多模态人工智能模型——让科技公司为残障用户开发新功能成为了可能。我认为这无疑是对人工智能的善用，也是人们的生活质量正在迎来实质性提升的领域之一。

I asked Steven Aquino, a freelance journalist who specializes in accessible tech, to recommend his top accessibility breakthroughs of 2023. He recommended Be My Eyes, a company that makes technology for people with impaired vision. In 2023, Be My Eyes announced a feature known as Be My AI, powered by OpenAI’s technology, that allows blind and low-sighted people to aim their smartphone camera at an object and have that object described for them in natural language.

我请专门研究无障碍技术的自由记者史蒂文·阿基诺推荐了他心中2023年无障碍技术的最重要突破。他选择了Be My Eyes这家为视障人士研发科技的公司。2023年，Be My Eyes推出了一项名为“Be My AI”的功能，在OpenAI的技术支持下，该功能允许盲人和弱视者将智能手机摄像头对准某个物体，然后用自然语言为他们描述该物体。

Mr. Aquino also pointed me to Apple’s new Personal Voice feature, which is built into iOS 17 and uses A.I. voice-cloning technology to create a synthetic version of a user’s voice. The feature was designed for people who are at risk of losing their ability to speak, such as those with a recent diagnosis of amyotrophic lateral sclerosis or another degenerative disease, and gives them a way to preserve their speaking voice so that their friends, relatives and loved ones can hear from them long into the future.

I’ll throw in one more promising accessibility breakthrough: A research team at the University of Texas at Austin announced this year that it had used A.I. to develop a “noninvasive language decoder” that can translate thoughts into speech — read people’s minds, essentially. This kind of technology, which uses an A.I. language model to decode brain activity from fMRI scans, sounds like science fiction. But it could make it easier for people with speech loss or paralysis to communicate. And it doesn’t require putting an A.I. chip in your brain, which is an added bonus.

When CRISPR, the Nobel Prize-winning gene editing tool, broke into public consciousness a decade ago, doomsayers predicted that it might lead to a dystopian world of gene-edited “designer babies” and nightmare eugenics experiments. Instead, the technology has been allowing scientists to make steady progress toward treating a number of harrowing diseases.

十年前，当荣获诺尔贝奖的基因编辑工具CRISPR进入公众视野之时，唱衰者预言它将带来一个属于基因编辑“订制婴儿”和优生学实验噩梦的反乌托邦世界。然而，这项技术让科学家在治疗许多痛苦疾病方面取得了稳步进展。

In December, the Food and Drug Administration approved the first gene-editing therapy for humans — a treatment for sickle cell disease, called Exa-cel, that was jointly developed by Vertex Pharmaceuticals of Boston and CRISPR Therapeutics of Switzerland.

12月，美国食品药品监督管理局批准了首个用于人类的基因编辑疗法，即由波士顿福泰制药和瑞士CRISPR医疗公司联合研发的“Exa-cel”镰状细胞疾病疗法。

Exa-cel uses CRISPR to edit the gene responsible for sickle cell, a debilitating blood disease that affects roughly 100,000 Americans, most of whom are Black. While it’s still wildly expensive and difficult to administer, the treatment offers new hope to sickle cell patients who have access to it.

“Exa-cel”疗法利用CRISPR工具来编辑镰状细胞的基因，这种导致人体衰弱的血液疾病影响了大约10万美国人，其中大多数是黑人。虽然该疗法仍然价格高昂且难以施用，但确实为有能力获得治疗的镰状细胞病患者带来了新希望。

One of the most fun interviews I did on my podcast this year was with Brent Seales, a professor at the University of Kentucky who has spent the past two decades trying to decipher a set of ancient papyrus manuscripts known as the Herculaneum Scrolls. The scrolls, which belonged to a library owned by Julius Caesar’s father-in-law, were buried under a mountain of ash in 79 A.D. during the eruption of Mount Vesuvius. They were so thoroughly carbonized that they couldn’t be opened without ruining them.

与肯塔基大学教授布伦特·西尔斯的对话是我今年做过最有意思的播客采访之一，过去二十年间，西尔斯一直在尝试破译一套古老的纸莎草手稿，即赫库兰尼姆卷轴。该卷轴原本藏于尤利乌斯·恺撒的岳父的图书馆，公元79年维苏威火山爆发时被埋于灰烬之下。这些卷轴已经完全碳化，一旦打开就将损毁。

Now, A.I. has made it possible to read these scrolls without opening them. And this year, Dr. Seales teamed up with two tech investors, Nat Friedman and Daniel Gross, to launch the Vesuvius Challenge — offering prizes of up to $1 million to anyone who successfully deciphers the scrolls.

现在，人工智能让人们不用打开卷轴就能阅读它们。今年，西尔斯与两位科技投资者纳特·弗里德曼和丹尼尔·格罗斯合作，发起了“维苏威挑战”——为成功破译古卷的人提供高达100万美元的奖金。

The grand prize has still not been won. But the competition sparked a frenzy of interest from amateur history buffs, and this year a 21-year-old computer science student, Luke Farritor, won a $40,000 intermediate prize for deciphering a single word — “purple” — from one of the scrolls. I love the idea of using A.I. to unlock wisdom from the ancient past, and I love the public-minded spirit of this competition.

大奖得主尚未产生。但比赛引发了业余历史爱好者的狂热，今年，21岁的计算机科学专业学生卢克·法里托破译了一幅卷轴上的一个词“紫色”，赢得了4万美元的二等奖。我喜欢利用人工智能来开启古老智慧的想法，我也喜欢这项竞赛面向公众的态度。

I spent a lot of time in 2023 being shuttled around San Francisco in self-driving cars. Robot taxis are a controversial technology — and there are still plenty of kinks to be worked out — but for the most part I buy the idea that self-driving cars will ultimately make our roads safer by replacing fallible, distracted human drivers with always-alert A.I. chauffeurs.

2023年，我花了很多时间乘坐自动驾驶汽车，在旧金山各地穿行。机器人出租车是一项有争议的技术——它还有很多问题需要解决——但在很大程度上，我相信自动驾驶汽车最终会让道路更安全，因为它用时刻警觉的人工智能司机取代了容易犯错、容易分心的人类司机。

Cruise, one of the two companies that were giving robot taxi rides in San Francisco, has imploded in recent days, after one of its vehicles struck and dragged a woman who had been hit by another car. California regulators said the company had misled them about the incident; Cruise pulled its cars from the streets, and its chief executive, Kyle Vogt, stepped down.

But not all self-driving cars are created equal, and this year I was grateful for the comparatively slow, methodical approach taken by Cruise’s competitor, Waymo.

但并不是所有的自动驾驶汽车都是一样的，今年，我很感激Cruise的竞争对手Waymo采取了相对缓慢、有条不紊的方法。

Waymo, which was spun out of Google in 2016, has been logging miles on public roads for more than a decade, and it shows. The half-dozen rides I took in Waymo cars this year felt safer and smoother than the Cruise rides I took. And Waymo’s safety data is compelling: According to a study the company conducted with Swiss Re, an insurance firm, in 3.8 million self-driving miles Waymo’s cars were significantly less likely to cause property damage than human-driven cars, and led to no bodily injury claims whatsoever.

Waymo于2016年从谷歌分拆出来，十多年来一直在公共道路上行驶。今年我六次乘坐Waymo的车，感觉比乘坐Cruise更安全、更平稳。Waymo的安全数据也很有说服力：根据该公司与瑞士再保险公司进行的一项研究，在610万公里以上的自动驾驶里程中，Waymo的汽车造成财产损失的可能性明显低于人类驾驶的汽车，也没有导致任何人身伤害索赔。

I’ll put my cards on the table: I like self-driving cars, and I think society will be better off once they’re widespread. But they have to be safe, and Waymo’s slow-and-steady approach seems better suited to the task.

实话说，我喜欢自动驾驶汽车，我认为一旦它们得到普及，社会会变得更好。但它们必须是安全的，Waymo缓慢稳定的方法似乎更适合这项任务。

One of the more surprising — and, to my mind, heartening — tech trends of 2023 was seeing governments around the world get involved in trying to understand and regulate A.I.

在我看来，2023年最令人惊讶、也是最振奋人心的科技趋势之一，是世界各国政府都在努力理解和监管人工智能。

But all that involvement requires work — and in the United States, a lot of that work has fallen to the National Institute of Standards and Technology, a small federal agency that was previously better known for things like making sure clocks and scales were properly calibrated.

但所有这些参与都需要工作——在美国，很多工作都落在了国家标准与技术研究院的肩上，这是一个小型联邦机构，以前是以校准时钟和称量设备之类的事而为人所知。

The Biden administration’s executive order on artificial intelligence, released in October, designated NIST as one of the primary federal agencies responsible for keeping tabs on A.I. progress and mitigating its risks. The order directs the agency to develop ways of testing A.I. systems for safety, come up with exercises to help A.I. companies identify potentially harmful uses of their products, and produce research and guidelines for watermarking A.I.-generated content, among other things.

去年10月，拜登政府发布了关于人工智能的行政命令，指定NIST为负责密切关注人工智能进展并降低其风险的主要联邦机构之一。它指示该机构开发测试人工智能系统安全性的方法，制定帮助人工智能公司识别其产品潜在有害用途的演练方法，并为人工智能生成的内容加水印等问题进行研究和指导。

NIST, which employs about 3,400 people and has an annual budget of $1.24 billion, is tiny compared with other federal agencies doing critical safety work. (For scale: The Department of Homeland Security has an annual budget of nearly $100 billion.) But it’s important that the government build up its own A.I. capabilities to effectively regulate the advances being made by private-sector A.I. labs, and we’ll need to invest more in the work being done by NIST and other agencies in order to give ourselves a fighting chance.

有大约3400名雇员的NIST年度预算为12.4亿美元，与其他从事关键安全工作的联邦机构相比，它的规模很小。（相比之下，国土安全部的年度预算接近1000亿美元。）但重要的是，政府要建立自己的人工智能能力，以有效地监管私营人工智能实验室的进步，我们需要在NIST和其他机构正在做的工作上投入更多，这样我们才有克服这个困难的机会。

And on that note: Happy holidays, and see you next year!

最后，祝大家节日快乐，明年再见！