11.12.23: DevDay
本周的主题当然应该是 DevDay。
Ben Thompson 在 The OpenAI Keynote 开头回顾了过去十年来科技产品发布会的迅速衰落,但怀旧情绪并不能把我们带回那个金光闪闪的年代。他把 Apple 的预录制视频当作反例,其内里并非 Tim Cook 缺少面对现场观众的勇气,而是时代不再会对小步快跑的迭代掌声。
Fast forward a decade and the tech keynote has diminished in importance and, in the case of Apple, disappeared completely, replaced by a pre-recorded marketing video. I want to be mad about it, but it makes sense: an iPhone introduction has been diminished not by Apple’s presentation, but rather Apple’s presentations reflect the reality that the most important questions around an iPhone are about marketing tactics. How do you segment the iPhone line? How do you price? What sort of brand affinity are you seeking to build? There, I just summarized the iPhone 15 introduction, and the reality that the smartphone era — The End of the Beginning — is over as far as strategic considerations are concerned. iOS and Android are a given, but what is next and yet unknown?
十年过去了,科技主题演讲的重要性已经减弱,就苹果而言,它完全消失了,取而代之的是预先录制的营销视频。我很想为此生气,但这是有道理的:iPhone 的介绍并没有因为苹果的演示而被削弱,而是苹果的演示反映了这样一个现实:围绕 iPhone 最重要的问题是关于营销策略的。你如何细分 iPhone 产品线?你们如何定价?您希望建立什么样的品牌亲和力?在那里,我刚刚总结了 iPhone 15 的介绍,以及就战略考虑而言智能手机时代——开始的结束——已经结束的现实。 iOS 和 Android 是既定的,但接下来会发生什么呢?
答案是异口同声的:当然是 AI,特别是 OpenAI。
Sam Altman 在开始有些紧张,眼睛不时向下看——很可能是是在寻找舞台上的提词器。舞台不大,更像是一次随意安排的分享活动。Altman 的语调并不夸张的戏剧性,用词也更加平实,更偏爱短句,和简洁的幻灯片十分配合。
Thompson 评价道:
New products and features, meanwhile, were available “today”, not weeks or months in the future, as is increasingly the case for events like I/O or WWDC; everything combined to give a palpable sense of progress and excitement, which, when it comes to AI, is mostly true.
与此同时,新产品和功能「今天」就可用,而不是未来几周或几个月,就像 I/O 或 WWDC 等活动越来越多的情况一样;所有的一切结合起来给人一种明显的进步和兴奋感,对于人工智能来说,这基本上是正确的。
OpenAI 越来越像一家产品公司,而非在年初我们所理解那样,像是一家纯粹的技术公司。这次发布会,正如很多人所批评的那样,和它的名字并不相符——很多人认为 OpenAI DevDay 对开发者并不友好,甚至于认为它足以杀死 Altman 老东家 YCombinator 整个 batch 的创业公司。特别是看到可定制的 GPTs 和 ChatGPT 可以自主选择模型的功能时,OpenAI 希望直接站在最终用户面前的欲望更加明显了。
甚至于 Altman 自己也在演讲中用了这样的表达:
Even though this is a developer conference, we can’t help resist making some improvements to ChatGPT.
尽管这是一次开发者大会,我们还是忍不住要对 ChatGPT 进行一些改进。
是啊,ChatGPT 本身是一个超级应用,当它可以自主选择模型、插件或其它 GPT 的时候,它就接近于一个通用用户界面(Universal Interface)——Ben Thompson 认为,它是自 iPhone 以来最重要的消费者产品:
In this case “browsing” or “image generation” are basically plug-ins: they are specialized capabilities that, before today, you had to explicitly invoke; going forward they will just work. ChatGPT will seamlessly switch between text generation, image generation, and web browsing, without the user needing to change context. What is necessary for the plug-in/GPT idea to ultimately take root is for the same capabilities to be extended broadly: if my conversation involved math, ChatGPT should know to use Wolfram|Alpha on its own, without me adding the plug-in or going to a specialized GPT.
在这种情况下,「浏览」或「图像生成」基本上是插件:它们是专门的功能,在今天之前,您必须显式调用;展望未来,他们只会工作。 ChatGPT 将在文本生成、图像生成和网页浏览之间无缝切换,而无需用户更改上下文。插件/GPT 想法最终扎根的必要条件是广泛扩展相同的功能:如果我的对话涉及数学,ChatGPT 应该知道自己使用 Wolfram|Alpha,而无需我添加插件或者去专门的 GPT。
这听上去既令人惊叹,也让人感到细思级恐。一直以来,平台自身与其参与者之间关系都很微妙:平台自身也面临竞争,也需要证明其商业模式,这就使得它可能会选择垂直整合作为竞争策略,但这可能极大的伤害平台参与者的利益。最简单的理解就是 iOS 不断从 App Store 开发者那里「学习借鉴」,并发布内置在 iOS 系统中的功能或应用。现在,随着 ChatGPT + GPTs 组合的进一步成型,OpenAI 的开发者生态可能会面临巨大的不确定性。
Dan Shipper 在 What I Saw at OpenAI’s Developer Day 中写道:
If OpenAI has to pick between ChatGPT and its developer ecosystem, it’s going to have to pick ChatGPT. ChatGPT is OpenAI’s most valuable source of high-quality training data, so it’s the best way to improve the quality of its models. Not only that, but OpenAI is moving in a direction that commoditizes and consumerizes development work. ChatGPT itself can turn anyone into a semi-competent programmer. And the features it launched yesterday allows anyone to build a chatbot without having to code at all.
如果 OpenAI 必须在 ChatGPT 及其开发者生态系统之间做出选择,那么它就必须选择 ChatGPT。 ChatGPT 是 OpenAI 最有价值的高质量训练数据来源,因此它是提高模型质量的最佳方式。不仅如此,OpenAI 正在朝着开发工作商品化和消费化的方向发展。 ChatGPT 本身可以将任何人变成半称职的程序员。它昨天推出的功能允许任何人构建聊天机器人而无需编写代码。
另一种理解是,OpenAI 正在创造一种新的平台生态模式:在这个模式中,AI 将代替用户做出更多选择,比如调用另一个 GPT,或从某个私有数据源中获取信息。把它和 App Store 模式做一个比较:
- App Store:搜索 / 浏览 → 安装 / 付费 → 使用
- ChatGPT:付费 → 对话 → 使用
有点像短视频或者音乐软件的无限滚动 / 播放,用户不需要做出什么选择,获得了更大的便利性。
让渡选择权,获得便利性,用户让渡的选择权,也就开放了自己的注意力和心智空间,而这几乎是所有科技产品商业模式的基础。类似 OpenAI 这样的平台中,实际上可能有四方(或者更多)参与:开发者、平台、用户(消费者)和广告主,多方利益的平衡是困难的,当平台认为自身的能力足够强大时,就可能希望更深度的控制生态中的用户体验,而开发者作为松散结合的第三方,就可能是被牺牲的一方。
关于 OpenAI 就讲这么多。
Benedict Evans 决定退出 Twitter,不再发表任何推文。在 Leaving Twitter 中他对 Twitter 近年来的发展表达了自己的看法。虽然 Twitter 的各种 clones 没有获得足够大的规模(Threads 除外?),但 Evans 认为,Twitter 的很多功能早已被其它平台替代:
Hence, Mastodon has been around since 2016 without getting much traction, but slices of conversation, content or industry have been unbundled to Reddit, LinkedIn, Instagram, Signal, Discord or, more recently, Substack, which someone joked was Twitter’s paywall.
因此,Mastodon 自 2016 年以来一直没有引起太多关注,但部分对话、内容或行业已被分拆到 Reddit、LinkedIn、Instagram、Signal、Discord 或最近的 Substack,有人开玩笑说 Substack 是 Twitter 的付费专区。
而 Elon Musk 在管理虚假信息上的做法,彻底破坏了人们对于 Twitter / X 的信任:
But it didn’t work out like that. The teams that looked for bots, scammers and state actors were mostly fired, and the scammers, Nazis and propagandists all bought the ‘Blue Ticks’. These little badges used to mean ‘notable person’ (in a chaotic and inconsistent way typical of the old Twitter) and are now supposed only to mean ‘real person’ (but often don’t) - and they give you both amplification in all the algorithms and a share of revenue if you drive a lot of replies. The more you troll, and the more furious replies you generate, the more Twitter promotes you and the more Twitter pays you. We saw this at its logical conclusion in the last week, with deliberate misinformation promoted by what we used to call ‘fake accounts’ that now get promoted by the algorithm because they pay their $8/month. It turns out that social networks are harder than rocket science.
但事情并非如此。寻找机器人、诈骗者和国家行为者的团队大多被解雇,诈骗者、纳粹和宣传人员都购买了「蓝 V」。这些小徽章曾经意味着「名人」(以旧 Twitter 典型的混乱和不一致的方式),而现在被认为仅意味着“真实的人”(但通常不是)——它们给你在所有方面都带来了放大。如果您获得大量回复,则可以使用算法和收入分成。你的恶搞越多,你产生的愤怒回复越多,Twitter 对你的宣传就越多,Twitter 支付给你的钱就越多。我们在上周的逻辑结论中看到了这一点,我们过去所说的「虚假账户」故意宣扬了错误信息,而现在这些「虚假帐户」却被算法所宣传,因为他们每月支付 8 美元。事实证明,社交网络比火箭科学更难。
FastCompany 发表文章回顾 Figma 和 Adobe 的未来:Figma’s future is about far more than design—or getting acquired。这个前景或许因为 AI 而变得更加乐观了,Figma 与生俱来的组件化设计可能和 Adobe AI / LLM 更好的结合,产生全然不同的创意工作流。
Rather than reducing the amount of human effort involved in product creation, such automation might free up time for more fruitful collaboration. To analyst Ray Wang of Constellation Research, that’s an argument in favor of the Adobe-Figma deal. “There are a lot of synergies between the future of what Adobe can deliver and what Figma has, and then the future of what a creative’s going to be doing,” he says. “You’re not going be spending all your time building pieces of content. What you’re going to be doing is assembling that content with other teams.”
这种自动化不仅不会减少产品创建过程中的人力投入,反而可以腾出时间进行更富有成效的协作。对于 Constellation Research 的分析师 Ray Wang 来说,这是支持 Adobe 与 Figma 交易的理由。 「Adobe 所能提供的未来和 Figma 所拥有的,以及创意人员将要做什么的未来之间存在很多协同作用,」他说。 「你不会把所有的时间都花在构建内容上。你要做的就是与其他团队一起组装这些内容。」
Figma 的 CPO Yuhki Yamashita 在 10 月 27 日发表了一则视频,分享了一种充满想象力的未来:
Since announcing the acquisition over a year ago now, we’ve talked a lot about what the future could look like, but we haven’t shown it.
— Yuhki Yamashita (@yuhkiyam) October 26, 2023
So, we came together with @scottbelsky and the @adobe team to imagine “what if ...”
We wanted to share some of our ideas with you. pic.twitter.com/utEWSxoqnK
来自设计师用户的反馈也很正面——但显然他们更喜欢 Figma 而不是 Adobe。
“You know, my fear was ‘What if Figma becomes more like Adobe’—but if the result of this acquisition is ‘What if Adobe becomes more like Figma,’ then I’m actually excited,” commented one designer.
「你知道,我担心的是‘如果 Figma 变得更像 Adobe 会怎么样’——但如果这次收购的结果是‘如果 Adobe 变得更像 Figma 会怎么样’,那么我真的很兴奋,」一位设计师评论道。
Figma 创始人 Dylan Field 对于投后整合非常关心,曾经去寻找 LinkedIn 创始人 Jeff Weiner 的建议:信任就是长期的一致性。
He says that the skeptical gut reaction to the deal among Figma customers encouraged him to seek counsel from LinkedIn’s Weiner, who led that company when Microsoft bought it and left it surprisingly intact. “His advice to me was ‘Trust is consistency over time,’” Field shares. “I even wrote that on my whiteboard at the time and looked at it every day for a while. That’s what we’ve got to do—just keep doing what we’re doing, keep building for [customers], and do that every day. And I think people figure it out if you keep doing that.”
他说,Figma 客户对这笔交易的怀疑本能反应促使他向 LinkedIn 的 Weiner 寻求建议,当微软收购该公司时,Weiner 领导着这家公司,并令人惊讶地完好无损。 「他给我的建议是‘信任就是长期的一致性’,」 Field 分享道。 「我当时甚至把这句话写在白板上,每天都会看一段时间。这就是我们必须做的——继续做我们正在做的事情,继续为 [客户] 建设,每天都这样做。我认为,如果你继续这样做,人们就会明白这一点。」
本周的 Links + Notes 就是这些 。