跨国深谈← 返回←Devansh Tandon

① 他说过的话

8 条

第一个是,推荐系统的扩展方式和 LLM 一样,而且这个领域在扩展曲线上还处于非常早期的阶段。
The first is that recommendation systems scale just like LLMs do, and that the field is very early in that scaling curve.
—— Devansh Tandon · [00:30]

指向原始笔记的链接

这条干净且可预测的曲线真正引发了 AI 前沿竞赛,因为你可以预测模型质量和能力的提升会是什么样子,这正是当今支撑巨额资本开支和 AI 建设浪潮的基础。
And this clean and predictable curve is what really set off the race for the AI frontier, because you can forecast what model quality and capability improvements will look like, and this is what’s underwriting the massive CapEx investments and the AI build-out today.
—— Devansh Tandon · [02:42]

指向原始笔记的链接

我不认为这条曲线会消失,但这里的模型开发会因为自动研究而加速,像特征工程这类事情将由智能体而不是真正的 ML 工程师来处理。
I don’t think this curve is going to go away, but model development here will be accelerated with auto-research, and things like feature engineering will be handled by agents rather than real ML engineers.
—— Devansh Tandon · [06:04]

指向原始笔记的链接

这个范式令人兴奋的地方在于,大部分算力是在所有产品表面之间共享的,所以你不必为每个产品表面从头训练单独的模型。
The exciting thing about this paradigm is most of the compute is shared across all of the product surfaces, so you don’t have to train individual models from scratch for every product surface.
—— Devansh Tandon · [09:24]

指向原始笔记的链接

所以我们会看到这个转变,我认为,是从黑盒推荐算法转向给用户更多对算法的控制权,让算法变得更可交互、更可引导。
And so we’re gonna see this shift, I think, from black box recommendations algorithms to giving users more control over their algorithm and algorithms becoming more interactive and steerable.
—— Devansh Tandon · [12:54]

指向原始笔记的链接

我非常兴奋的是,用户可以用语言表达自己的目标来引导算法,而不仅仅是点赞或评论。
I’m really excited that users can direct it towards their own goals that’s expressed in language rather than just likes or comments.
—— Devansh Tandon · [13:07]

指向原始笔记的链接

所以如果你尝试比较内容信息流和 AI 聊天应用,我认为 LLM 推荐器在结构上其实比 AI 聊天更省 token。
And so if you try to compare content feeds and AI chat apps, I think that LLM recommenders are actually structurally more token efficient than the AI chat.
—— Devansh Tandon · [14:46]

指向原始笔记的链接

存在一个巨大的结构性差距:内容信息流要便宜得多,比 AI 聊天应用便宜高达 100 倍或更多,因为它们解码的是指向内容的指针,而不是内容本身。
there’s a huge structural gap where content feeds are significantly cheaper, up to 100 times or more cheaper than AI chat apps because they’re decoding pointers to content rather than content itself.
—— Devansh Tandon · [16:10]

指向原始笔记的链接

② 出现在这些集

1 集

• 《Tokens 进、互动出：推荐系统正在像 LLM 一样扩展》 — 作为嘉宾

③ 他谈到的

点进去有真内容 —— 本页主要出口

Meta · Instagram · YouTube · Spotify · DoorDash · 推荐系统 · LLM · 语义 ID · 算力扩展 · 推理

④ 也在聊「智能体」的人