惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Hacker News: Ask HN
Hacker News: Ask HN
月光博客
月光博客
Martin Fowler
Martin Fowler
人人都是产品经理
人人都是产品经理
有赞技术团队
有赞技术团队
Hugging Face - Blog
Hugging Face - Blog
T
Tailwind CSS Blog
爱范儿
爱范儿
博客园_首页
Last Week in AI
Last Week in AI
N
Netflix TechBlog - Medium
NISL@THU
NISL@THU
H
Help Net Security
A
Arctic Wolf
K
Kaspersky official blog
Y
Y Combinator Blog
G
GRAHAM CLULEY
T
Threatpost
I
Intezer
Attack and Defense Labs
Attack and Defense Labs
Jina AI
Jina AI
Recent Announcements
Recent Announcements
D
Docker
博客园 - 聂微东
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
www.infosecurity-magazine.com
www.infosecurity-magazine.com
GbyAI
GbyAI
The Hacker News
The Hacker News
C
Cyber Attacks, Cyber Crime and Cyber Security
Help Net Security
Help Net Security
Microsoft Azure Blog
Microsoft Azure Blog
G
Google Developers Blog
F
Fortinet All Blogs
S
Security Affairs
Security Archives - TechRepublic
Security Archives - TechRepublic
量子位
AWS News Blog
AWS News Blog
Google DeepMind News
Google DeepMind News
罗磊的独立博客
V
Vulnerabilities – Threatpost
阮一峰的网络日志
阮一峰的网络日志
T
Tor Project blog
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
H
Hacker News: Front Page
Security Latest
Security Latest
The Cloudflare Blog
cs.CV updates on arXiv.org
cs.CV updates on arXiv.org
M
MIT News - Artificial intelligence
B
Blog
云风的 BLOG
云风的 BLOG

Matthias Ott

Hello Again, World This, Still Not for Everyone The Shape of Friction WeissKlang L1 – Punching Above Its Weight Continvoucly Morged Value Webspace Invaders To Affinity and Beyond The Mystery of Storytelling Amateurs! Echoes of Connection Linear() Is Not (That) Linear View Transitions: The Smooth Parts Adding AVIF and WebP Support to My Craft CMS Site Challenge Acoustic Room Treatment and Building Sound Panels, Part 1: Planning Play On Overshoot The HTML Output Element Listening Closely Compressed Fluid Typography The Lifeblood of the Web What Could Go Wrong? That’s My Rank Making Space CSS :is() :where() the Magic Happens Visual Regression Testing for External URLs With Playwright Jane Goodall’s Famous Last Words European Tech Alternatives 🇪🇺 Independent Type Foundry Advent Calendar – Day 24: NaN Independent Type Foundry Advent Calendar – Day 23: Typotheque Independent Type Foundry Advent Calendar – Day 22: 205TF Independent Type Foundry Advent Calendar – Day 21: HvD Fonts Independent Type Foundry Advent Calendar – Day 20: Frere-Jones Type Independent Type Foundry Advent Calendar – Day 19: Fontwerk Independent Type Foundry Advent Calendar – Day 18: Vectro Independent Type Foundry Advent Calendar – Day 17: Studio René Bieder Independent Type Foundry Advent Calendar – Day 16: R-Typography Independent Type Foundry Advent Calendar – Day 15: David Jonathan Ross Independent Type Foundry Advent Calendar – Day 14: Interval Type Independent Type Foundry Advent Calendar – Day 13: Newglyph Independent Type Foundry Advent Calendar – Day 12: Swiss Typefaces Independent Type Foundry Advent Calendar – Day 11: Sharp Type Independent Type Foundry Advent Calendar – Day 10: Colophon Foundry Independent Type Foundry Advent Calendar – Day 9: Commercial Type Independent Type Foundry Advent Calendar – Day 8: Letters from Sweden Independent Type Foundry Advent Calendar – Day 7: Lineto Independent Type Foundry Advent Calendar – Day 6: Ohno Type Company Independent Type Foundry Advent Calendar – Day 5: Milieu Grotesque Independent Type Foundry Advent Calendar – Day 4: TypeMates Independent Type Foundry Advent Calendar – Day 3: Klim Type Foundry Independent Type Foundry Advent Calendar – Day 2: Dinamo Independent Type Foundry Advent Calendar – Day 1: Grilli Type The Independent Type Foundry Advent Calendar 2022 A Conversation With ChatGPT ChatGPT, please explain websites in the words of William Shakespeare Transient Frameworks Leaving Twitter Behind Converting Your Twitter Archive to Markdown The Wrong Question It Wasn’t Written Syndicating Posts from Your Personal Website to Twitter and Mastodon Suspension None of Your Business Doing Our Part Patch That Package Brain Dump Generating Accessibility Test Results for a Whole Website With Evaluatory The CSS Cascade, a Deep Dive Updates About Updates How to Delete Your Commit History in Git Unblocking Your Writing Blocks, Part 2: I’m Not an Expert nor a “Thought Leader” Connections No Wrong Notes Better Options Design Debt Finite and Infinite Games Don’t Assume, Validate. Necessity Is the Ultimate Teacher One Egg Go Deep There Is No Secret Code Balancing Risk Blue Eyes, Brown Eyes The Shortcut Boomerang My RSS Feed Collection of Personal Websites Frequency The Illusion of Control The Decisions Journey Write It Down Nownownow Into the Personal-Website-Verse Considering the Opposite What is it for? Unlimited Bowling. Never done. We Are Team Internet. We Need to Save #NetNeutrality. Progressive Search Data loss (also) by JavaScript Books I Will Definitely Maybe Read in 2017 Starting to Write Notes
Ad Infinitum
Matthias Ott · 2026-05-21 · via Matthias Ott

At Google I/O this week, the company announced the biggest change to Search in 25 years. The ten blue links? Gone. Instead, you get — first at times? soon always? — “generative UI”, an “intelligent search box” with custom interactive widgets, built on the fly by Gemini. You get “information agents” that monitor the web for you around the clock. You get mini-apps you can build right inside the search box, using natural language. And with Gemini Spark, you get a personal AI agent that runs 24/7, connected to your Gmail, your Drive, your calendar, your photos, and soon your local files and third-party services.

It’s ambitious and also impressive at times, no question. It’s also a staggering gamble. Google is dismantling the very product that built its empire – and rebuilding it into something entirely different. Watching the demos, I kept asking myself two questions. First: what if people don’t actually want to use search this way? And second:

Where did the ads go?

Google made $295 billion in advertising revenue last year, out of $403 billion in total. And yet, across all these announcements – the new search box, the generative UI, Spark, the agents – there wasn’t a single mention of how any of this will be monetized. Not one word about ads.

That seemed odd. I mentioned this on Mastodon, and Sijmen replied with a link that sent me down a rabbit hole.

It turns out there’s a small but growing body of research – some of it from Google’s own researchers – exploring exactly how ads could work inside LLM-generated output. And the ideas are as fascinating as they are unsettling.

One approach, proposed by Google Research, is what you might call a “token auction.” In this model, advertisers don’t buy ad slots on a page. Instead, they bid, token by token, on the actual text the model generates. Each advertiser brings their own LLM, and an auction mechanism decides whose model gets to influence the next word. The output is a weighted blend of competing interests, shaped by who’s willing to pay more.

Another approach – also from Google researchers – fits the new “Search” much more precisely. It’s called “prominence allocation.” Here, when a user submits a query with commercial intent, the system runs an auction that doesn’t just decide which ads appear, but how prominently the LLM writes about each one. The auction outputs a prominence score for each advertiser, essentially telling the model: give this product 35 words, that one 20, and this one zero. The ad isn’t next to the answer. The ad is the answer. Or rather, it shapes how much space and enthusiasm each product gets within the answer.

Now think about what Google just announced. Generative UI means there are no more discrete ad slots. The search result is a dynamically generated experience – a synthesized narrative, an interactive widget, a curated set of recommendations. In that world, you can’t place a banner ad next to the output. The only thing you can auction is prominence within it.

But it’s not just the ads that disappeared from the keynote. The links did, too.

For 25 years, Google Search was built on a contract. The web provided the content – billions of pages, freely linked, freely crawled. In return, Google sent people back. The link was the unit of exchange. It’s what made the Web thrive as an information system: you publish, Google indexes, users click through, and value flows back to the source. Win-win.

That contract is now broken. Generative UI doesn’t link to your article, necessarily. It absorbs your article, synthesizes it into a widget, and presents it as Google’s own answer. Information agents don’t send users to websites. They deliver “synthesized updates” with maybe a link or two buried at the bottom. The web was the scaffolding Google needed to build its index, to train its models, to accumulate the world’s information, and put ads next to it to get filthy rich. Now that the content is inside the system, the scaffolding is no longer needed. Google is creating its own context.

Google thinks it no longer needs the Web to deliver answers. And it no longer needs ad slots to deliver ads. What it needs is you. Your emails, your files, your calendar, your purchase history, your travel plans – all flowing into Spark, all building the richest possible picture of who you are and what you’re likely to click on. That’s exactly the kind of personal context those auction models need to work. The prediction module in the prominence allocation framework doesn’t run on keywords. It runs on knowing you.

None of this was mentioned at I/O. The entire keynote was framed as a story about usefulness and possibility and trust. Will you opt in to sharing your data? Will you let Spark read your inbox? The pitch is productivity. But the infrastructure being built – a deeply personalized agent, connected to commercial services, generating natural-language recommendations with no visible boundary between organic and paid content – is a monetization engine.

Then, one day later, Google held Marketing Live, its annual keynote for advertisers. Same campus, same week, different audience. And there, the language changed completely. Vidhya Srinivasan, VP of Google Ads, put it plainly: “The best ads must be answers.” Ads are already running inside AI Mode. They’re not banners next to the output. They’re generated by Gemini to read as part of the conversation. And advertisers who want to appear in the new AI search? They must hand over creative and targeting control to Google’s system. “You can’t choose keywords anymore,” Srinivasan said.

Edit: Google’s own blog post now spells out exactly what this looks like in practice. When AI Mode gives you a list of recommendations, paid placements can now appear as items on that list. Google calls them “Highlighted Answers.” When you search for a product, Gemini writes a custom explainer for the advertiser, framed as objective advice about why this product “may be the right choice for you.” The language throughout is about being “helpful” and “building trust.” The mechanism is paid content inside AI-generated recommendations, written by the same model that’s supposed to give you neutral answers.

So yes, Google didn’t talk about ads in the I/O keynote. Everything they announced is the new ad system. One that, in the eyes of Google, no longer needs the open web as an intermediary. And now, they’re gradually turning it on.

But there’s still that first question. The one about whether people actually want any of this. The people building these systems seem so deep inside their own world that they sometimes forget to check. You might have heard that, over the past few days, commencement speakers across the US, including former Google CEO Eric Schmidt, have been booed off stage the moment they mentioned AI. Not just politely ignored. Booed.

Google is betting its entire future on a world where people eagerly hand over their emails, their files, their habits, and their trust – to an AI system that will quietly auction off their attention, word by word, to the highest bidder. The ad infrastructure is ready. The question is whether the audience is.

~