Topics API demo and colab walkthrough

The Topics API enables interest-based advertising, without resorting to tracking the sites a user visits.

This video shows how topic inference and calculation works, and explains how to use chrome://topics-internal for debugging. You'll also learn how to try out the API's machine learning model using the Topics colab.

Chapters:
0:00 - Introduction
1:25 - Topics taxonomy
1:58 - Topics demo
8:44 - chrome://topics-internals
10:54 - Topics colab
12:22 - Find out more

Resources:

#ChromeDeveloper

[randfur]

Well paced explanation, easy to follow.

[brentmcd]

so it profiles a dataset gives some random input and then that's what the ad tracker can read to give you more suggestive ads but somewhat keeping your privacy more hidden ? interesting I don't know too much about this I can read it but I'm not a code writer, keen to see more like this and about ChromeOS tho ! 😎💻

[Ostap]

Nice detailed overview, as always. Feels very easy to use. Do I understand correctly that the basic difference from 3rd party cookies is that instead of advertizers need to use cookies to track where user has been and create the site-mapping themselves, they will get this info preprocessed by Google? And if user logs into the site that is related to the iframe-provider, user's name can be easily added to the data?