It’s a nice commentary about the business and consumer sides of Google Maps. On the personal/consumer side, Google is adding more advertisements when people search for an address – for example, asking if you want delivery via DoorDash if you search for a restaurant. There are also “promoted pins” which are map pins that appear showing an advertiser’s location, even if the advertiser isn’t relevant to the search. On the business side, Google recently increased the cost of using Google Maps embedded into a website/app/mobile app.
I predict we’ll see more ads and more monetization of the Maps service. Google needs to diversify its revenue – most of it currently comes from ads – and Maps is a great source of data. It wouldn’t surprise me to see Google expanding the Google Local service, encouraging more business reviews from customers so they have more local data to display along with Maps.
It’s an interesting and difficult problem to algorithmically categorize videos – in this case, it’s easy to see how the Google AI fouled up: both 9/11 and Notre Dame fires were in tall buildings, with a cityscape surrounding them – from an AI perspective both look very similar. YouTube is such an important source for fresh news, I almost think that they should have people on 24/7 monitoring popular livestreams.
During today’s Notre Dame fire, Google updated the search results for Notre Dame to indicate the emergency. Here is what Google Maps showed (note the little fire icon above the cathedral):
If you Google searched for the fire, or clicked the link from Maps, you’ll see the following screenshot. Note the emergency alert at the top, and the map of the fire’s affected area underneath the news section.
Google recently posted the below Twitter thread, introducing the before: and after: search operators. Now instead of using advanced search to limit searches by date, you can use these operators straight in the search box.
For example, if I need to search for Star Wars content before 2018, I can try:
star wars before:2018
Compare that to a regular search for Star Wars content, which includes news released in this year – 2019.
I use PhantomJSCloud to take screenshots of web apps. While writing new code, I noticed this error coming back from the server:
"message":"Error extracting userRequest. innerException: JSON5: invalid character 'u' at 1:1",
"no stack unless env is DEV or TEST, or logLevel is DEBUG or TRACE"
The problem came because the request wasn’t JSON-encoding the object; the fix looks like this (using the requests library):