Peggy K's Creator News: Google Photos Memories, YouTube Live, AI-Generated Summaries

Author

Peggy K
August 19, 2023

Peggy K's Weekly Newsletter Header

Dear readers, I hope you are having an excellent summer and are safe from fires, heat domes, hurricanes and other natural dangers.

This week there are new updates for YouTube creators, including updates to the medical misinformation policy, a music Shorts feed in the YouTube Music app, and new live streaming support for AV1 and HEVC codecs. Google is testing AI-generated “key concept” summaries for YouTube videos and sites on the web.

Google Photos has a new feed with automatically created vertical album-like Memories.

Plus there’s drama at X (formerly Twitter), because isn’t there always?

Creator Weekly Live:
Do you think blocking is a necessary social media feature?
Join the live Creator Weekly on Sunday, 10:30AM Pacific time (5:30PM UTC).

#OnEBoardChat : Managing Online Identities
I’m hosting this week’s #OnEBoardChat on Twitter/X on Sunday at 11AM Pacific time (6PM UTC). The topic is Managing Online Identities.

Follow me on Twitter/X or check out the #OnEBoardChat hashtag to join the discussion (you must be signed in to Twitter/X to see the content on those links).

Take this week’s quiz
Do you know the history of image editing software? Test your knowledge in this week’s YouTube Quiz.

YouTube Testing Automated “Key Concepts”

YouTube is testing “automatic key concepts for learning-related videos”. The idea is that this automatically identifies the “key concepts” and then shows images and text snippets with more information related to those concepts.

This is currently a limited test, with only a small number of English language education videos (school topics like biology or chemistry) showing this information. And only a subset of viewers in the mobile app will see the key concepts.

But the reason why I’m highlighting this limited test is that anyone can opt out individual videos.

  1. Open YouTube Studio (studio.youtube.com) on desktop

  2. Click Content on the left menu

  3. Open the Details for the video you want to change the setting for

  4. Click “Show More”

  5. Deselect “Allow Automatic Concepts”

  6. Save the changes

Google is also testing a similar “SGE while browsing” feature that lets you see “key points” of publicly available web pages. That links each key point to the relevant part of the article. Maybe that’s more similar to YouTube Chapters?

Shorts in the YouTube Music App

The YouTube Music mobile app has a new “Samples” tab, with vertically cropped music video samples in a Shorts-like feed. If you hit on a song you like, you can take several actions:

  • Tap the Play icon to watch the full music video

  • Tap the thumbs up “Like” button to save the song to your library

  • Tap Save to save to a playlist

  • Tap the Shorts icon to switch over to the YouTube app and view Shorts that used the song

  • Tap the Camera icon to create your own Short with the song (not available for all songs)

  • Tap the Share icon to share the video on another platform or get the video link

  • Tap the More (...) icon to

    • start a radio station based on the song

    • play the song next

    • add the song to your queue

    • or go to the album or artist page

This actually seems like an interesting way to find new music. And the option to immediately create a Short sounds nice, since you wouldn’t need to try to find the song again in the Shorts editor (if you can find a video that offers that option).

Update to YouTube’s Medical Misinformation Policy

YouTube has updated and reorganized its medical misinformation policy. It has been boiled down into three types of misinformation: Prevention misinformation, Treatment misinformation, and Denial misinformation. Content that contradicts “health authority guidance” or that denies the existence of specific health conditions is prohibited.

See the YouTube help center article for specific examples.

Content that violates this policy may be removed and result in a Community Guidelines strike. Channels that get three strikes within 90 days, or that severely abuse the policy, may be terminated.

YouTube has noted that they will be focusing on cancer treatment misinformation, removing content that promotes harmful or ineffective treatments or that discourages people from getting medical treatment. They have also created a playlist of cancer-related videos from authoritative sources.

Live Stream to AV1 and HEVC live video to YouTube

This is an update for advanced live streamers. Up until now YouTube only supported the H.264 codec (AVC) for live streaming through the RTMP protocol. Now AV1 and H.265 (HEVC) live video are also supported.

This has been in beta for at least the past few months, so you may have already seen reports of this feature.

Why does this matter? AV1 and HEVC let you stream higher quality video at a lower bitrate. That means higher quality live streams for viewers, and a lower bandwidth requirement for streamers.

If you use a live streaming platform like StreamYard or Restream, implementing this would be a change made on their end.

(If you want to get fancy, you can also live stream 5.1 surround sound for viewers watching on TV and live stream HDR video for “more vibrant and realistic colors”.)

Google Transparency Center Update

Google has launched a new version of their transparency center where you can find policies across Google’s products and services, and easily report a violation. Just go to transparency.google and choose the option to “Explore by Product”.

You can also find the Google privacy policy and terms of service and learn more about their “AI Principles”.

More Video Creator Updates

YouTube Creator Liaison Rene Ritchie was on The Editing Podcast and shared a lot of useful tips. Matt Southern at Search Engine Journal has a summary. Grab your audience immediately, deliver on the promise of your thumbnail, and try to get viewers to come back by leaving them with a cliffhanger.

Twitch is rolling out event prompts that lets viewers engage with the events (like polls) even if they don’t have chat open.

Also on Twitch, if you block or ban someone, they will no longer be able to view your stream.

Google responded to a report from Adalytics that claimed that cookies on “made for kids'' content on YouTube violated government regulations. They go on to detail how they manage ads on content for kids and in the YouTube Kids app. They conclude “not only did Adalytics repeatedly ignore our offers to meet with them, but also their report — either accidentally or intentionally — draws misleading conclusions based on a deeply inaccurate understanding of our privacy practices.”

Web Publishers

Tumblr updated its web interface with improved navigation and labeled buttons (nice!).

Photos on the Web

Flickr’s World Photography Day contest is open, and there are prizes including camera gear, Flickr Pro subscriptions and nice prints. Entries must be submitted by September 8.

In recent updates Flickr has also made it easier to add tags and people to a photo page and to moderate comments. These changes especially improve Flickr on mobile.

Google Photos has introduced a new Memories tab that is “automatically curated and organized with the help of AI.” For several years Photos has shown automatically-created “Memories” at the top of Google Photos, even as it stopped creating albums automatically. Now there is a chronological Memories feed in the Photos app.

  • Save automatically created Memories to the “Memories” view.

  • Save Memories shared with you.

  • Rename memories, optionally using AI-generated title suggestions.

  • Add or remove images or videos from Memories

  • Create your own Memories.

  • Collaborate on Memories with friends and family.

  • And coming soon: share a Memory as a vertical video.

This is initially rolling out in the US.

Social Media

Threads has added a Reposts tab on your profile and Reposts will appear in your Following Feed. Andrew Hutchinson at Social Media Today notes that while active users on Threads has dropped significantly, engagement is high and adding more features (like the desktop version coming “soon”) might increase active users as well.

X (Formerly Twitter)

On Friday, Elon Musk said X will be removing the ability to block people, other than in DMs. It’s not clear if this is just an off-the-cuff response to a fan that will never be implemented or if it will really happen. The spin from CEO Linda Yaccarino says they are “building something better than the current state of block and mute.” But people aren’t happy about the idea. Blocking the head Twit improved my timeline immensely, and for some who face harassment it’s the only way to use the platform at all.

Tweetdeck (now X Pro) is now only available to paid subscribers. This change was announced in July.

Jeremy Merrill and Drew Harwell reported for The Washington Post that X added a 5 second delay when clicking a link to certain competing (or apparently disliked by management) sites. This affected links to the New York Times, Reuters, Facebook, Instagram, Bluesky and Substack. This seems to have been implemented around August 4, when X owner Elon Musk railed against the New York Times for an article it published on South African politics. Shortly after the Washington Post article was published, the delay was removed from at least some of those sites.

Logged out viewers will now see X profiles with the most popular “liked” Tweets/posts at the top, even if they were posted years ago. (I still can’t see my own profile at all when I’m logged out)

The journal Nature surveyed more than 9,000 scientists (or more correctly, sent a survey to 170,000 scientists and 9,200 responded) and found that more than half report spending less time on Twitter/X and 40% have joined new platforms (Mastodon, Bluesky, TikTok, etc).

Mashable published a study of Musks’s followers on X and found that most of his 152 million followers have less than 10 followers and have posted less than 10 Tweets. A significant percentage have zero followers or content. I don’t find this particularly surprising as Musk is a default follow when people sign up, and I believe he has been a recommended follow for a long time. On most platforms there are many people who sign up and then never really use it. And it’s likely that on Twitter/X such people also follow Musk. I’ll note that there isn’t any easy way to tell whether an account is “fake” or just inactive, as real people also use the default profile icon and name (with a bunch of numbers).

Privacy and Security

Chrome browser is moving towards a default HTTPS First Mode. HTTPS links are encrypted for added security. Changes coming soon include automatically navigating to the https:// page, even if a clicked link uses http://. If there isn’t a HTTPS version of a site, it will fall back to HTTP. But that means that a link that is outdated or inadvertently doesn’t use HTTPS will lead to the secure site. As an example of why this could happen, if I type www.google.com in Google Docs, it’s automatically linkified, but uses the HTTP, rather than HTTPS URL. Chrome will also start showing warnings on insecure downloads.

More Generative AI Updates

Google’s experimental Generative AI in Search (GSE) added a few new features this week.

  • Hover over selected words in the AI-generated response to get a definition.

  • Enable the “SGE while Browsing” experiment to get a summary of key points while viewing a freely accessible web page with links to the relevant part of the article.

  • In Search, AI-generated results with segments of code will be color-coded with syntax highlighting to make it easier to read and understand.

Sign up to try SGE at labs.google.com/search/experiments .

John Nack shared a few new updates in Adobe Firefly, including dark mode, performance improvements on mobile, more supported languages, an easier way to send images to Adobe Express, and access through the Web section of Adobe Creative Cloud.

Open AI published an article on using GPT-4 for content moderation. Kyle Wiggers at Yahoo!Finance points out that automated moderation isn’t new (for example Google and Jigsaw’s “Perspective” launched in 2017) and they all suffer from biases created by human trainers. And the appropriateness of a comment or other content is highly context dependent.

Header image background: Sunset by pattama choomsree on Pexels via Canva (free for commercial use)