Hello again from Brussels,
I've been sharing a lot of ideas over here with a lot of excellent folks from the publishing side of the industry. No, I don't plan to publish a DTNS broadsheet anytime soon, but I'm definitely feeling some creative juices flowing.
Alas, I must leave tomorrow morning and head back home. But the good side is getting back to the DTNS crew and all of you. In the meantime here is what I think about some of the tech news we've seen in the last day or so.
Cheers,
Tom
"Google and major mobile carriers want Europe to regulate Apple's iMessage platform"
"Google turns to regulators to make Apple open up iMessage - The Verge"
Short catch up: The Digital Markets Act requires gatekeepers of messaging services to allow inter-operation. To qualify as a gatekeeper you need to have a large user base. Specifically the company must have annual revenues of more than €7.5 billion, and at least 10,000 monthly active business users in the EU. Apple argues its user base for iMessage is too small to qualify. You also don't have to use it. You can keep the app on SMS only if you like. To argue against those points, Google, Vodafone and Orange teamed up to write a letter to the EC saying its big enough and the fact that enriched messages are only available to Apple users of iMessage means you don't have an alternative. So the question is how many users is enough and are enriched messages enough to qualify it as a unique service.
"Things are going from bad to worse for Cruise's robotaxis"
This is a bad story. But not as bad as you may think if you only see the headline. So let me try to clarify. Cruise autonomous vehicles can see children. They just don't know they're children. This has caused a lot of confusion. Cruise vehicles can see a child and see that it's moving. And try to avoid hitting them. But it would be desirable for the system to identify it as a child, since children move in more unpredictable ways than other people. Hence the statement that Cruise made that the problem was that it added "no additional care" around children.
When I see a person at the side of the road I keep an eye on whether they appear to be about to move in front of my car. When I see a child I might pay more attention and even slow down a little since I know children can do impulsive things. Cruise is not doing the machine equivalent of that second thing. No children have been struck by Cruise cars. However in simulations it couldn't always avoid an accident. That's a little deceptive because in simulations autonomous vehicles don't avoid lots of other kinds of accidents either. But the rarity of the accidents is what allows cars to be used on the road. And even the simulations seem to have a better safety record than actual humans.
So do I think this story is bad? Yes. Cruise knows it should recognize children and take additional care and it didn't. That's bad. But I don't think this alone would be a reason to pull the cars from the road.
"YouTube’s ad blocking crackdown is facing a new challenge: privacy laws - The Verge"
Alexander Hanff filed a complaint with the Irish Data Protection Commission that YouTube violates his and other user's privacy when it detects ad blockers. Hanff contends that Adblock detection software is no different from spyware. Adblocker detection works by either analyzing javascript code to see if anything changed or just observing that ad elements weren't loaded. Either way it has to look at the document downloaded by the user. Hanff argues that either method violates Article 5.3 of the ePrivacy directive requires user consent before "storing or accessing information on a user’s device." So anything that analyzes the user page has to happen on the user device.
"Samsung’s Galaxy S24 will likely include on-device generative AI called Samsung Gauss - The Verge"
Samsung showed this off during an AI forum, so this isn't a leak. Gauss is on device. On device is the new hotness in AI. The generative models it can run locally are limited but could make for some interesting feature promotion. Currently Gauss is used for "employee" productivity but it seems confident that ti will show up in next year's Galaxy S24.
"Google is rolling out tools that let advertisers create AI-generated content"
Prepare for people to catch whactever few examples there are of ads with three-armed people or Ghengis Khan in the background or whatever generated oddities make it through.
"Google, Meta, Discord, and more team up to fight child abuse online - The Verge"
This is essentially intelligence sharing. And while it may increase the potential for false positives, overall I think it's fairly non-controversial move. Combating false positives will be an ongoing effort for each platform. And honestly sharing information will help them work to reduce false positives as much as increase the number of true positives caught and prevented.
Interesting Reads
"Device keeps brain alive, functioning separate from body : Newsroom - UT Southwestern, Dallas, Texas"
"Bing Chat is so GPU-hungry, Microsoft will rent Oracle's • The Register"
"Who will write the rules for AI? How nations are racing to regulate artificial intelligence"
"Coupang’s (CPNG) Revenue Beats Estimates After Customer Numbers Hit Record - Bloomberg"