Apple gets hammered by the EU once again while there’s a threat in the US of breaking up the big tech giants. Google appears to have another problem AI implementation, Steve Wozniak is back as an unlikely critic of the TikTok ban, a new open source AI that runs on your computer an an Amazon executive has a different take on Artificial General Intelligence.
These stories and more on the “Breaking up is hard to do” edition of Hashtag Trending. I’m your host, Jim Love, let’s get into it:
Big tech is taking a big beating on a number of fronts, with some asking if the ultimate end game isn’t the breaking up of some of the giants – Meta, Google and even Apple.
Apple got hammered in the European Unio again this week. The EU announced that it’s not satisfied that Apple’s App Store changes are compliant with the Digital Markets Act and the company is now officially under investigation for non-compliance.
What it really all boils down to is that the EU is not convinced that Apple is complying with their anti-trust laws. These laws are supposed to ensure that tech giants can’t use their market dominance to give their products and services an unfair advantage over their competitors.
Apple has been flagged for some of what the EU regards as anti-competitive behaviour in the way it forced developers to not be able to recommend any payment method other that through the Apple store and the EU also forced Apple to allow for other stores that could sell software for iPhones and other devices.
But each time, the “solutions” that Apple offered were criticized for not, to put it kindly, being anything resembling a true compliance to the EU requirements.
The rules and restrictions were so complex that they would take regulators a lot of time to analyze.
But for some of them, when you worked through the convoluted rules, left Apple competitors worse off than before the changes. For example, one of the provisions they implemented to allow developers to participate in another app store had provisions that could bankrupt smaller developers just for being successful in attracting a lot of downloads.
So the EU is pushing back, big time, opening another investigation into Apple’s potential “non-compliance.”
This is serious stuff in the EU. Apple was fined 1.8 billion dollars for supposedly breaking rules for supposed anti-competitive behaviour in music streaming. But there’s a much bigger potential set of penalties for these latest charges.. Companies can be fined up to 10% of their world-wide sales, or “turnover” is what the legislation calls it, and this increases to 20% for repeat offences.
And to add to the bad news, Apple is under investigation in the US as the Department of Justice launched legal action along with 15 state governments against the company, again for what they feel is anti-competitive behaviour.
But unlike the EU which relies on monetary penalties, the US has had a history of breaking up companies that engage in anti-competitive behaviour. The biggest of these goes back to the breakup of AT&T in the US. AT&T was broken up into 7 different companies, the so-called Baby Bells.
But Google and Meta can’t be smug. The EU is also turning its sights on them. The news release reads:
Today, the Commission has opened non-compliance investigations under the Digital Markets Act (DMA) into Alphabet’s rules on steering in Google Play and self-preferencing on Google Search, Apple’s rules on steering in the App Store and the choice screen for Safari and Meta’s “pay or consent model”.
And while the EU leads in this area, so far the US may take time, but has traditionally moved in the same direction, like it did with Apple.
Sources include: Reuters and 9to5Mac and the BBC
Google’s new AI-powered “Search Generative Experience” is supposed to be a big move forward in AI powered search. But it turns out that this new offering seems to be promoting scams and spam sites.
The new feature gives a summary of search queries followed by recommendations of sites to visit. But it turns out, as one SEO consultant, Lily Ray spotted, those conversational responses are making it easier for people to fall for scams and sites with malicious behaviours.
Bleeping Computer did some of its own testing and found similar results. Many sites that were part of what is called SEO poisoning, where legitimate sites are taken over and used to promote scams with seemingly legitimate ads and links. They can also be taken through a series of redirects to the scam sites.
Whatever the tactic, it appears that Google’s new feature is prone to promoting these sites, making this yet another time that Google has “stepped in it” as it tries to promote its use of AI.
Sources include: Bleeping Computer
An new open source AI caught my attention. I was recently interviewed by Red Hat for a series called Todd Talks where I talked about Open Source AI as potentially democratizing AI by making it available to everyone and then, lo and behold, I saw a story on Jan.
Jan is being described as an Open Source ChatGPT alternative that runs completely offline on a regular computer.
Developed by a team of researchers from Jan Labs intent on democratizing AI, as well as giving users greater control over the privacy of their information, Jan is described as having the “power of ChatGPT locally on the desktop.”
Jan can run on a wide range of different computers from Apple’s M series to Nvidia GPU’s.
Jan’s founders envision AI as “an extension of human intelligence” and have developed it based a four principles:
- Ownership: After being freed from external data tracking or storage, users can maintain complete ownership of Jan. Jan is a tool created by users and for users that promotes independence and self-determination.
- Privacy: Jan prioritizes user privacy by minimizing reliance on external servers, primarily operating in a local context. The user’s device securely stores data, providing unmatched privacy control.
- User Support: Jan has a community-driven development approach, allowing users to access, alter, and add to its codebase. Because of this cooperative approach, Jan is guaranteed to change in response to user requirements and preferences.
- Ethical Design: Jan upholds user welfare and agency by prioritizing ethical design principles over deceptive methods or proprietary lock-ins.
Jan is avaiable under AGPLv3 license, and is being developed via collaborative platforms like Discord and tools like Kanban boards, to foster what the team says is information sharing and creativity.
Support for Python runtimes and mobile platforms are some of the planned additions to
Sources include: Martech Post
Steve Wozniak, the co-founder of Apple Computer is back in the news as an unlikely defender of TikTok. Or at least as someone who thinks that its hypocritical to single out TikTok, while letting big players like Meta and Google are also gathering our data and tracking us. As he said in the CNN interview,
“If you have a principle [that] a person should not be tracked without them knowing it, you apply it the same to every company, or every country. You don’t say, ‘Here’s one case where we’re going to outlaw an app, but we’re not gonna do it in these other cases.’ So I don’t like the hypocrisy, and that’s obviously coming from a political realm.”
He went on to say, “And tracking you – tracking you is questionable. But my gosh, look at what we’re accusing TikTok of, and then go look at Facebook and Google and that’s how they make their businesses,I mean, Facebook was a great idea. But then they make all their money just by tracking you and advertising, and Apple doesn’t really do that so much.”
Wozniak was actually quite complimentary to Apple’s “walled garden” approach as being more protective of user data and privacy.
Wozniak has always been out on the leading edge, as far back as the founding of Apple. He’s led the charge for digital rights as one of the founders of the Electronic Frontier Foundation.
But you might be forgiven for seeing him as an unlikely champion of TikTok. But actually, Wozniak says he prefers prefers it to the other social networks.
In this CNN interview he said, he largely avoids “the social web,” but gets a lot of fun out of watching TikTok “even if it’s just for rescuing dog videos and stuff.”
There was another brighter moment to Wozniak’s interview. He was hospitalized last fall in Mexico City with a stroke and much to the joy of fans (myself included) he looks like he has fully recovered.
Sources include: The Register
What is Artificial General Intelligence. Is it AI that’s smarter than us? Is it AI that is self-aware? Or is it simply, AI that can hold down a job?
The problem is that we don’t have an agreed upon standard for what differentiates what we think of as an algorithm to what we think of as a more human type of intelligence.
The lines are getting blurrier all the time. Autonomous agents, we’ve covered these, are able to plan, learn and execute tasks. If you’ve followed the conversations with Anthropic’s Claude or Inflection’s pi.ai it’s getting harder and harder to convince yourself that there is not some kind of emergent behaviour that is, at least possible.
AGI has a certain fear factor which accompanies it, with visions of it taking over from humans like some bad sci-fi novel. Even a report prepared for the US government warned of the dangers of AGI.
But while might not have agreement on what the definition of Artificial General Intelligence is, but that’s not stopping Amazon from projecting where they think it will play in our future.
To Amazon, AGI may be a more personal thing, not a monolithic intelligence, but instead a series of functions that serve humans.
Vishal Sharma, Amazon’s VP for artificial general intelligence announced at the South by Southwest conference that “the future is a personal AGI for everyone.”
It’s this idea of “ambient intelligence” which is personified, if that’s the right word, but the Alexa assistant. It’s an AI that presents itself when needed and then fades into the background.
Today with Alex, there are something in the neighbourhood of “30 models powering more than 130,000 skills,” Sharma said. And 40% of smart home interactions are initiated by Alexa, according to Amazon.
According to Sharma this is the path that will lead to what he called “embodied AI” – more than simply a speaker in your room.Amazon is experimenting with Astro, an AI powered house robot with the same name as the 1960’s futuristic cartoon the Jetson’s.
But closer to reality are services like Alexa’s “hunches” which can lock your door if you forget to.
Sharma noted that he thought we are still a ways away from AGI and that we might hit a wall in ceiling in that development where we find that our language is too abstract to train AI to achieve AGI.
But it’s brighter future than what we have been hearing, more like the Jetson’s than Terminator.
Sources include: Axios
And that’s our show for today.
Remind your friends that they can get us anywhere you get audio podcasts Google, Apple, Spotify, whereever, and even on their smart speakers – and remind yourself that if you like the podcast, please give us a good review – it matters. And as I’m sure you know, there is a copy of the show notes at itworldcanada.com/podcasts
I’m your host, Jim Love. Have a Wonderful Wozniak Wednesday.