Marzipan πŸ“§

AI's Call to Arms, AI's American Front, AI's Audible Augmentation

Happy Great British Bake Off Day, to those who celebrate – well, those in the UK anyway, as the show won't air in the US (on Netflix) until Friday, per usual. Never in a million years would I have imagined I would like such a show, let alone have a countdown for it coming back each year. I chalk a lot of it up to nostalgia – my wife and I watched it endlessly in the early days after our first child was born and we needed some good, hours-filling content that didn't require following a plot. But some of it is also just how well it's presented. Between the mock-up drawings of the creations to the baker backstories to the handshakes, it's just well done.

Anyway, lots of AI writing and thoughts below...


Some Analysis...

An AI Call to Arms
Sam Altman wraps a plea in some grandiose (and vague) promises…
The Fellowship of the AI
9 companions: the US teams up with the biggest American AI firms…
Amazon Finds a Path for AI to Augment Creatives
Audible’s AI voices can scale audiobooks in a way that wasn’t happening…

Some Thoughts...

πŸ”Š How Sonos Botched an App and Infuriated Its Customers – A lot has been written about the shitshow that has been the Sonos app update, but Dave Lee digs into why it happened. And, at least from my own experience with the app, it rings true. I've long hated the Sonos software. The hardware is great which just makes the software that much more infuriating. Set-up (or re-set-up) was the worst. It felt like a patchwork process that would randomly work sometimes and not at all at other times. All of that seemingly caught up with them in a major way while they were distracted by shiny new hardware, like headphones. Meanwhile the service has one job and it has been failing to do it. [Bloomberg πŸ”’]

πŸ’¬ Spurned by Social Media, Publishers Chase Readers on WhatsApp – In our post-Twitter world, where audiences have fractured into a million social pieces, it's fascinating to watch things evolve to fill such voids. Before moving to the UK, my WhatsApp usage was just about zero. Now it feels like my entire life runs through it (which should be a signal of something to Apple, btw). I didn't expect content delivery to flow through these pipes as well, but it also sort of makes sense – it's... like a newsletter of sorts! Just one built of far more modern rails. Anyway, I created a channel for Spyglass just to add to my link-sharing misery, if that's your thing... [NYT]

πŸ’° Anthropic Has Floated $40 Billion Valuation in Funding Talks – Yeah I mean, why not? Make hay while the sun shines – especially if your fellow farmers are making a lot more hay. If nothing else, given the reports that Microsoft, NVIDIA, and Apple are a major part of OpenAI's round, you have to imagine that Amazon and Google are not going to let that go unanswered – especially if Anthropic really is about to power a new, "Remarkable" Alexa. But at these prices, it will also make it near impossible for either to "hackquire" Anthropic down the road, if it comes to that... [The Information πŸ”’]

πŸ€– Meta's AI Chatbot to Start Speaking in the Voices of Judi Dench, John Cena, Others – The impending launch of these celebrity voices was already known thanks to a couple reports last month, but this adds Cena and Kristen Bell to the roster as well as confirming that the voices are launching tomorrow at Meta's event. I'm still unclear why Meta is pushing this (and clearly paying millions to do so). Amazon tried this years ago with Alexa and it never really went anywhere and they shut it down. And Meta itself tried a different version with different celebrities earlier this year and it lasted just a few months (with Meta eating the millions spent on the talent in that case). Maybe they really just want to upstage OpenAI, showing that they can do celebrity voices without the lawsuits? Or maybe Bell's voice will lure millions of children into the product thinking Anna AI is here... [Reuters]

πŸ“Ί Google TV Streamer: Smarter Than Your Average Set-Top Box – The review (and score) suggests The Verge really likes this new product, but they also complain multiple times about how slow it is when compared to both the Apple TV and even the five-year-old NVIDIA Shield. What they actually seem to like here is the Smart Home integration as they spend way more time on that element than the actual, you know, streaming. I'm still not sold and I'm skeptical on how well this will actually sell. (Especially with a new, faster 'Roku Ultra' also out there in the wild at the same price.) Wake me up when someone creates the killer unified streaming service/interface. [The Verge]

🎢 Apple Music Classical 2.0 Adds Thousands of Full Album Booklets – While this app's mere existence is still slightly confusing to me (I know people are very vocal about why this simply must be separate from Apple Music itself for all the metadata, I'm still unconvinced), now I'm even more confused why album liner notes aren't a part of Apple Music proper?! Some will recall that Apple used to offer such niceties as a part of "iTunes LP" which launched back in 2009, but was discontinued in 2018. Give us back our album art! [MacRumors]


  • Microsoft says it now has 34,000 full-time engineers working on their Secure Future Initiative – aka: Make Microsoft Secure Again (or for the first time). Presumably they're not all just working on security, but making sure it's a focus on whatever they are working on full-time. [The Verge]
  • After five years in hibernation, Montgomery, LoveFrom's official bear mascot is now crawling all over their spartan website, chasing your cursor around. Cute as a button. [Fast Company]
  • It sounds like Apple Intelligence, when it finally launches, is going to require at least 4GB of storage to be able to install (and that's going to grow over time). It's undoubtedly too cynical to think Apple will push this as another reason to upgrade, but, well, they need to find more reasons to upgrade! [9to5 Mac]
  • I just learned that TikTok had a streaming music service from the news that TikTok was shutting down their streaming music service. [Bloomberg πŸ”’]
  • Warner Bros Discovery is using Google's Vertex AI (which is not Gemini AI, because why have one AI product when you can have two or more) to automatically create captions on Max content – but similar to what Amazon is doing with Audible and AI, this is about content that doesn't already have captions right now, such as unscripted shows. [CNBC]
    • Meanwhile, Snap is integrating Gemini for their AI chatbot. While the service has been using OpenAI's tech (and mentioned them in their recent keynote), Google's multimodal features are key here given Snapchat's heavy focus on images and video. [Bloomberg πŸ”’]
    • Meanwhile, Gemini is also set to become a standard part of the Workspace productivity suite for Google, no longer just an add-on, as the AI Wars continue unabated... [The Verge]
  • Another day, another service backing down on its purported ideals under government pressure – this time, Telegram says it will work with France (and other countries) on reporting illegal content, which seems more straightforward than Xitter backing down in Brazil. [Bloomberg πŸ”’]
  • The NYT is following Sirius and others in trying to sell premium podcasts through Apple (and Spotify). If this continues, I think a 'Podcasts+' bundle becomes more likely, and perhaps it leads to a bigger splash from Apple. Services! [Axios]

And I Quote...

"It has become a huge source of traffic actually, larger than X."

-- Marta Planells, the senior director of digital news at Noticias Telemundo (the news arm of Telemundo), talking about the rise of sharing news (and links) via Channels on WhatsApp (per above).

It's wild how badly Xitter shit the bed when it comes to link sharing. In many ways, it was the lifeblood of the service, and they bled that poor little blue bird dry.