Happy November. I spent much of the day running around London on a James Bond experience my family set up ahead of my birthday. Good fun. No martinis though. Yet. 🍸 🇬🇧 Sent from Oxford, England
I Think…🔎 OpenAI Adds Search to ChatGPT, Challenging Google – While I still think the "challenging Google" angle remains a bit much for now, one of the more interesting/aggressive moves is prompting you to install a Chrome extension to make ChatGPT your default search engine. It sort of feels Google Toolbar-y in Internet Explorer back in the day... That said, this still seems more like a "enough queries are looking for real time information, so we need this" type thing rather than a "let's kill Google" type thing. (See also: Meta's potential search engine.) I have access and my early take is that it's an interesting additional layer but it's a bit too cumbersome to replace Google if you're trying to do an actual web search. So this only takes on/out Google if it changes ingrained behavior (i.e. people stop being interested in doing web searches – which is why something that looks like Google was never going to kill Google). The linking/referencing here is pretty good. And while some of it is done through bespoke deals with various publishers, a lot of it is seemingly just using Bing. Regardless, should be fun to watch Google and OpenAI come at this from the opposite angles. [WaPo 🔒] 🍎 Apple’s Q4 Results: $95B Revenue – With a Twist – As always, Jason Snell's charts give a good visual overview of the quarterly trends with Apple. Wearables + Mac + iPad = 25%. Services = 26%. iPhone = 49%. Services continues to chip away at the iPhone's fiscal dominance but that was also one segment slightly under analysts' forecasts. But the real story remains about the iPhone and as you can tell from the earnings call, nearly every question was about Apple Intelligence and how it was/would impact sales. At one point, Tim Cook humorously pointed out that the service runs on iPads and Macs too. No one cared. iPhone. iPhone. iPhone. The fact that guidance is fairly light there does suggest the company doesn't think it will have a big impact in the immediate future – which makes sense since it literally just rolled out, as they kept noting. That hasn't stopped them marketing it though! This was also the swan song for Luca Maestri – the CFO oversaw the company go from "merely" a $600B company to a $3.5T one. Oh yeah and the "twist" above was the $14.4B that the EU forced Apple to pay to Ireland for back taxes. [Six Colors] 🗣️ Alexa’s New AI Brain Is Stuck in the Lab – Reading this account of what's going on behind-the-scenes on the Alexa and AI teams at Amazon is almost stressful. It pairs well with my piece 'Being Too Early Is Worse Than Being Late' which laid out why everything that made Alexa great is now the very reason she's being held back – quite literally, it seems, as this report talks specifically about an event that was to take place on October 17. Instead, we learned about new Kindles – which almost felt like a "cupcake game" as it were, to get new exec Panos Panay ready for the "main event". That event, it seems may not happen until next year now. Andy Jassy promised just yesterday in Amazon's earnings call that we'll hear about an agentic Alexa "in the near future". I do agree with the call to delay, the stakes are too high in the post-ChatGPT world to airball it. [Bloomberg 🔒] 🌨️ AI Demand Drives Growth and Higher Spending on Amazon’s Cloud – Sticking with Amazon, their earnings seemed mostly solid and they look to re-join the $2T club when the market opens as Wall Street clearly likes the overall story (and, of course, the cost-cutting that has been happening). AI still seems more complicated as the headline – and capex spend, $75B! – would seem to indicate that it's going well, but it largely reads like Amazon is benefiting simply because so many businesses already use AWS and so plug into their AI offerings, which mainly means plugging into third-party AI services, and that is leading to more capacity and compute needs. That seems fine for now, and could be the right call, but also leaves AWS more vulnerable than they have been to Google and Microsoft – both of which are still growing their clouds faster (from far smaller bases, naturally) with more bespoke AI offerings. [FT 🔒] 💸 Intel Posts $16.6 Billion Quarterly Loss, Its Biggest Ever – The headline makes it sound like this is the end of Intel, but actually, Wall Street quite liked these results. Mainly because things aren't quite as bad as feared. Still bad. But perhaps starting to show some break in the clouds. Also, the crazy loss was the result of one-time write-downs, taking a hit up-front on restructuring (though a large chunk of the write-down was due to overspend on equipment to meet capacity that was never needed, which was dumb). Meanwhile, the shift to AI is way behind schedule, PC CPUs are down, server chips are doing okay, but the foundry remains a cash incinerator. 2025 will be the make or break (up) year for Intel. And a lot of that rides on the CHIPs act money from the government, with an election in four days, which could also set up a new M&A environment... [NYT] 📝 Writes and Write-Nots – Paul Graham sees a world in which AI pushes writing to the extremes: people who like to write will continue to write while those that don't, won't. Whereas before, you had to do some level of writing for various things, AI will eliminate that. The problem there is that if writing is thinking – something I certainly agree with as it's basically the entire premise for this site – the non-writing part of society will become dumber over time, while those that continue to write will seem that much smarter for it. That would seem like it would be a huge advantage for the writers – and perhaps it becomes a skill even more prized than it currently is – except you could easily see a world in which the two sides grow to dislike and distrust the other just as with wealth, political parties, etc. So we perhaps have that to look forward to. [Paul Graham]
I Wrote…
I Link...- Microsoft recalls the hope that in October you'd recall their Recall feature as it's now November and they're again recalling the preview of the hopefully now secure Recall until December. What a total mess for a feature that seemed set to be Microsoft's big Windows AI jewel. [Verge]
- Forget Peter Todd, Stephen Mollah might be my new favorite would-be Satoshi Nakamoto. This one says he invented the Twitter logo too. [FT 🔒]
- Waymo's valuation? Something just north of $45B. Which still feels cheap in a world where OpenAI is north of $150B, xAI is trying to raise at $40B, and Cybercab is being framed as Tesla's future. What I'm saying is that the valuation should be way... ah, nevermind. [Bloomberg 🔒]
- Comcast decided to leak their own may-or-may-not happen news that they're exploring spinning out their cable networks to stop leaks about it. Also, perhaps to appease Wall Street. Staying put? NBC and Peacock – both of which saw nice bumps thanks to the Olympics. But they're still ready and willing to partner the latter with another streamer to combat Netflix. Will it be Paramount+ or Max? Or will those two partner up leaving Peacock's feathers ruffled? [THR]
- Uber also beat their earnings targets, but there's concern about future growth of the core business as rising prices naturally dampens some usage. (Also, the massive surge in net profit was mainly due to investments being marked up.) Re: that Expedia rumor, "We’re not looking at this point to do big deals, transformational deals, one way or the other." [CNBC]
- The AI elements that Google is adding to Maps actually seem useful – parsing natural language queries, summarizing reviews in aggregate, etc – not just shoving AI into the product to say they did. [Verge]
- Google's Play Store is clearly doing some work to get ready for XR content/headset... [9to5Google]
- I just wanted to take a minute to acknowledge Louis Ashworth's snarky attempt to explain the UK's fiscal "black hole" by referencing Soundgarden songs. [FT 🔒]
- Pixelmator feels like a good get for Apple as it already seemed like an Apple product, their version of Photoshop. Interesting that it needs to clear "regulatory approval" – that either means it's a large deal, well north of $100M, or, more likely, because the company was started in Lithuania, an EU country, Apple has to dance with their old friends once again... [9to5Mac]
- The "Moanapocalpyse" is upon us, with early ticket sales breaking records. It's a favorite in my household too, though I didn't realize Moana was the "most streamed movie across all platforms" last year. Seems even more wild Disney has been willing to wait 8 years between Frozen sequels. [THR]
I Quote..."We have some very good releases coming later this year! Nothing that we are going to call GPT-5, though."
-- Sam Altman, during a Reddit AMA he did yesterday alongside other OpenAI executives. (I went ahead and updated his (non) use of capitalization for my own sanity.) While this seemingly ends the GPT-5-in-2024 debate, I suppose it's possible that the project named 'Orion' is something tangential. Certainly it seems based around "agentic" capabilities, some of which could still come later this year, in the face of increased competition, one imagines...
I Spy...One of these is not like the others... [via Kendall Baker]
|