2024-11-23 04:02:09
I'm biased because I wrote it, but I think this is a pretty good encapsulation of what people are – and are not – talking about with regard to Google's selling off/spinning out Chrome. That's not going to happen, but it does raise a number of points worth actually thinking about and through...
M.G. Siegler
💰 Amazon Invests Another $4 billion in Anthropic – Well, they took the deal. Beyond the money, the key here is that the startup is going to use Amazon's Trainium and Inferentia chips "to train and deploy its largest AI models." This is a little slippery as it's not clear if that means they'll use those chips instead of NVIDIA ones – presumably not, but Amazon's own "primary training partner" wording implies they'll being using the chips first and foremost. Or perhaps they simply mean that Anthropic will now be the largest user of those chips, but much of the training work will continue on NVIDIA? I suspect we'll hear more about this soon as it would obviously be a massive deal if Anthropic was moving the training of Claude from NVIDIA to Amazon's chips – even just most of the training. Frankly, even some of the training. Are those chips really comparable for a company that is in an arms race with OpenAI and Google and everyone else? The $8B Amazon has invested in Anthropic is approaching the almost $14B Microsoft has into OpenAI. Both for minority positions (not that they would want majority positions as regulators would not like that – have they mentioned there's no board seat here? lol) – presumably Amazon's position is far more in the minority given the later entry point, but unlike OpenAI, Anthropic doesn't seem to like to tout their valuation. What does Google, another Anthropic partner with billions invested here think of this? Will more money be coming from them to ensure GCP has at least one seat at the table too, if not in the board room? Back to work everyone, we have a "Remarkable" Alexa to ship. [CNBC]
❌ Mass X-odus: Professionals Desert Elon Musk’s Network – Xwitter has been in a state of decline for a long time now. Beyond the anecdotes we all know and constantly hear, the actual data from multiple sources seems to back this up as well. The only data that doesn't comes from Xitter itself, but they're an unreliable narrator at best seemingly with weasel-y interpretations of usage (how much activity is bot/spam-related now?). Still, when something big is happening in real-time, many rush back to the one service that nails this. But yesterday was interesting in that the Matt Gaetz shitshow seemed to play out in real time just as well on Bluesky as it did on Xitter. Perhaps even better minus the spam and nonsense. If Xitter loses that crown, watch out below. The ultimate Faustian bargain may have been made in steering the network explicitly to help with one end. And it has perhaps given people the means to leave, finally. [FT 🔒]
🔎 The Startup Powering Perplexity, Meta's AI Search – I feel like this has been massively underplayed, but while Bing may be powering ChatGPT's new search abilities, it's the startup Brave which is powering such functionality for many of the other AI startups – and Meta. And yes, Brave is the startup co-founded and run by Brendan Eich, the cofounder of Mozilla, the makers of Firefox. (And co-creator of JavaScript.) That feels like an interesting position to be in – especially with all the renewed talk around search deals, and web browsers for that matter! [Information 🔒]
🙏 In Memory of Enrique Allen – A lovely, thoughtful bit of writing from Craig Mod in honoring his good friend and former roommate, who passed away at the far, far, far too young age of 38. I only had the privilege of meeting Allen a couple of times over the years as we overlapped on a handful of deals, but Mod's piece feels like a window into his life and makes it clear why the outpouring of love for him on social media these past few days was overwhelming. Hugs. RIP. [Roden]
Apple seems to think they can take 18-months or two years to ship in AI...
Who is going to tell them?
(I am.)
M.G. Siegler
"Turns out for the 'Twitter Files' crew, 'creeping authoritarianism' isn’t so creepy when it’s your team doing the creeping."
-- Mike Masnick, in an op-ed for MSNBC, pointing out the hypocrisy from those up in arms when they believed the Biden administration was manipulating Twitter and the other social networks to silence certain speech (COVID information, Hunter Biden's laptop, etc) who now are completely silent with regard to Elon Musk and crew manipulating Xitter to push various other aims.
It's one of those things that is so overt, the strategy, in so far as there is one, is clearly just "move along, never mention it". And sadly, that generally works, especially in our age where news and information move so fast.
A daily newsletter by M.G. Siegler
Sent from afar on weekdays...
2024-11-23 03:38:56
The latest from Gurman starts promising enough...
Apple Inc. is racing to develop a more conversational version of its Siri digital assistant, aiming to catch up with OpenAI’s ChatGPT and other voice services, according to people with knowledge of the matter.
"Racing"! Good! Fire lit!
The new Siri, details of which haven’t been reported, uses more advanced large language models, or LLMs, to allow for back-and-forth conversations, said the people, who asked not to be identified because the effort hasn’t been announced. The system also can handle more sophisticated requests in a quicker fashion, they said.
Yep, sounds great. The first stabs at this by OpenAI and Google are pretty interesting, but early. Seems like something Apple could nail. It's UI, after all, just a vocal kind...
The new voice assistant, which will eventually be added to Apple Intelligence, is dubbed “LLM Siri” by those working on it. LLMs — a building block of generative AI — gorge on massive amounts of data in order to identify patterns and answer questions.
Yeah, I mean that name needs work, even as a codename, but cool, it's clear what it's all about and indicates that Apple is taking moving such capabilities in-house (versus just outsourcing them to OpenAI and the like) seriously.
Apple has been testing the upgraded software on iPhones, iPads and Macs as a separate app, but the technology will ultimately replace the Siri interface that users rely on today. The company is planning to announce the overhaul as soon as 2025 as part of the upcoming iOS 19 and macOS 16 software updates, which are internally named Luck and Cheer, the people said.
Testing already?! Amazing. We're ready and waiting Apple. Good on ya. 2025 is mere weeks away...
Like Apple Intelligence this fall, the new features won’t immediately be included in next year’s crop of hardware devices. Instead, Apple is currently planning to release the new Siri to consumers as early as spring 2026, about a year and a half from now. Given that Apple is still several months away from announcing the plan, the timing and features could still shift.
<record screech>
2026? What the fuck? Why not 2040?
Apple. Dear, sweet Apple. Open your eyes. WAKE UP. By 2026, OpenAI, Google, Anthropic, Amazon, Microsoft, xAI, honestly probably even Cisco – no offense, Cisco – is going to have run laps around whatever it is that you're cooking up. Because what you're cooking up is competitive to what is state of the art right now. But this field is moving so incredibly fast that there's no chance it's state of the art in – checks calendar – two years.
I'm not saying we're going to have humanoid robots by then – Elon Musk may be saying that, but he says a lot of things – but we're obviously going to be a few steps beyond this type of conversational AI by then. Hopefully it's a part of the arsenal, but it really should be next year. Not 2026.1
Yes, yes, "not first, but best". This isn't 2004. It's not even 2022. This is a different world. The cadence done changed. You have to adapt.
Culturally, that will be hard, I get that. But it can be done. Money never sleeps, pal. Astonish me, pal. This is your wake up call, pal. Go to work.
1 Related: those smart glasses? Next year. Not 2026.
2024-11-22 21:50:13
In this whole debate about the demand from the US Department of Justice that Google sells and/or spins-off Chrome it feels like we're talking right past a few obvious elements and realities.
The first is that a web browser remains the most-used and thus, most important piece of software for nearly everyone on the planet. This is especially true on desktops/laptops, but on mobile too as it remains the common denominator in our app environment. Through that lens alone, you can see why the DoJ thinks this is a good and powerful remedy. It's not clear they understand the broader ramifications of such an ask or even the feasibility of it, but still, you can sort of see what they're thinking. They wanted to take a big swing without demanding all of Google be broken up. This is a big swing.
The second is that unlike the default search placement and deal they're (rightfully, I think) worried about within browsers, using Chrome itself is actually a choice almost all users on desktops/laptops make. That's because beyond ChromeOS devices, most such devices are Windows or Mac machines. And these devices generally do not have Chrome installed out of the box.1 You have to boot up Microsoft Edge or Apple Safari and yes, search for Chrome in order to install it. This is sort of a pain. And sure, Google uses prompts on Google Search at times to try to ease such actions, but this is something many users explicitly do.
Mobile, of course, is different since Chrome is the default browser on many, if not most,2 Android devices around the world (but no iPhones, of course). And you could certainly argue that using Chrome on mobile creates a halo effect which causes people to use Chrome on desktop/laptop too. But the reality is that Chrome had won the browser wars long before this was the case. Chrome won the browser wars because it was a better browser, full stop. We can (and will below) argue about whether that remains the case, but that's largely because other browsers have copied many of the features and techniques Google introduced with Chrome – and/or are using Chromium itself, the underlying open source technology, to power their browsers. Including now... Microsoft!
Speaking of, while I'm now of the mind that there are better overall browsers than Chrome, including the one I use, Arc, there are a few issues/current market realities for why new entrants can't displace Chrome in the same way that Chrome displaced Internet Explorer (and Firefox). The market is simply far more mature now and inertia is a hell of a drug. People use Chrome because everyone uses Chrome – which really means everyone knows how to use Chrome. And websites are optimized to work with Chrome. Arc (and some others) get around that latter point by building on top of the aforementioned Chromium project. But the way they try to differentiate is through new features and UI and while I'm now a fan of what they do and offer, it took some learning and getting used to. Most people aren't willing to do that in 2024.
Perhaps if there was some fundamental shift in performance, but all of us have computers now that are beyond capable of running any browser without noting any real performance differences in day-to-day usage. Battery life was perhaps the one vector of attack against Chrome (largely because of the way they handled background tabs, which was a dual-edge sword) but while even though Safari was clearly lapping Chrome here on Macs for a while – and despite the fact that Safari is the default browser on the Mac – it wasn't enough to overtake Chrome. And now Chrome has taken some steps to alleviate some of those battery challenges.
Here as well, of course, mobile is a different story. Safari's huge market share is not because of its default status on the Mac but because of its default status on the iPhone. This is even more nuanced, but essentially Chrome on iOS is not really Chrome because it's forced to use Apple's WebKit technologies versus their own. Users don't care about this, but it obviously isn't an ideal situation for Google. Regardless, defaults matter more on mobile – hence, Google paying over $20B a year to Apple to set Google Search as the default on Safari!
More broadly, the reality remains that what is going to break the dominance of Google Search is not going to be another search engine, it will be something tangential out of left field. As such, asking Google to sell/spin-off Chrome is seemingly irrelevant to the DoJ's goal here. Perhaps if they forced a sale to a competitor that actually had the resources and will to support Chrome, but that would mean selling it to another tech giant, like Microsoft, or Apple, or Meta and there's no way they want that – all of those companies are also under investigation for other forms of antitrust, market dominance, etc. Also, as the testimony from Apple and others has made clear in various trials, the main search competitor right now, Bing, simply isn't good enough to replace Google even if they were offered more money to do so. Hell, even Microsoft admitted this! Perhaps after years of default placement and thus, usage, Bing would surpass Google, but the reality here remains that most people would probably just go back to using Google.
Far more interesting here is Microsoft's partner, OpenAI. And that's because ChatGPT seems like the most likely product capable of impacting Google Search. While the data suggests this hasn't happened yet, it's early. My own anecdotal data suggests that there is something to this notion, especially as ChatGPT grows more capable including with, yes, web search capabilities (currently powered by Bing, naturally, but that will undoubtedly shift over time).
I know many people that now at least say they use ChatGPT more often than Google. ChatGPT's new web search and Google's new 'AI Overview' results aside, they mostly still serve different use cases. If you're looking for a quick and simple answer for something, ChatGPT has been the way to go. If you're looking for a range of options across the whole web, it's Google. But it used to be that Google was the main option for all of this, ChatGPT is eating into some element of what search has been used for because it's better for those use cases. And again, ChatGPT keeps gaining more and more features. Chatting about something with a bot. Using your voice. Generating an image. Coding. Etc.
This was my answer when asked by The Washington Post yesterday what could actually make a difference here. It's not just the end of the default search payments; that likely has to be coupled with a product/service that's actually good enough to replace Google Search from Apple's perspective:
But companies such as Apple could still choose to use Google without paying for it. “It would only really change things if someone else is installed as the default,” Siegler said. The “wild cards” could be AI companies such as OpenAI improving their own search chatbots quickly enough to be able to replace Google on iPhones and other platforms, Siegler said.
That would be a hell of a second phase of an Apple/OpenAI partnership... But it would also require a lot more trust than the current deal for Siri integration. And, well, that is seemingly going in the opposite direction...
To beat the dead horse, what displaces Google Search is not going to be a search engine, but it's instead likely to be something that makes you use a search engine less. My own usage of ChatGPT suggests this is slowly happening to me. And if this spreads to other, more mainstream users... Here, OpenAI's main problem may be the many other rivals in AI including Claude, Perplexity, and others that fragment these new use cases. And, of course, Google, Meta, Amazon, even Apple (and yes, OpenAI's good friend Microsoft!) are working on their own products with basically unlimited resources in a world where resources matter.
Something in this current jumble of chaos feels like it will be what eventually displaces Google Search. But none of those have anything to do with Chrome other than they mainly run inside of it, just as other web apps do. A move towards "agents" could change this dynamic, with some early entrants likely focused on using the browser as the hub from which such AI operates – that likely includes both OpenAI and...
All that said, if Google were to leverage Chrome to give Gemini some sort of unfair advantage, that may actually matter. I'm honestly not sure the court is savvy enough to play this out and it's mainly theoretical for now – and you shouldn't litigate theoretical future problems as they may or may not happen, obviously. Still... it's worth considering in this whole debate.
And with that in mind, I do think it makes sense for OpenAI to have a browser. They shouldn't buy Chrome, nor should they be allowed to for all the reasons discussed. But they should do what Google did back in the day and build it from the ground up to tailor it for their own vision of the future of the internet – complete with a new "Omnibox" that is truly "omni". And so the reports (and hires) suggesting that they're working on such a browser makes sense, of course.
Now, you might argue that their resources are better spent working on what's next after a web browser. And yes, they should be doing that too – and seemingly are. But an OpenAI web browser product right now reminds me not of Chrome, but of Google's old toolbar for Internet Explorer back in the day. One of the early projects which, much like Chrome itself, was led by none other than... current Google CEO Sundar Pichai. This was their way to get a toehold in the market at the time. And it worked.
You might say that the ChatGPT Chrome extension – pretty aggressively installed and implemented alongside the new Search product – is OpenAI's version of this. And it may be. But I think a re-imagined browser tailored around ChatGPT would be better placed to gain a toehold within our current tech environment. Until we get to the AI wearable product or whatever is actually next.
Go back to that first point: a web browser remains the most-used and thus, most important piece of software for nearly everyone on the planet.
1 I believe some OEMs install Chrome for out-of-the-box usage, maybe due to deals with Google, maybe not, though I'm not sure it's set to the default browser at first...
2 Certainly in most Western markets, but not China...
2024-11-22 01:51:57
I'll keep this brief as it's freezing cold (literally) in London at the moment and my basement office heat leaves a bit to be desired... More when I thaw...
🦚 Comcast Will Issue its Cable Spinoff a Hunting License – While most spin-offs become acquisition targets, it does seem like Comcast is setting up "SpinCo" to be a buyer in a market that is seemingly about to get quite acquisitive. Some of Paramount's properties feel obvious as that company works through their merger. Warner Bros Discovery can always use more cash, but they may also want to be a buyer to continue to bulk up Max. Everyone clearly wants to partner – including Comcast itself with the left-behind Peacock – which would be cheaper, so the market dynamics will be interesting. Remember too that beyond the cable channels, this new company has tangential properties like Fandango and Rotten Tomatoes, so they could buy more of those types of things. I'm still confused by the cleaving of MSNBC and CNBC from NBC itself – "Microsoft NBC" now is no longer Microsoft nor NBC – and I'm seemingly not alone there. But some think it will be good for those properties as they'll get more resources. [Semafor]
💰 NVIDIA Doubles Profit as A.I. Chip Sales Soar – Good news everyone, Jensen emerged and saw his shadow; AI's season of exuberance will continue for at least another quarter. While the stock is down a bit in pre-market trading, that mainly feels like a law of large numbers problem: NVIDIA thinks they'll grow revenue a mere 70% this quarter, versus the 94% last quarter. These remain staggering numbers for a company at this scale. And profit is even more impressive – that grew 106% y/y and the company passed both Amazon and Meta in that regard on a quarterly basis. The company, once again the most valuable in the world, makes up nearly double the percentage of the S&P 500 than what Cisco did at their in their Dot Com Bubble peak 2000 – but profit and earnings are also far stronger than Cisco saw back then. Meanwhile, any sort of AI training "slow down" (much contested in recent days) doesn't worry Jensen as he believes they're also perfectly primed for a shift to inference (versus the chip upstarts more focused on that element of AI compute). He, of course, would say that. But he's given people – and certainly Wall Street – no reason not to believe him yet. [NYT]
⚖️ DoJ Asks Court to Force Google to Spin Off Chrome – With the official filing now out there, there's not too much to add beyond the initial reports, but there are a few more details and nuance. Notably, the DoJ did consider also asking for the spin off of Android, but thought that might be a bridge too far here, for now – they're keeping the option on the table for a later date were these remedies not to work (or not to be implemented). There are asks to decouple Android and Search and Play Store. Google, naturally, is not happy with any of these "staggering" and "radical" proposals. In particular, the data sharing/licensing concepts have them bringing up the security and privacy angle – how very Apple of them! And they really can't be happy about the elements that would "chill" their AI work. Interestingly, though as expected, they also signaled being open to talking about the default search deals. Their proposal will be due on December 20. [CNN]
📺 Amazon’s new Echo Show 21 – I can't decide which is more likely: 1) If Amazon was sitting on these waiting to debut them alongside 'Remarkable Alexa', but since she's been delayed now multiple times, they still wanted to get them out for the holidays. 2) Amazon knows Apple's similar-sounding home hub thing is coming early next year and again, aimed to get these out now to get a nice install base with the holidays when Apple's device hits. In the immortal words of the meme: why not both? I will say, a 21" screen seems pretty large, at what point do we just call this a TV? Perhaps another question for Apple too at some point... [Verge]
⚾️ Around the Horn ending in 2025 after 23-year run on ESPN – This hasn't been confirmed by the network yet, but The Athletic also corroborated it so it seems like a done deal. Which sucks. I've been watching the show since it started 22 years ago. I started watching as a college kid and continued over 5 cities in 3 states and 2 countries. Now I'm a father of two in London, still watching it. While I also enjoy PTI, there's something about ATH that I just find so comforting and enjoyable and entertaining. In fact, my first ever tweet was about the show. And, coincidentally, last night I sent out a skeet because the show itself wrapped with Mina Kimes using her 'Face Time' to talk about Bluesky. [New York Post]
Just a few thoughts today on writing/blogging platforms and owning your home on the internet...
M.G. Siegler
"The titles, numbers, they’re there. People probably know that. The way I’d like to be remembered more is like a good person from a small village in Mallorca."
-- Rafael Nadal, speaking after his final match as a professional tennis player. One of the three best ever. Best ever on clay, period.
I feel lucky to have seen the primes of Nadal, Novak Djokovic, and Roger Federer as they battled each other to win nearly every single Grand Slam final every year, for basically 20 years.
A daily newsletter by M.G. Siegler
Sent from afar on weekdays...
2024-11-21 22:58:41
Next week – in 9 days, to be exact – I will mark 20 years writing on the web. I know this because of a calendar entry I made when I hit 10 years that reminds me. Sadly, not because of said post from 20 years ago, which is no longer online. A lot of what I have written over the two decades – thousands of posts – remains online, but a lot also doesn't. Publications come and go. And platforms come and go. Only the web remains. And while I wasn't savvy enough to recognize that in 2004 when I was just a bored young person living by myself in a new city, I do now.
That's why I went with Ghost when it came time to start this publication last year. I had been writing a newsletter on Substack which had a decent following. And I had a publication on Medium with a far larger one. But in trying to set this up for long term success I decided a (nice and modern) set of tools that were open source and provided a lot of flexibility would be best. Yes, it lacks some of the network effects of a Substack,1 but again, I've been doing this a long time. These days, I favor longevity and control.2
Anil Dash pulls no punches. "Don't call it a Substack" is a rant both against associating your work with a corporate brand and against this particular startup:
We constrain our imaginations when we subordinate our creations to names owned by fascist tycoons. Imagine the author of a book telling people to "read my Amazon". A great director trying to promote their film by saying "click on my Max". That's how much they've pickled your brain when you refer to your own work and your own voice within the context of their walled garden. There is no such thing as "my Substack", there is only your writing, and a forever fight against the world of pure enshittification.
Long before Substack, I actually started playing around with newsletters on a platform called Revue. Sadly, those newsletters no longer exist because Revue no longer exists. Twitter bought it and killed it. But wait, those posts actually do still exist! In my inbox. Email, the true cockroach of the internet, will never die.
That's the good news for those publishing on Substack. And yes, that platform has elements of it which allows writers to take their readers (as I did) and even payments (via Stripe) with them. That's good. But it doesn't negate the fact that Substack has more recently been practicing in some of the dark arts in order to keep writers (and their subscribers) locked-in to the network. Things like features conflating the notion of "subscribers" and "followers". The whole weirdly insulated Twitter-like network "Notes" thing. Etc.
Regardless, the reality remains that while Substack has some heat at the moment, all such fire inevitably fizzles for one reason or another. And again, all that's left is the web. And email. Anyone who has been around long enough has seen it time and time and time again. And it has led me to look enviously at folks like John Gruber – who has his own take on this matter today, more objecting to the way all sites using Substack are presented, which I definitely agree with – and Dave Winer and Jason Kottke, who have planted one flag in the web and stayed put.3 Not jumped from new publishing platform to the next to the next.
Back to Dash:
Links are powerful — that's why Instagram and Twitter and Threads punish and limit them, and why Substack tries to take credit for them. And that's why "wherever you get your podcasts" is such a radical concept — like email, it's a medium that the tech tycoons don't, and can't, own. People can read your writing "wherever they get their email".
Yes, though links only retain their power if the content on the other end of them doesn't vanish. Email is a nice insurance policy. But you should aim to house your writing in a home on the web that won't vanish.
1 We'll see if what they're attempting to do with "Fediverse" integration changes this at all.
2 And to be fair to myself, and Medium, and even Substack, I thought that by using my own domain, which I insisted on and which all eventually supported, I was protecting myself and my content just in case I ever had to move it...
3 Dash himself has been at it since 1999, though I believe moved his site a few times, but took his content with him!
