Apple's first day of their annual World Wide Developer Conference (WWDC) was today where they showed off their latest OS updates. As the rumors had discussed for a while, this year is mainly a...
Apple's first day of their annual World Wide Developer Conference (WWDC) was today where they showed off their latest OS updates. As the rumors had discussed for a while, this year is mainly a stability/catch-up year for Apple. Their v26 OSes were all considered to be a big downgrade compared to previous versions as they introduced the Liquid Glass aesthetic, which hurt performance, optimization, legibility and usability. The main upgrades this year are improved Apple Intelligence and Siri features (detailed below), child safety features, and a focus on details. They specifically call out refining Liquid Glass to be more readable and customizable (with a slider from ultraclear to tinted) and faster performance across the board in the OS (from app launches, to AirDrop transfers, and more).
The big update that Apple focused on was Siri AI, their newest version of their Siri personal assistant. In partnership with Google, Apple built their Apple Foundation models on Gemini's technology. This doesn't mean that Apple are simply using Google's models but are building their own models while learning from what Google did for theirs (I could be wrong on this but this is how I understood it). This new Siri AI brings a lot of "personal intelligence" features that we've seen Google Gemini, Google Lens, ChatGPT, Claude, OpenClaw, Poke, etc. all brought over the last 3 years or so. The new Siri AI can go through your photos, messages, emails, and more to let you interact with your content in a more natural way, much like what Google has been doing with Gemini. One thing that seems to set Siri AI apart though is that Apple placed very heavy emphasis on local/on-device AI being used, as opposed to primarily using cloud models.
Apple Intelligence also bring a couple of other useful features around the OS. A big one was Spatial Reframing for photos, where AI can let you extend your photo and move the camera around to better frame your shot. Safari can group tabs for you by topic and can notify you when something changes on a page (like a price or restock). The Passwords app can also notify you when a password has been detected in a breach and also automatically change your password in that app. This feature I'm a little weary of but does sound like the type of "magic" that Apple would build. There's also Call Context, which will surface relevant info for you from messages or email while you're in a call, similar to Magic Cue on the Google Pixel phones.
Overall, a seemingly minor update feature-wise compared to other years but its good to see Apple return to focusing on minor details and performance enhancements. It was quite shocking installing the first iOS 26 beta last year and seeing my phone crawl like it were a 2014 Android phone. I've installed the iOS 27 beta so far and it's such a delight to use. Things feel much smoother than even iOS 26.5 that I had installed previously. The updates to Liquid Glass are subtle but its definitely way more legible and honestly reminds me a lot more of the Aqua aesthetic that Mac had for ages. Excited to see how the OSes pan out through the betas.
I'm heartened to hear Liquid Glass is stabilizing. I've been stubbornly refusing to upgrade off iOS 18 for that reason alone. I'll probably hold out until 27 drops, maybe 27.0.1 or whatever...
I'm heartened to hear Liquid Glass is stabilizing. I've been stubbornly refusing to upgrade off iOS 18 for that reason alone. I'll probably hold out until 27 drops, maybe 27.0.1 or whatever because I'm still superstitious about about x.0.0 releases. But it's good to know I'm not going to be just hanging out here forever, lol.
I've been playing around with the iOS 27 developer beta and it is much improved. Legibility, even with the most clear option available, is excellent. With the most clear liquid glass option, for...
I've been playing around with the iOS 27 developer beta and it is much improved. Legibility, even with the most clear option available, is excellent. With the most clear liquid glass option, for components without text, the "glass" is much like real glass with colors refracting and bending as you'd expect. However, when there's any amount of text, the glass is mildly frosted. With the least clear option, the glass everywhere is frosted.
I'd still hold off, Apple spent the entirety of last year adjusting Liquid Glass constantly so I imagine adjustments are in the pipeline still.
My understanding was that they are using the Gemini models as is, and doing some additional reinforcement learning and runtime optimisation on top, not retraining from scratch. I do hope we get...
This doesn't mean that Apple are simply using Google's models but are building their own models while learning from what Google did for theirs (I could be wrong on this but this is how I understood it).
My understanding was that they are using the Gemini models as is, and doing some additional reinforcement learning and runtime optimisation on top, not retraining from scratch. I do hope we get some more details, but my guess is we won't.
Ah perhaps. I saw that Federighi put out a statement saying that these models are not simply what Google uses for its own Gemini product so I think I just interpreted that wrong.
Ah perhaps. I saw that Federighi put out a statement saying that these models are not simply what Google uses for its own Gemini product so I think I just interpreted that wrong.
I could see it going either way based on what Apple has said. I am more basing it on the fact that Apple has been trying for years to train their own models and has seemingly failed. I'm not sure...
I could see it going either way based on what Apple has said. I am more basing it on the fact that Apple has been trying for years to train their own models and has seemingly failed. I'm not sure they have anything to add to the Google special sauce on the training step that Google can't already do better. Given Apple's stance on privacy, their raw training data is almost certainly much worse and much smaller than Google.
What Apple has been good at is optimising for local models. They have a public "mobileclip" series of models. (Fun fact, mobileclip powered the search on audiobookcovers.com for a while; the model I use now is worse for search, but it has better infrastructure for running it cheaply) These are for image embeddings, and are somewhat unrelated to LLMs, but, as far as I am aware, they are the only CLIP style model that anyone even attempted to optimise for mobile devices. And even right now, Apple photos is the only photo service that does the AI features on device.
I can also see Apple being interested in post training to tune the Gemini models to how they want to act. This is also where small sample data from Apple employees could work well. It's not enough data for a full training run, but it is likely high quality enough for post training.
Either way, it's going to be interesting. There are some interesting quirks of Gemini models, and I wonder if they will make it into these models. For example, Gemini models are incredibly bad at tool calling. In my own testing, giving a Gemini model only a single tool, it still hallucinated and tried to call other tools around 60% of the time. That doesn't strike me as something apple would be okay with using, but can possibly be fixed with post training. At the same time, Gemini is also the only western lab doing audio input tokens. Anthropic and OpenAI use a speech to text model and input text into the model. We are not sure yet how the new Siri works, but it's possible it's sending audio directly to the model instead of through a transcription layer first. If so, that could be a very powerful advantage for Apple.
It’s nice that they’re focusing on performance and stability this year. That said, it’s ridiculous how much padding is there in macOS now. The shape around the buttons was already so large and now...
It’s nice that they’re focusing on performance and stability this year.
That said, it’s ridiculous how much padding is there in macOS now. The shape around the buttons was already so large and now there is another blown up shape for the tab bar! This looks like the biggest downside of the rumple rumoured upcoming touchscreen MacBook.
I don’t think I’ll ever trust agentic tools enough to let it attempt to log into websites and mess around with the password settings completely unmonitored.
They seem to have selected a good use case for it where it’s specifically going through long-neglected and compromised passwords. So these are likely to be accounts you don’t access often,...
I don’t think I’ll ever trust agentic tools enough to let it attempt to log into websites and mess around with the password settings completely unmonitored.
They seem to have selected a good use case for it where it’s specifically going through long-neglected and compromised passwords. So these are likely to be accounts you don’t access often, unlikely to have critical information in them, and they’re already compromised/insecure so it’s hard to do additional harm that hasn’t already been done through neglect. All that and “change password” flows tend to already be sanitized and put through validation checks so you have a very narrow window of actions it’ll actually be able to do.
From the speculation online, its seemingly just on life support unfortunately. They've shelved the lower-cost version, moved engineers all around the company, and are apparently instead working on...
From the speculation online, its seemingly just on life support unfortunately. They've shelved the lower-cost version, moved engineers all around the company, and are apparently instead working on a pair of smart glasses much like the Meta glasses. People are speculating the focus on Visual Intelligence during the keynote today sort of hints at capabilities in Apple's smart glasses. The features they've introduced into visionOS for just looking at something and then looking at Siri and asking what you're looking at is something they say you would be able to do on their glasses.
Out of curiosity, how often do you use it? From what I gather even people who have one typically aren't using it with regularity. I don't have a Vision Pro but I do have a VR headset that's just...
Out of curiosity, how often do you use it? From what I gather even people who have one typically aren't using it with regularity. I don't have a Vision Pro but I do have a VR headset that's just collecting dust. I feel conflicted about it because I love the idea of VR and honestly I always have a good time with it when I use it. But the extra work of setting it up, clearing space in my room for it, and the (fairly minor) physical exertion and eye strain involved present enough of a psychological barrier that I usually don't want to bother with hooking it up.
I suspect Apple's sales figures and usage metrics indicate a lack of interest, which deters further development. By all accounts using the Vision Pro is a very pleasant experience, so that's disappointing. I imagine people would be more inclined to use XR in daily life if it became frictionless... but that will never happen if R&D gives up while the tech is still nascent. I'd love to see Apple invest in a 10+ year project to move things in that direction, since the field as a whole seems kinda plateaued at the moment. After the whole car thing, I can see why they'd be reticent though. But they're sitting on a mountain of cash, seems like that would be a good use for a small fraction of it.
This update is mostly focused around Apple Intelligence, however the standout new feature to me is the EQ for AirPods. It’s something that most third parties already do for their earbuds through...
This update is mostly focused around Apple Intelligence, however the standout new feature to me is the EQ for AirPods. It’s something that most third parties already do for their earbuds through their own apps, and it was something AirPods needed in my opinion. I know not many care about equalising their earbuds, but for the ones who do it’s a great addition. It seems very basic from the small image they put on their website, but I guess we’ll have to see when it releases.
Yeah, I was excited by that announcement too. I don't need an intensively customisable EQ, and three bands should be absolutely fine for my needs. Looking forward to having a play with it.
Yeah, I was excited by that announcement too. I don't need an intensively customisable EQ, and three bands should be absolutely fine for my needs. Looking forward to having a play with it.
This all sounds really great. I have been wishing for Siri to become smart - particularly, conversationally smart - for a while now, and I would be oh-so-happy for it to be able to do so as...
This all sounds really great. I have been wishing for Siri to become smart - particularly, conversationally smart - for a while now, and I would be oh-so-happy for it to be able to do so as locally as possible.
I’ve also been really nervous about Rosetta 2 being discontinued and it looks like 27 will keep it around for the time being. Until I bother to update some stupid tools I still use I will be sticking to 27.
Yeah with macOS 27 discontinuing support for Intel-based Macs and Rosetta 2 remaining, I wonder how much longer Apple will dedicate resources into x86->Apple Silicon software. I'm pretty sure the...
Yeah with macOS 27 discontinuing support for Intel-based Macs and Rosetta 2 remaining, I wonder how much longer Apple will dedicate resources into x86->Apple Silicon software. I'm pretty sure the Game Porting Toolkit also remains so if Rosetta suddenly dies in macOS 28, I can imagine there being some workarounds for the GPTK to run x86 applications (unless I'm misunderstanding how it works).
I don't use a huge amount of Apple Intelligence, so this year's updates seem potentially a little on the underwhelming side. I like the sound of describing a shortcut, as well as using on-screen...
I don't use a huge amount of Apple Intelligence, so this year's updates seem potentially a little on the underwhelming side. I like the sound of describing a shortcut, as well as using on-screen context. I have apps that don't hook into shortcuts, and it would be really cool to automate them with Siri AI. Perhaps I'm misunderstanding what's going to be available to us though.
Otherwise, the big standout for me is the overhaul of the screentime/parental controls. My wife and I have both lamented how poor and overly complex Apple's implementation of this stuff is when it comes to our kids, and this redesign looks like a really promising step in the right direction.
I was hoping for an announcement for the m5 versions for Mac mini and mac studio, but it looks like they ain’t happening. I want to build my own server for LLM processing, but I’ll have to wait...
I was hoping for an announcement for the m5 versions for Mac mini and mac studio, but it looks like they ain’t happening.
I want to build my own server for LLM processing, but I’ll have to wait longer it seems
I think Mac stuff wouldn't happen until much later in the year, perhaps around August or October. WWDC is purely software focused. I did read somewhere recently that there has been a big sell-off...
I think Mac stuff wouldn't happen until much later in the year, perhaps around August or October. WWDC is purely software focused. I did read somewhere recently that there has been a big sell-off of Mac Mini's from people who bought into the OpenClaw hype a few months ago to setup their own local LLM with a M4 Mac Mini but never ended up doing so, you might be able to snag one for cheap but I haven't checked myself.
There was absolutely zero chance of Apple rolling back Liquid Glass. Like it or not, if you want Apple devices, you are stuck with it until the next big redesign. But in the 27 OSs, you can set...
There was absolutely zero chance of Apple rolling back Liquid Glass. Like it or not, if you want Apple devices, you are stuck with it until the next big redesign.
But in the 27 OSs, you can set the transparency slider to fully tinted, which improves a lot of the usability problems with Liquid Glass.
Nope but it’s tweaked a bit. I elaborated in my other comments but any text/icon elements now have a frosted glass appearance behind them, the liquid aspect remains unchanged.
Nope but it’s tweaked a bit. I elaborated in my other comments but any text/icon elements now have a frosted glass appearance behind them, the liquid aspect remains unchanged.
Disclaimer: guess I'm being rather critical of things today- so a forewarning. I am glad to see they're refining liquid glass a bit more because I hated (and still hate) it but there was zero else...
Disclaimer: guess I'm being rather critical of things today- so a forewarning.
I am glad to see they're refining liquid glass a bit more because I hated (and still hate) it but there was zero else in the announcement for me. I decided to abandon my Mac rather than upgrade to macOS 26. Seems 27 brings small refinements but still is maddeningly fisher-pricey so I'm glad to have completely abandoned the desktop Mac ecosystem (for the second time in my life)
Most of the announcements are all a bunch of AI garbage I could not be less interested in. FWIW, I am glad I'm still on a model of iPhone that can't run most of that stuff anyway. I do have an iPad Pro but I might keep it on iOS 18 as long as I can hold out, then begrudgingly update when some app I need requires it
Though as garbage as Siri has been as a thing, what's sad is this might actually improve it
Can't blame you for being critical, the Liquid Glass update is very polarizing. The refinements in the 27 do rollback some of the silly updates like the floating sidebar. I overall do enjoy the...
Can't blame you for being critical, the Liquid Glass update is very polarizing. The refinements in the 27 do rollback some of the silly updates like the floating sidebar. I overall do enjoy the Liquid Glass update on my Mac. I don't see myself abandoning the Mac anytime soon though, unless the upcoming touch screen and dynamic island updates truly ruin it. I'm on a Mac laptop so the hardware is just unmatched (at least for now, we'll see how the new Nvidia/Microsoft ARM chip goes, apparently its just the same basic chip from their DGX Spark). My phone, watch, and Airpods all working in conjunction with each other is nice to have compared to the Android equivalents I used to have, where they were all separate, which was fine but the Apple niceties are hard to give up.
I just think for MacOS they fully committed to the "tons of whitespace" "make it look kinda like mobile" types of design language, in addition to all the ways it was unrefined and just awful in...
I just think for MacOS they fully committed to the "tons of whitespace" "make it look kinda like mobile" types of design language, in addition to all the ways it was unrefined and just awful in some places where it was clear they paid no attention to detail (things that were unreadable or transparent in the least usable way, etc). It just looked like an even less professional, "toy" operating system with even less information density than it used to have and just lots of weird transparency in places that didn't need it. So I decided it wasn't for me. That plus Apple's overbearing "we know better than you" stuff that power users have to work around, and some other restrictive things I ran into with my M4 mini just made me decide to give it up and go back to a machine that I can easily upgrade storage / components in and so on. I specifically used it as my music / creative production machine, and decided to go back to a de-bloated Win10 machine (actually specifically chose 10 because it's no longer getting updates). Linux just isn't there for music production yet (at least for the plugins/etc I've invested in) so maybe someday I'll go that direction because it would be my preference- I'd love to eventually get the last non-Linux machine in my home switched over to Linux, just not gonna happen to that one specific system any time soon. It's also plenty powerful. Most of the "power" of the M4 for me was going unused anyway. Now I have easily upgradeable storage, way more RAM, add-in cards if I need them including a more powerful discrete GPU if I want to use that machine for editing photos/videos/etc too, etc.
I definitely understand the nicety of being fully in the ecosystem where they own the entire vertical and everything works together well, so I don't blame people for liking that, I just never use almost any of those benefits anyway, so it wasn't a big deal for me to leave. I don't use any other Apple hardware than iOS/iPad OS now. Got rid of my AppleTV, don't want a watch or airpods, etc.
I definitely also understand having high quality hardware (despite my MANY qualms with their hardware in other ways) in a desktop or laptop form is nice for sure. Great touchpad, build quality, etc. The tradeoffs have just tipped the scales for me. Apple is not a company that makes things for me anymore.
I don't even know if my next phone will be an iPhone. Probably because of inertia/investment yes, but I'm going to use my current devices until I just can't anymore.
I still prefer iOS / iPadOS to their Android equivalents but begrudgingly so. I do hate Liquid Glass. Aside from the abject ugliness, and me despising the aesthetic choices in general, on my iPhone, even multiple updates in, it's still less stable than iOS 18. Still, apps occasionally freeze that never did before, or the home screen / app icons occasionally uncache themselves and pop back in, or sometimes the animations on the phone are really laggy and slow. None of this happened in 18, and they began immediately after upgrade (and still happen), so it was definitely OS-based, not a symptom of my hardware being too old. This is why I've avoided it on the iPad Pro. I do not want to experience there what I've experienced on my iPhone.
Apple's first day of their annual World Wide Developer Conference (WWDC) was today where they showed off their latest OS updates. As the rumors had discussed for a while, this year is mainly a stability/catch-up year for Apple. Their v26 OSes were all considered to be a big downgrade compared to previous versions as they introduced the Liquid Glass aesthetic, which hurt performance, optimization, legibility and usability. The main upgrades this year are improved Apple Intelligence and Siri features (detailed below), child safety features, and a focus on details. They specifically call out refining Liquid Glass to be more readable and customizable (with a slider from ultraclear to tinted) and faster performance across the board in the OS (from app launches, to AirDrop transfers, and more).
The big update that Apple focused on was Siri AI, their newest version of their Siri personal assistant. In partnership with Google, Apple built their Apple Foundation models on Gemini's technology. This doesn't mean that Apple are simply using Google's models but are building their own models while learning from what Google did for theirs (I could be wrong on this but this is how I understood it). This new Siri AI brings a lot of "personal intelligence" features that we've seen Google Gemini, Google Lens, ChatGPT, Claude, OpenClaw, Poke, etc. all brought over the last 3 years or so. The new Siri AI can go through your photos, messages, emails, and more to let you interact with your content in a more natural way, much like what Google has been doing with Gemini. One thing that seems to set Siri AI apart though is that Apple placed very heavy emphasis on local/on-device AI being used, as opposed to primarily using cloud models.
Apple Intelligence also bring a couple of other useful features around the OS. A big one was Spatial Reframing for photos, where AI can let you extend your photo and move the camera around to better frame your shot. Safari can group tabs for you by topic and can notify you when something changes on a page (like a price or restock). The Passwords app can also notify you when a password has been detected in a breach and also automatically change your password in that app. This feature I'm a little weary of but does sound like the type of "magic" that Apple would build. There's also Call Context, which will surface relevant info for you from messages or email while you're in a call, similar to Magic Cue on the Google Pixel phones.
Overall, a seemingly minor update feature-wise compared to other years but its good to see Apple return to focusing on minor details and performance enhancements. It was quite shocking installing the first iOS 26 beta last year and seeing my phone crawl like it were a 2014 Android phone. I've installed the iOS 27 beta so far and it's such a delight to use. Things feel much smoother than even iOS 26.5 that I had installed previously. The updates to Liquid Glass are subtle but its definitely way more legible and honestly reminds me a lot more of the Aqua aesthetic that Mac had for ages. Excited to see how the OSes pan out through the betas.
I'm heartened to hear Liquid Glass is stabilizing. I've been stubbornly refusing to upgrade off iOS 18 for that reason alone. I'll probably hold out until 27 drops, maybe 27.0.1 or whatever because I'm still superstitious about about x.0.0 releases. But it's good to know I'm not going to be just hanging out here forever, lol.
I've been playing around with the iOS 27 developer beta and it is much improved. Legibility, even with the most clear option available, is excellent. With the most clear liquid glass option, for components without text, the "glass" is much like real glass with colors refracting and bending as you'd expect. However, when there's any amount of text, the glass is mildly frosted. With the least clear option, the glass everywhere is frosted.
I'd still hold off, Apple spent the entirety of last year adjusting Liquid Glass constantly so I imagine adjustments are in the pipeline still.
My understanding was that they are using the Gemini models as is, and doing some additional reinforcement learning and runtime optimisation on top, not retraining from scratch. I do hope we get some more details, but my guess is we won't.
Ah perhaps. I saw that Federighi put out a statement saying that these models are not simply what Google uses for its own Gemini product so I think I just interpreted that wrong.
I could see it going either way based on what Apple has said. I am more basing it on the fact that Apple has been trying for years to train their own models and has seemingly failed. I'm not sure they have anything to add to the Google special sauce on the training step that Google can't already do better. Given Apple's stance on privacy, their raw training data is almost certainly much worse and much smaller than Google.
What Apple has been good at is optimising for local models. They have a public "mobileclip" series of models. (Fun fact, mobileclip powered the search on audiobookcovers.com for a while; the model I use now is worse for search, but it has better infrastructure for running it cheaply) These are for image embeddings, and are somewhat unrelated to LLMs, but, as far as I am aware, they are the only CLIP style model that anyone even attempted to optimise for mobile devices. And even right now, Apple photos is the only photo service that does the AI features on device.
I can also see Apple being interested in post training to tune the Gemini models to how they want to act. This is also where small sample data from Apple employees could work well. It's not enough data for a full training run, but it is likely high quality enough for post training.
Either way, it's going to be interesting. There are some interesting quirks of Gemini models, and I wonder if they will make it into these models. For example, Gemini models are incredibly bad at tool calling. In my own testing, giving a Gemini model only a single tool, it still hallucinated and tried to call other tools around 60% of the time. That doesn't strike me as something apple would be okay with using, but can possibly be fixed with post training. At the same time, Gemini is also the only western lab doing audio input tokens. Anthropic and OpenAI use a speech to text model and input text into the model. We are not sure yet how the new Siri works, but it's possible it's sending audio directly to the model instead of through a transcription layer first. If so, that could be a very powerful advantage for Apple.
It’s nice that they’re focusing on performance and stability this year.
That said, it’s ridiculous how much padding is there in macOS now. The shape around the buttons was already so large and now there is another blown up shape for the tab bar! This looks like the biggest downside of the rumple rumoured upcoming touchscreen MacBook.
I don’t think I’ll ever trust agentic tools enough to let it attempt to log into websites and mess around with the password settings completely unmonitored.
They seem to have selected a good use case for it where it’s specifically going through long-neglected and compromised passwords. So these are likely to be accounts you don’t access often, unlikely to have critical information in them, and they’re already compromised/insecure so it’s hard to do additional harm that hasn’t already been done through neglect. All that and “change password” flows tend to already be sanitized and put through validation checks so you have a very narrow window of actions it’ll actually be able to do.
As someone who owns a Vision Pro, I wasn’t expecting much but I was still shocked at how little attention it got. Not a great sign of things to come.
From the speculation online, its seemingly just on life support unfortunately. They've shelved the lower-cost version, moved engineers all around the company, and are apparently instead working on a pair of smart glasses much like the Meta glasses. People are speculating the focus on Visual Intelligence during the keynote today sort of hints at capabilities in Apple's smart glasses. The features they've introduced into visionOS for just looking at something and then looking at Siri and asking what you're looking at is something they say you would be able to do on their glasses.
Out of curiosity, how often do you use it? From what I gather even people who have one typically aren't using it with regularity. I don't have a Vision Pro but I do have a VR headset that's just collecting dust. I feel conflicted about it because I love the idea of VR and honestly I always have a good time with it when I use it. But the extra work of setting it up, clearing space in my room for it, and the (fairly minor) physical exertion and eye strain involved present enough of a psychological barrier that I usually don't want to bother with hooking it up.
I suspect Apple's sales figures and usage metrics indicate a lack of interest, which deters further development. By all accounts using the Vision Pro is a very pleasant experience, so that's disappointing. I imagine people would be more inclined to use XR in daily life if it became frictionless... but that will never happen if R&D gives up while the tech is still nascent. I'd love to see Apple invest in a 10+ year project to move things in that direction, since the field as a whole seems kinda plateaued at the moment. After the whole car thing, I can see why they'd be reticent though. But they're sitting on a mountain of cash, seems like that would be a good use for a small fraction of it.
This update is mostly focused around Apple Intelligence, however the standout new feature to me is the EQ for AirPods. It’s something that most third parties already do for their earbuds through their own apps, and it was something AirPods needed in my opinion. I know not many care about equalising their earbuds, but for the ones who do it’s a great addition. It seems very basic from the small image they put on their website, but I guess we’ll have to see when it releases.
Yeah, I was excited by that announcement too. I don't need an intensively customisable EQ, and three bands should be absolutely fine for my needs. Looking forward to having a play with it.
This all sounds really great. I have been wishing for Siri to become smart - particularly, conversationally smart - for a while now, and I would be oh-so-happy for it to be able to do so as locally as possible.
I’ve also been really nervous about Rosetta 2 being discontinued and it looks like 27 will keep it around for the time being. Until I bother to update some stupid tools I still use I will be sticking to 27.
Yeah with macOS 27 discontinuing support for Intel-based Macs and Rosetta 2 remaining, I wonder how much longer Apple will dedicate resources into x86->Apple Silicon software. I'm pretty sure the Game Porting Toolkit also remains so if Rosetta suddenly dies in macOS 28, I can imagine there being some workarounds for the GPTK to run x86 applications (unless I'm misunderstanding how it works).
I don't use a huge amount of Apple Intelligence, so this year's updates seem potentially a little on the underwhelming side. I like the sound of describing a shortcut, as well as using on-screen context. I have apps that don't hook into shortcuts, and it would be really cool to automate them with Siri AI. Perhaps I'm misunderstanding what's going to be available to us though.
Otherwise, the big standout for me is the overhaul of the screentime/parental controls. My wife and I have both lamented how poor and overly complex Apple's implementation of this stuff is when it comes to our kids, and this redesign looks like a really promising step in the right direction.
I was hoping for an announcement for the m5 versions for Mac mini and mac studio, but it looks like they ain’t happening.
I want to build my own server for LLM processing, but I’ll have to wait longer it seems
I think Mac stuff wouldn't happen until much later in the year, perhaps around August or October. WWDC is purely software focused. I did read somewhere recently that there has been a big sell-off of Mac Mini's from people who bought into the OpenClaw hype a few months ago to setup their own local LLM with a M4 Mac Mini but never ended up doing so, you might be able to snag one for cheap but I haven't checked myself.
Is Liquid Glass fully rolled back now? Everyone I know who's updated hates it
There was absolutely zero chance of Apple rolling back Liquid Glass. Like it or not, if you want Apple devices, you are stuck with it until the next big redesign.
But in the 27 OSs, you can set the transparency slider to fully tinted, which improves a lot of the usability problems with Liquid Glass.
Nope but it’s tweaked a bit. I elaborated in my other comments but any text/icon elements now have a frosted glass appearance behind them, the liquid aspect remains unchanged.
Disclaimer: guess I'm being rather critical of things today- so a forewarning.
I am glad to see they're refining liquid glass a bit more because I hated (and still hate) it but there was zero else in the announcement for me. I decided to abandon my Mac rather than upgrade to macOS 26. Seems 27 brings small refinements but still is maddeningly fisher-pricey so I'm glad to have completely abandoned the desktop Mac ecosystem (for the second time in my life)
Most of the announcements are all a bunch of AI garbage I could not be less interested in. FWIW, I am glad I'm still on a model of iPhone that can't run most of that stuff anyway. I do have an iPad Pro but I might keep it on iOS 18 as long as I can hold out, then begrudgingly update when some app I need requires it
Though as garbage as Siri has been as a thing, what's sad is this might actually improve it
Can't blame you for being critical, the Liquid Glass update is very polarizing. The refinements in the 27 do rollback some of the silly updates like the floating sidebar. I overall do enjoy the Liquid Glass update on my Mac. I don't see myself abandoning the Mac anytime soon though, unless the upcoming touch screen and dynamic island updates truly ruin it. I'm on a Mac laptop so the hardware is just unmatched (at least for now, we'll see how the new Nvidia/Microsoft ARM chip goes, apparently its just the same basic chip from their DGX Spark). My phone, watch, and Airpods all working in conjunction with each other is nice to have compared to the Android equivalents I used to have, where they were all separate, which was fine but the Apple niceties are hard to give up.
I just think for MacOS they fully committed to the "tons of whitespace" "make it look kinda like mobile" types of design language, in addition to all the ways it was unrefined and just awful in some places where it was clear they paid no attention to detail (things that were unreadable or transparent in the least usable way, etc). It just looked like an even less professional, "toy" operating system with even less information density than it used to have and just lots of weird transparency in places that didn't need it. So I decided it wasn't for me. That plus Apple's overbearing "we know better than you" stuff that power users have to work around, and some other restrictive things I ran into with my M4 mini just made me decide to give it up and go back to a machine that I can easily upgrade storage / components in and so on. I specifically used it as my music / creative production machine, and decided to go back to a de-bloated Win10 machine (actually specifically chose 10 because it's no longer getting updates). Linux just isn't there for music production yet (at least for the plugins/etc I've invested in) so maybe someday I'll go that direction because it would be my preference- I'd love to eventually get the last non-Linux machine in my home switched over to Linux, just not gonna happen to that one specific system any time soon. It's also plenty powerful. Most of the "power" of the M4 for me was going unused anyway. Now I have easily upgradeable storage, way more RAM, add-in cards if I need them including a more powerful discrete GPU if I want to use that machine for editing photos/videos/etc too, etc.
I definitely understand the nicety of being fully in the ecosystem where they own the entire vertical and everything works together well, so I don't blame people for liking that, I just never use almost any of those benefits anyway, so it wasn't a big deal for me to leave. I don't use any other Apple hardware than iOS/iPad OS now. Got rid of my AppleTV, don't want a watch or airpods, etc.
I definitely also understand having high quality hardware (despite my MANY qualms with their hardware in other ways) in a desktop or laptop form is nice for sure. Great touchpad, build quality, etc. The tradeoffs have just tipped the scales for me. Apple is not a company that makes things for me anymore.
I don't even know if my next phone will be an iPhone. Probably because of inertia/investment yes, but I'm going to use my current devices until I just can't anymore.
I still prefer iOS / iPadOS to their Android equivalents but begrudgingly so. I do hate Liquid Glass. Aside from the abject ugliness, and me despising the aesthetic choices in general, on my iPhone, even multiple updates in, it's still less stable than iOS 18. Still, apps occasionally freeze that never did before, or the home screen / app icons occasionally uncache themselves and pop back in, or sometimes the animations on the phone are really laggy and slow. None of this happened in 18, and they began immediately after upgrade (and still happen), so it was definitely OS-based, not a symptom of my hardware being too old. This is why I've avoided it on the iPad Pro. I do not want to experience there what I've experienced on my iPhone.