Siri was the first widely available AI bot but they forgot to open it to developers. The ecosystem would have grown a million fold in all these years<p>Never underestimate the power of a bored programmer<p><pre><code> Siri turn on the lights
Siri code a Doom clone in Swift
Siri design a landing page for a Consulting Agency
Siri summarize the latest news about AI
Siri find a solution to the Riemann hypothesis
</code></pre>
Now that's an agent I'd use. Can we start with Siri for the terminal?<p>$ Siri 'List the files in this folder'
Amazon did exactly this with Alexa skills and it hasn’t exactly been a booming success. Turns out that all of the things you listed as examples are stuff that most consumers don’t actually want or will use.<p>“AI as a chat bot” has always been cool and fun in sci-fi movies, but it’s the least imaginative real world use for it. We need to move past it.<p>> $ Siri 'List the files in this folder'<p>Why in the world would I ever use this instead of just ‘ls’?
> Turns out that all of the things you listed as examples are stuff that most consumers don’t actually want or will use.<p>I've never understood this obsession with trying to invent the talking ship computer from Star Trek TNG. The purpose of the talking computer in the series was not to be a "cool 24th century technology advancement". It was a plot device to keep the story going and explain things to the audience. Having a voice conversation with a computer is terrible, imprecise, plodding UX. It's not a good interface, it's not a good invention, it's not good technology. Yet $billions are being spent to try to invent it.
Amazon and I guess Siri's original design was completely undiscoverable. Both had limited vocabulary and actions you could take. I had to learn from others what I could even ask the devices. Also both were audio only and for things like, "turn on the lights" that required talking to cheap 3rd party devices that barely stay connected, it was frustrating.<p>Instead of quickly iterating to find how to take voice to the next level, both projects quickly locked in on whatever that was and became too costly to pivot.
Alexa's big issue for me was that you had to remember very specific commands and unless you remembered the exact syntax you were out of luck. I genuinely think if the AI had been better and had been able to interpret commands better, skills would have been infinitely more useful.
Granny: "cobolcomesback, what's that command I have to write in this black box thingee to list all the files in this folder"<p>cobolcomesback: "ls, granny"<p>Granny: "it said No such file or directory"<p>cobolcomesback: "no, that's... just write ls, granny.<p>Granny: "I wrote ls granny it said no such file or directory."<p>cobolcomesback: "just.. from now on... just write Siri 'list all files in this directory', granny"<p>Granny: "I wrote siri list all files in this directory granny and it told me all the files and then it also asked me who's granny and I told it and it said nice to meet you"<p>cobolcomesback: "SIGH"
The problem with Alexa is that by the time it appeared anywhere other than Echo devices, it was already deeply enshittified. Siri is in your pocket, on your wrist, and in your ears. Alexa was stuck on your kitchen counter asking you if you wanted to try Amazon Music.
> Siri is in your pocket, on your wrist, and in your ears<p>and more often than not isn't reliable unless you have a midwestern US accent.<p>Siri wasn't good. It still isn't good. Apple hasn't delivered on its promises with it and that's why it's being replaced with Gemini.
distribution. MacBooks are everywhere and already used by devs
<i>> Can we start with Siri for the terminal?</i><p>It already exists in macOS 27: fm(1)<p>It is an extremely powerful tool, but here are the basics:<p><pre><code> # Ask the default on-device model
fm respond 'Explain this shell error in plain English: permission denied'
# Start a conversation
fm chat --instructions 'Let's talk about economics.'
# Use the larger Private Cloud Compute model
fm respond --model pcc 'Are zebras and horses related?'
# Pipe stuff to a model
echo 'What is Swift?' | fm respond</code></pre>
I realized my piping stuff example wasn't good enough. Here's a better one:<p><pre><code> pbpaste | fm respond --instructions 'Summarize in five bullets. Preserve dates, owners, and decisions.'</code></pre>
fm = Foundation Models<p><a href="https://developer.apple.com/videos/play/wwdc2026/334/" rel="nofollow">https://developer.apple.com/videos/play/wwdc2026/334/</a>
I'd settle for a siri that can at least turn the lights off properly.<p>Google assistant on the nvidia shield left Siri in the dust years ago. Siri can't even chain basic commands in your smart home, it can't retain any kind of context (e.g. dim the lights...a bit more).
Just today running latest developers beta, I asked Siri to turn on the coffee machine and it started playing some song instead. I found out later that I had named the switch coffee machine plug.
Similar thing happens when I ask Siri to play my favorite songs and it starts to play a song named favorite from my library
I gave up on using Siri for my lights because 1/3 of the time it would get caught in a “who is speaking? Please go through the voice configuration to identify yourself.”<p>This was frustrating for me as it always prompted “who cares? Just turn off the lights.” As well as would drive my parents and other guests batty.
> This was frustrating for me as it always prompted “who cares? Just turn off the lights.”<p>I guess you should be able to create a bucket of low-risk tasks that require zero authentication, like letting your guests turn off the lights. I do get why they are cautious, though, because in some cases this could be a security risk, like a burglar turning off lights from outside the home.
I have used Siri for years and turning on and off lights has never been an issue.
The year is 2026 and "Siri mark my medication as taken" returns "I'm sorry I can't do that"
That’s the weird part. For me it used to work very well, then after some update it just stopped. Typical Apple.
Can we continue it to be somehow room-aware?
Siri is Gemini now (Semeni?) so expect similar stories from both major phone OS providers going forward, since we have lost more diversity in the mobile market (seems anti-trustworthy to me)
You haven’t used the beta
I tried "turn off the lights in [room a] and [room b]" in the new Siri app and rather than Siri saying they can't do compound commands they just turned off the lights in one of the two rooms listed.<p>So, progress?<p>Any agent with the Home Assistant MCP hooked up can do it fine.
Now playing "Turn Off the Lights" by Teddy Pendergrass on Apple Music.
The command might have changed.<p><a href="https://youtube.com/watch?v=W63oJtX0dP8" rel="nofollow">https://youtube.com/watch?v=W63oJtX0dP8</a>
Most people do not know tools to their problems already existed before AI.<p>Example: "I can use AI to combine multiple PDFs into a single PDF."
Lmao.<p>Okay let me just install a tool to do that because the OS doesn't usually have it built in in 2026.<p>Oh, it's riddled with ads. Oh it's paid for. Etc.<p>It's the same thing with the recent opening of knowledge thanks to AI.<p>Before: search for recipe with search riddled with ads, find website riddled with ads cookies notification requests forced sign up and a million other dark patterns, then an entire fucking life story before the recipe.<p>Now: "find me a good recipe for pasta that uses the stuff I have left over from the party last weekend, I don't mind buying a few more bits".
$ Siri 'rename this thousand of photos in these folders describing their content and make sure to use the same name for duplicates regardless of their size'<p>I am waiting for this function to sort my folders
The thing is, Apple will have never lost this opportunity. The distribution advantage of the user base is so deep they can basically always implement this at will.
Apfel filled that role: $ apfel "List the files in this folder."<p><a href="https://apfel.franzai.com/" rel="nofollow">https://apfel.franzai.com/</a>
What about: just not Siri at all. What about that. What about MrSunshine. I would rather have my own bot that I code. But no, vendor lock in.
It’d be short sighted to not have a wrapper around which AI to use, even without implementation for them. Like, having an abstract parent class with a Siri implementation of it would be baseline good software engineering.
Yeah; plus we know they're designing with potential future local-only Vs. hybrid Vs. cloud-only in mind depending on hardware/market changes. So having it be this flexible buys them a ton of future potential.<p>People have speculated that with FPGAs or even an ASIC you could take some smaller models we have today, turn the chip into essentially that model-as-hardware, and get some really compelling performance/power-usage characteristics.<p>At that point running a GOOD quantized Q8 model on a phone is not beyond the realm.
I’m not even sure how they would implement these features <i>without</i> them being interoperable. All the AI systems use pretty similar if not identical API standards.
That exists already for Apple devices. They just never give third party models first class access to user and device data.<p>That arrangement will change, somewhat, later today.
And given how Apple switched out OpenAI for Gemini (to power Siri), they experienced firsthand the importance of having said wrapper.
And you can also sideload apps (only in the EU), change you default navigation app from apple maps (only in the EU), and use more features of the NFC reader (only in the EU).<p>The code for all of those things is there, but not enabled if for most iOS users. There's a non-zero chance this code is there for Apple to try and convince the EU to let them release a Siri AI, and then just like sideloading have some impossible-to-actually-use alternative path to make the DMA happy.
Apple can hold the AI customers hostage by locking this level of AI integration to their app store, taking a 30% cut of AI subscriptions while at the same time having someone else fix the reputation of their devices' voice assistant for them.<p>There's profit to be made by offering customers choices here, I'm sure Apple will be ready to scoop up all of those delicious slop dollars.
I pray for the possibility of not only choosing the provider, but being able to choose the thinking level.<p>I've been using the new siri, and it seems more or less stuck in instant-answer mode, which gives really terrible answers with any provider.<p>I wish more people were talking about this, but if you can't crank up the models to extended thinking, the technology is more or less stuck at 2004 levels of usefulness.
This is a very clever platform play. If you make Siri / your mac the central access point for all AI related workflows and make models interchangeable, then Apple has a huge weapon in the AI race.
I don't disagree, but I do feel the need to point out this was not Apple's plan A. Luckily for them they have so many tools and options at their disposal, that they are able to massively fail at developing their own models and still end up in a good position.
Apple “genius” has been to accept their plan A failed (the original Apple Intelligence), and instead of doubling down, let the others explore the space, then offer a good interface (I say that, I haven’t seen if it’s actually good).<p>So many other actors in the industry would have just doubled down and follow the race to the bottom
Genius would have been for Apple to get Siri in a better state after so many years.<p>This looks more like a participation medal, if at all.
I imagine that Google assistant only makes money for Google because they’re an advertising company. Alexa loses money for Amazon and it’s becoming more gated to Prime subscription features as a result. Amazon thought everyone would be buying stuff with their voice and that future never panned out.<p>It turns out that voice assistants aren’t all that important. Customers aren’t willing to pay for them and they’re more of a nice to have convenience feature.
They didn’t invest in Siri and became the most valuable company in history. I don’t think Siri being better would have mattered much to be honest. Not joining the financial black hole that is AI training seems to have benefited them well so far. But yes, it’s definitely a case of failing upwards.<p>Adding quotes around “genius” as it seems you misunderstood what I meant
Meh, AI is deep in it's diversification stage -- there are soooooooo many players in the game right now and basically one maybe two will survive once the dust settles. All these frontier labs are fighting tooth and nail to be the one left standing. Apple is making a measured approach saying "hey, we're probably not going to be the Google of AI, but the smartest thing for us to do is to put infrastructure in place so that once there is a contraction (which there will be, think the search engine landscape in the 2000s) we're relatively ahead for relatively little effort/cost."
maybe wait until their plan B succeeds before calling it genius. all we know so far is that plan A has failed.
I don't know why you're getting downvoted. You're spot on.<p>Almost everyone developing their own AI is doomed to fail. I would not be surprised that when the dust settles, Microsoft CoPilot and Google Gemini will be left standing while OpenAI and Anthropic fade away. They will get what they need and then throw them away.<p>"Embrace, extend, extinguish" is the same playbook for decades, and you can only do that when you make money. Yes, making money still matters and post-capitalism is a big fat lie.
It's the same plan. Apple's model is:<p>Handle the query on device. If the device cannot handle the query, push the query to Private Cloud, If Private Cloud cannot handle the query, push the query to a third party of the users choosing.<p>Which is pretty much the same as what they announced a few years ago.<p>Apple underestimated the amount of time and work needed to get to work and really the only bit that I think changed is that Private Cloud (and possibly the on device model) is a white label Google model rather than being developed in-house by Apple.
The problem with that approach is that “how do I handle this query”, and “did I handle that query ok” are both AI-hard problems, and must be tackled near the user, where the large model is not.<p>It also requires a relatively complex distributed architecture.<p>I worry about how successful they can be with this approach, in terms of best user experience.
I’ve been thinking recently about moving my notes from Notes.app to e.g. Notion because the latter jacks into ChatGPT and Claude, thereby giving them way more power as a personal assistant. This is the exact kind of move that would keep someone like me locked in.
The EU forced this on Apple and other manufacturers. Apple fought this bitterly, but there is only so much market share they are willing to lose before capitulating. <a href="https://futurium.ec.europa.eu/es/apply-ai-alliance/community-content/android-and-applesiri-simple-explanation-how-phones-could-let-ai-assistants-work-together-safely" rel="nofollow">https://futurium.ec.europa.eu/es/apply-ai-alliance/community...</a>
[dead]
That’s actually a good solution. Create a platform where app actions can be invoked with a permission model by any AI model.<p>Maybe apple can even have something similar to MCPs for tool invocations of apps and publish it so all models can actually be trained to use that framework
My favorite Siri behavior is telling it, on my phone, to turn on a focus state only for it to transfer the request to a nearby HomePod that tells me it can't help with that. I hope the new changes to it are an improvement, but the bar is currently buried in the dirt.
I would pay a good chunk of money if Apple released a local AI hub to coordinate all AI usage locally (including photo indexing).
That is rumored to exist.<p><a href="https://www.macrumors.com/guide/apple-command-center/" rel="nofollow">https://www.macrumors.com/guide/apple-command-center/</a><p>Mark Gurman seems to get extraordinarily accurate and detailed leaks. I wonder how Apple got this leaky. Pretty much everything about the Duo was known months in advance.
Photo indexing is already local. What specifically are you looking for?
You should look at Osaurus.
Once Siri gets MCP support, it’ll be my main orchestrator.<p>I built an app that lets me run terminals on my MacBook. It’s a Next app so I can go to that website on my phone using Tailscale. It has a built-in MCP server, so I can connect it to ChatGPT and have it create terminal sessions, orchestrate terminal sessions, all that good stuff.<p>Down the line, I’d like to be able to connect it to Siri and just talk to my phone and it can control terminal sessions and projects that I’m working on on my laptop.
Oh, cool. I used tailscale to provide a private web app for my cross-harness orchestrator, too. (<a href="https://github.com/BrokkAi/mjolnir/" rel="nofollow">https://github.com/BrokkAi/mjolnir/</a>)
I'd love it if Siri does, but I don't expect big corps to keep supporting MCP since MCP gives so much control to the end user, can't charge for eg. Google integration when someone can just add an MCP to handle it.
Apple Foundations Models Framework already aimed towards some generalized API for models. Personally I like it, just as I would be a little happier if instead of codex/claude/antigravity (rip gemini cli) I would have single standardized interface.<p>With mobile devices especially where ideally models could be shared across several apps.<p>But this specific code I'd assume is targeted to comply with EU regulations more than anything.<p>OT:<p>Since I've moved to Spain my principal Apple account is still in my home country but I've made another one for local app-store (since some local apps aren't available if you're not in the Spanish store).<p>iPhone is acting really strange, I'm logged to my principal account for iCloud+ and anything and local App Store.<p>If I log back to my principal account, I loose HCE mode and my building NFC app stops working until I re-log in the Spanish Apple Store.<p>If I'm logged at the Spanish account, I loose Siri AI which is blocked in the EU...
One can imagine a scenario where 1) this requires a Claude or ChatGPT subscription, 2) this subscription needs to be purchased through the iOS app. AI providers get new users, Apple gets its App Store tax, everyone wins.
Seems Apple's first attempt AI failing two years ago is turning into a blessing. Now they can buy AI as a commodity from the cheapest supplier. Well played.
Third-party calls won't eat into Apple's compute, so it's not entirely against the interests of Apple as they put caps on people's AI use.<p>I wonder what the margin on this stuff is for them:<p><pre><code> Some Apple Intelligence features in the Home app, such as
video summaries, are available for a limited number of
cameras depending on your iCloud+ plan — up to 1 camera
in the 2 TB plan, up to 2 cameras in the 6 TB plan, and
up to 5 cameras in the 12 TB plan. Supported camera
counts and other iCloud+ plan features may change over time.
</code></pre>
<a href="https://support.apple.com/en-us/127901" rel="nofollow">https://support.apple.com/en-us/127901</a>
Quite annoying that you need to get the top plan for 5 cameras, and you cannot get more.<p>I like tinkering with this stuff, everything around me is a perpetual lab, so I have more than 5 cameras that I'd like to hook up to this. But I REALLY don't NEED or WANT 12TB or storage.<p>I want all of these integrations to be possible locally, without having to send my data to the cloud.<p>But I cannot. A lot of these nice integrations aren't offered for fully-local setups.
I don't want to tell Siri to ask Claude or GPT - I want Siri to automatically delegate to another model of my choice when it can't answer something, instead of showing irrelevant web searches.
this does not provide much value on Mac, as Claude and ChatGpT apps both have global key activation.<p>iPhone is different, one button hold on phone without touch the phone is arguably Apple's last and only moat.
Don't they need that for China for compliance with their law?
That would be really cool, if you can route your local model for it.
In my opinion the OS should broker the communication between harnesses (which is part of Siri), LLMs and app functionality. It can take into account the permissions, the user could configure which LLM to use at an OS system settings level and complex workflows can integrate permissions like FaceID only once as the harness then interacts with APIs exposed by applications (arguably already there with app intents).<p>Self promotion blog article:<p><a href="https://braw.dev/blog/2026-04-06-mobile-os-as-agentic-broker/" rel="nofollow">https://braw.dev/blog/2026-04-06-mobile-os-as-agentic-broker...</a>
Feels like this is based on the idea that any model is interchangeable with any other, while all of them are better at some stuff, worse at some stuff.<p>Besides, the modern concept of "agent" is what, some year old? Bit fast to ossify stuff in the OSes already.
Hmmm then why is it not available in EU?
This appears to be the current feature prototype intended for the EU and potentially China. Apple has chosen not to release the new Siri AI in the EU because our Digital Markets Act requires them to allow interoperability. I.e. that users be allowed to choose another AI provider.<p>Ostensibly they'll activate this in the EU when it's ready. Given the current functionality, it looks ready (or close to being ready). My guess is it <i>is</i> ready and Apple is doing their usual negotiation tactics with the EU to see who blinks first.
Apple proposed a solution to the EU after reportedly extensive talks and asked the EU if it would satisfy them.<p>The EU said “build it then we’ll tell you”, which is crazy.<p>The problem is not having Siri use different models beneath. That’s not what the EU wants. They want Siri to be replaced by Gemini/etc completely. Which fits with the DMA.<p>Apple’s proposal was for something akin to an agent that could gather the necessary data for the 3rd party AI to use. Because what they refused to do is let any other AI have unfettered access to everything whenever it wants.<p>So I suspect this is really for Apple’s benefit. I don’t think the EU would be happy if you could still use Siri, but it was grok underneath. Because that would still be Siri.
My secret hope is that Tim Cook was the main weight behind cutting off half the new features from Apple software in the EU an attempt to weaponize their users against legislators. Even features that have had analogues where nobody would've ever asked for interoperability (i.e. Handoff/Continuity exists in the EU, but iPhone mirroring doesn't)<p>This is not how politics works here.
I think it is because the EU is requiring broader access to the deeper layers that Apple doesn’t want to give.<p>EDIT: by broader I mean all apps the user installs on their phone.
There’s a difference between Apple being able to swap out the model in code and Apple letting end users choose which model to use.
What does a full integration of claude or openAI models into the phone even mean? What additional functionality can one get and what are those use cases? Does it actually pull in more users to the voice/Siri interface? I mean even with vastly superior models, does it move the needle enough to attract more users?
Nice, thanks EU for forcing Apple. I guess it was possible all along.
Is there any hint that Siri can be connected to arbitrary MCP providers?
Siri, and Alexa, would be significantly more useful if it could be connected to ChatGPT etc.
It happens now – when I ask certain questions Siri asks if I want ChatGPT to answer it.
That’s gone in 27. You can still ask ChatGPT by prefixing your request with “ask ChatGPT“ like you’ve been able to for the last two years or so.<p>But it will never prompt you anymore, because Siri AI can handle that stuff now when it couldn’t before.<p>Whether you think it’s better or not is up to you. I do. I also found the version of ChatGPT it used to be worse than the one in the app.
Good news, Alexa+ (included with Prime) is already an Anthropic model, probably Sonnet.
Just playing the devil advocate: what happens when the answers from a not-Apple LLM are garbage? Who the users are going to blame, Apple or OpenAI, Anthropic, etc..?
I'm not 100% on Apple side on this, I would love to chose whatever I want, even local models, but I get why trey trying to control 100% of the user experience. If they could trust any provider with customers data, they would let them in, since they're not competitors. They would even let Apple not spend a ton of money on developing LLMs....
Are people blaming Apple when Google search in Safari gives garbage results? Some probably are, but that really shouldn’t be a relevant discussion, in particular when the user explicitly chose which AI backend to use.
If you are smart enough to change the default, you would know which LLM you picked and could point the finger at it.<p>Apple is in the game of providing an application platform. Maybe in the future they will provide their own LLM.
Now Claude will be listening to your audio 24/7 via the AOP logs:<p><pre><code> [PhraseSpotter] Detected wake word “hungry” with a score 0.942.
[com.apple.voiceactivation] Proceeding with full wake.
</code></pre>
And serve you ads accordingly (or sell the log data to other ad networks)
They could definitely take that risk.<p>They could instead make a good assumption that you'll search the web, open the Yelp or OpenTable app. Take your phone to your favorite restaurant and/or pay with a credit card at the grocery store.<p>If you walk to a taco truck without your phone and pay with cash, you probably don’t do that every time and you generate enough data through other means to be accurately advertised to.
There’s a reason everyone knows the various wake words.<p>All the big companies know how suicidal that would be. They already get accused of it by people who don’t understand how targeted web advertising works.<p>Some of the new AI start-ups don’t seem to understand that, but Apple absolutely will not make a mistake that stupid.
Right because they’re so benevolent?<p>High rep accounts like yours always come to defend, and you will probably dead this comment.<p>You claim that the AOP (always on processor) - which <i>must</i> always run in order to listen for “Hey Siri” - doesn’t actually listen for “Hey Siri”.<p>Or you claim that it <i>does</i> but they would never ever use the PhraseSpotter for that purpose because they’re too benevolent.
Apple is highly sensitive to dangers to its reputation. They would nuke any service that seemed to be spying on users, just like they revoked Facebook's cert back in 2019 over the same issue.
Glad I’ve never turned on Siri
It would be great if AI could be completely removed from iOS altogether, including any underlying frameworks.
this is likely true for most applications accessing the standard APIs, a good thing in my book
News at 11: Developers write decoupled code.
If this consistently has happened at every place you've worked in your career, I think you're luckier than you realize. Independent of the implementation, it's also still interesting as a sign of what might be supported for users down the line; I don't agree that this is how they'd write the code if they never expected to allow another provider, and other comments in the thread have already provided some insight into what the context might be (e.g. EU market).
Apple had enough experience relying on and being screwed by third parties that they have a built-in allergy to it. They do not like being boxed in and always want to be in control of their destiny.<p>This would’ve been useful during their bake off where they were trying to choose if they wanted to go with Google, Anthropic, or OpenAI. But even without that, they would absolutely not want to be in a situation where they had to redo everything to get off a bad partner.
Neat. Now let me point it to an OpenAI-compatible API endpoint of my choosing, like a Mac Mini running Hermes on my LAN.<p>I do not want my personal data or the queries against it to leave my network boundaries if at all possible. This should be a default option.
You can download an app that does that. You don’t have to use Siri, and it can be turned off entirely, or decoupled from AI entirely, and the wake word can be disabled entirely. A lot of options are available.<p>Make a Shortcut that will open your local voice assistant app of choice when you press the action button.
[dead]
[dead]