[internet2000]

> Apple runs on Anthropic at this point. Anthropic is powering a lot of the stuff Apple is doing internally in terms of product development, a lot of their internal tools…They have custom versions of Claude running on their own servers internally.

--Mark Gurman, Bloomberg https://x.com/tbpn/status/2016911797656367199

[rustyhancock]

Apple seems to purposefully have decided to sit out the arms race.

Probably smart time to rent and not buy if they plan on buying in a downturn.

[stefan_]

Okay, but why is the Siri team sitting out transformers. I really wanna move past the „Dragon Naturally Speaking“ experience with a bolted on decision tree.

[acdha]

Who’s doing it better? I have yet to hear from a Google or Amazon user who has a transformatively better experience, and I think that’s why they haven’t jumped so far because they have hundreds of millions of users who have daily habits that they don’t want to lightly disturb.

[simgt]

> I think that’s why they haven’t jumped so far because they have hundreds of millions of users who have daily habits that they don’t want to lightly disturb.

I don't think that's part of their decision making, Liquid Glass moved most things around for seemingly not much else than novelty and that's not the first time.

[nebula8804]

Liquid Glass makes sense if this is what they are working towards: https://www.macrumors.com/guide/20th-anniversary-iphone/

They have done this before, release something large early in anticipation of a major shift and iron out issues before the shift happens. Liquid Glass started off a little janky but they appear to have been ironing out initial issues with each update.

[steve1977]

From what I understand (which might be wrong), Liquid Glass was at least partially inspired by visionOS and "spatial computing". And I guess on that platform it might make sense for some use cases.

That doesn't change the fact that I can hardly read some of the user interface in Apple Music for example.

It's not that the idea is bad, but it's badly executed.

[dangus]

It only “makes sense” if you believe this concept also “makes sense.”

Nobody asked for a phone with fake buttons and a fragile wrap around screen.

Nobody asked for the UI to drastically change at random.

I wish smartphone companies would treat their products like they were completed devices with no innovation required. They are fully mature.

Instead, work on making them actually improved in ways that matter rather than trying to find “the next big thing.”

Be more like Toyota and less like Tesla.

[DonHopkins]

The worst problem they finally ironed out was Alan Dye.

[conartist6]

Really? None of my issues are fixed. The settings panel still has a massive gray empty chunk hanging off the bottom which makes it look like a 13 year old coded it...

[coastermug]

Is Liquid Glass not just a means to slowly force old phones to be obsolete? - My iPhone 11 is fairly slow now and they’ve probably bought forward my next phone purchase by a year

[bmitc]

Yea, if anything, it's Apple's normal mode to heavily disturb and move things around.

[bzbz]

Agreed. I vaguely remember another HN link that said Apple tried a competing-team approach to building a better siri, but it fell apart due to internal politics reasons?

[acdha]

Liquid Glass was also noteworthy for being the first macOS release since 10.1 which was worse across the board in a deliberate manner. They have shipped bugs before but this time it got such poor reception because all of the regressions were intentional and there wasn’t an expectation that they’d be patched.

I don’t think that cavalier attitude is universal at Apple and I don’t think the Siri PM wanted to break with their past respect for UX.

[alfiedotwtf]

When a company doesn’t have anything to innovate on, or hires a new marketing exec, the first thing they do is change the company logo.

Liquid Glass was Apple’s logo change moment

[pram]

There have been many metaphorical and literal logo change moments in Apple’s 50 years champ

[wenc]

Right now Alexa+ and Gemini are objectively better.

The best is ChatGPT voice mode. It understands non English words and accents amazingly well, and even though the LLM model isn’t the full fledged one, I can have deep conversations with it for an hour without it missing a beat.

[barumrho]

Siri doesn't need to have conversations with you. ChatGPT can do that. But, it should be able to do actions you'd do on your phone.

[cachius]

Speech to text should work. I regularly have to manually edit the transcribed input. The more special words the more frequent. Completely disregards the context of the current input, for example, on Hacker news might involve special technical and IT vocabulary.

[skeledrew]

Pretty straight forward on Android at least to wire up a harness that talks to Tasker[0] or another full automation app.

[0] https://tasker.joaoapps.com/

[HereticLocke]

I agree, ChatGPT voice mode is pretty impressive. Almost similar to Samantha in 'Her', laughably.

[bobthepanda]

Scarlett Johansson is suing OpenAI, in fact

[nozzlegear]

> Almost similar to Samantha in 'Her', laughably.

Things that Sam Altman would prefer people not say lol

[alfiedotwtf]

This! I talk to ChatGPT every morning, and will listen and navigate my feeds while I drive, summarises posts, answer my questions. It just works.

[CountHackulus]

Alexa+ has been a massive downgrade for me. It's extremely laggy and constantly misunderstands me, whereas the old one never did. "Set a timer for 20 minutes" used to be instant and just work, I did this the other day and it took 10 seconds to respond and set a timer for 10 minutes.

[mh-]

I had the same experience until I upgraded my Echo (we have a few, but the one in the kitchen gets 99% of the voice commands).

Just looked it up in my order history: I went from an "Echo Show 5 (1st Gen, 2019 release)" to a "Amazon Echo Show 8 (newest model)".

Whether I should have needed to upgrade is a separate question, but, yeah.

[hamdingers]

Same here. I can see why LLM-driven voice assistants makes sense to product people in the abstract, but introducing non-deterministic behavior into a device I primarily use to help with timekeeping and control lights is nothing but a regression.

[virgil_disgr4ce]

I concur that the ChatGPT voice mode is excellent. I can't even think of anything to knock it for other than for whatever reason it never 'hears' my kids, but that's probably because it's not intended to be used in multi-participant chats?

But for one-on-one, it is a really outstanding experience. Especially since they tamped down the way over-the-top humanisms.

[DaiPlusPlus]

"objectively better" is a subjective statement :)

My preference, however, is for a voice-control UX just like I get with my Amazon Echo and "classic" Alexa like I have been for the past 10 years I've been using it: I think I can best describe it as a "voice-driven command-line" just like your OS' CLI shell, which makes its interactions predictable, even if it means I need to "know" what commands are valid in a given context. We all need predictability and reliability when it comes to my home-automation integrations.

...but computer interaction with a LLM / transformer-driven / "AI agent" is anything but predictable. When Amazon opted everyone into Alexa+ I agreed to give it a go and see if it really made things better or not - and it did not. I opted-out of Alexa+ and went back to something actually reliable.

[thrtythreeforty]

Here's a question: I don't understand the gap between these LLM powered voice agents vs CLI coding agents, the latter of which are obviously useful and quite resourceful at getting something done when asked in plain English.

Seems like an agent given 20-30 tool calls like "read_sms" "matter_command", and "send_email" would be able to work out what to do for things like "set the house to 72° and text Laura that I did it."

[redwall_hp]

Siri's one job I care about is doing exactly what I want while I'm driving. I need it to check my text messages, take dictation, start phone calls and deal with music. I don't need to have conversations with it, I need deterministic responses to known commands.

[wat10000]

"Objectively" has become a generic intensifier. It's literally infuriating.

[ShyCodeGardener]

Whenever I see one of these comments, it's always from someone that tried it at the start and then gave up because of a bad experience. And many times there are more people commenting back that this was essentially the 1.0 version and that the current 2.0 version is much better. So as someone that uses none of these products (old voice assistants vs. ai ones) it's really hard to evaluate if any of these anecdotes mean anything.

You could have tried Alexa+ at the start when it was shitty compared to plain Alexa, and maybe it's better now. But equally none of the people that comment that it is "amazing" in its current iteration qualify their statements with their experiences comparing and contrasting the old version vs. the new version making them seem either unqualified to make statements based on how much "better" it is than the old version or at worse they are shills (paid or not). The best take is that they are comparing (e.g.) day-one Alexa+ vs. the current Alexa+ without a comparison to the original Alexa.

... which is to say that it really feels like there are no clear conclusions that could be drawn from all of this.

[stronglikedan]

Alexa+ is terrible compared to Alexa. It's so bad that I've dusted off my v1 echos cuz they're too old to run Alexa+. Complete shit show that is.

I do like Gemini better than Assistant, even though it's not quite there yet. But that's just a matter of time because they actually designed it from the ground up to be a drop in replacement for Assistant.

[acdha]

I’m curious, what are you talking about for that long? It sounds like that’s moving out of the home automation space into something else.

[recursive]

Google user here. My experience with the new assistant is worse. The old one could pretty reliably set timers. The new one could not.

[BeetleB]

Oh man. I made the mistake of converting my Google Home devices to Gemini.

The first problem is that it's just slow. If I want it to turn off some light, it takes a long time before responding.

But yeah, the failure to do basic tasks. I have a routine that I used to have it run (controls several devices at once). Now:

10-20% of the time it runs it.

60% of the time it says it's running it but it doesn't do anything.

20-30% of the time it says it can't do it unless I opt in to invasive permissions. And when I opted into them, it still failed about a third of the time. So I opted out again.

[recursive]

I don't know if it's related to Gemini, but sometimes the Android Auto tells me "I don't have permission to do that" simultaneously with actually doing the thing that is allegedly lacking permission. Sometimes I want to move off the grid.

[stickfigure]

Strong disagree. The upgrade was a little bit rough at first (mostly because of slow response) but now it's a million times better than the old assistant. The old assistant basically just repeated "I don't know how to do that" over and over.

I have never had trouble setting timers with either.

[recursive]

The new one was 100% failure to do anything with timers for me. I never saw it work once. If I had ever gotten that to work at all, I may not have uninstalled it, and might have a different impression now. I cannot account for why our experiences are so different.

[cmckn]

I like that I can ask more nuanced questions without getting dumped to a web search, and I can have a back-and-forth conversation.

But timers and smart home actions are definitely less reliable and sometimes take absurdly long to respond (like 20-30 seconds p99).

[recursive]

Your experience is valid of course, but I never once have had the inclination to have a conversation with my phone. I'm not sure which of our experiences is more common.

[netsharc]

I hate it too.. the old assistant is pretty smart, obviously it has some language processing, but not "AI", but it's very fast for things like "Set alarm for ... ", "Remind me at X about Y", "Add calendar event on x at y about z", or "Navigate home".

And now if I want to use Gemini on my phone I have to replace Assistant. Nah, I'll keep Assistant thanks, and just have a shortcut to load the Gemini in the browser.

Except the browser experience is so fucking buggy, constant reloads needed..

[bakies]

Claude.. I switched my phone assistent to claude and it does everything that google (used to) do like set alarms and timers, but also does everything claude can do.

[saratogacx]

The only thing I haven't been able to get it to do is read from my phone's local calendar. The claude app can but the voice assistant cannot (Why? No idea). Perplexity has no issue doing it so I actually use them for my rare needs to do voice commands with my phone.

[arnavpraneet]

How did you do that?

[bakies]

Settings > Apps > default apps > assistant

[infecto]

I was looking myself and it appears only certain regions (Japan) have that option.

[nerdsniper]

On just the transcription side:

WhisprFlow produces much better speech-to-text for long text messaging-by-voice (dictation / transcription) than apple's speech-to-text does. Whisper models in general seem to do a lot better than most built-into-OS/app models. Which is interesting, because there's nothing stopping them from just using Whisper models.

I love MacWhisper personally. Also, Gumroad is a fantastic app distribution platform for my personal values.

https://goodsnooze.gumroad.com/l/macwhisper

As far the "decision tree" side ... there's not much that can be done about that now. Agentic agents still go too "off-the-rails" to be productionized out to the billions of smartphones of the world. I'm working on voice-controlled agentic-with-rails AI features for my HomeAssistant, because Alexa / Google Home suck. But that's a hobby project and rogue AI actions only affect me, not billions of customers.

[EwanToo]

As a Google user in a household of iPhones, my opinion is Gemini on android is radically better than Siri.

[naravara]

It’s not “transformatively better” but it definitely involves fewer frustrations to interact with. That’s always been Apple’s main value proposition, you’re not getting the most cutting-edge stuff but you’re supposed to have something that “just works” not something that makes you go “GODDAMN IT!” when it inexplicably seems to fumble normal things.

So if you buy Apple products based on that value proposition it’s a big problem for Apple if they can’t seem to keep their brand-promise in this area.

[taude]

My android phone was so much better for voice-to-everything. Whether it was transcribing my voice for text messages, or doing looks on the internet. Siri is just so bad.

Still love not having google's paws all over my data, though, so not going back.

[hrimfaxi]

> Who’s doing it better?

Any of the Whisper-based apps on the App Store.

[Wowfunhappy]

Actually, could you recommend one? The ones I've found all seem to want subscriptions. I'm okay paying a few dollars for a well done frontend, but an ongoing sub to run an open weights model locally is nuts...

[oceansweep]

This is one I use with no tracking or ads:

https://apps.apple.com/us/app/id6447090616

[oceansweep]

I’ve found this one to be useful offline with no ads:

https://apps.apple.com/us/app/id6447090616

[BoredPositron]

Here have an anecdote: Gemini Assistant is pretty good.

[nozzlegear]

Feeling #blessed that I apparently have the exact same upper midwest accent they must've trained Siri on, because I've literally never had an issue with dictation or being misheard. And I use it a lot!

(It misunderstands my wife from California all the time, though.)

[Cthulhu_]

Plus, if someone else does it better (or different), I bet they've got a team and technology at a 90% done state waiting to jump on it, pick it apart and make it better. I don't think they're not doing anything.

[bmf]

Wasn’t planning on it, but Show HN: JoinIn.AI?! We’re working on making a better audio interface to LLMs that is socially adroit enough to handle even multi-party conversation.

[brightball]

Grok integration in a Tesla is doing it muuuuuuuch better.

[phrotoma]

Yesterday my google home mini gave me the current temperature in farenheit. I live in Canada and use a pixel. Dumbest fucking AI going. May as well give it to me in coulombs per hectare.

[stetrain]

Not sure "sitting out" is the right way to put it. They've been publicly trying to ship a next-gen Siri for years and haven't been able to get something good enough to release. The latest plan is to base it on Gemini so we should be seeing progress on that next month at WWDC.

[readams]

Gemini will be replacing the legacy Siri:

https://blog.google/company-news/inside-google/company-annou...

[pxc]

The experience of using LLMs as digital assistants so far is not great. Gemini on Android sucks so bad it's hard to describe. It can't tell what its own capabilities are, it can't inspect the states of the apps it manipulates, it hallucinates constantly, and it needs more handholding than the crappy old decision tree to do the right thing. I much more often have to pull over to make sure Google Maps is doing the right thing than I ever used to before, because trusting the LLM to be "smarter" so often fails for me.

Be careful what you wish for.

[GeekyBear]

They did create a chatbot version of Siri small enough to run locally, but decided that hallucinations were a big enough issue to push the release.

[colechristensen]

I think they could never make it good enough at the right price.

You have to remember all of the AI companies are making cash bonfires. People aren't going to stop buying iPhones because Siri can only do what it does now.

If Apple focuses on hardware and skips the pay-for-inference bubble they'll come out the other side with the best consumer hardware everybody already has for local inference which is going to eat the whole industry's lunch.

nvidia is going to have a hard time convincing people they need to buy $1000 LLM inference hardware. Apple isn't going to have a hard time convincing people to buy the next generation of phone/tablet/laptop.

[yieldcrv]

Because the competitor voice models sound good but are dumb upon any scrutiny

ChatGPT’s voice model has a great user experience and seems like it is seamlessly integrated into the chat, but its actually a far smaller and dumber model. @husk.irl on instagram has videos displaying how dumb and undiscerning it is

People were wowed by the magic at one point, but its faded. Apple avoids those things and the limitations havent been solved

[insane_dreamer]

in my experience voice recognition on Claude (using the iPhone app) isn't that great -- maybe even worse than Siri (I'm referring to voice transcription specifically, not the inference of course)

[gchamonlive]

I think it's the same reason why MacOS and iOS degraded a lot in terms of UX the past decade. The focus of Apple shifted towards hardware independence.

The 2010s was marked by Intel's lazy product lineup, year after year pumping rehashes of older products, iterating on top of their 14nm lithography with increasingly minor improvements on its architecture until AMD overcame them. In the process, Apple's partnership with Intel became a liability it had to solve, and a push for the unified ARM architecture was no small feat.

If you ask me I don't think it's justified to degrade the user experience for the sake of focusing on this. It's a trillion dollar company, and has been for a while. Sure it could have tackled both, but what do I know.

In any case I think it explains really well why Siri feels so abandoned.

[threetonesun]

I dunno, Apple has always had a pretty high level of hardware independence, and one could imagine even if Intel did produce great chips for longer the ARM architecture would replace it eventually. Certainly the timeline got shifted (and I'm glad for it) but I don't know if that really impacted Siri. If anything it seems like it got pushed to the bottom of the pile in favor of projects like the Apple Car and Vision Pro OS one on side and the demand to increase services revenue on the other.

[Wowfunhappy]

Also: Before, Apple was dependent on Intel (whose "product" is an integrated chip design and the fab to make it). Now they're dependent on TSMC (whose "product" is a fab). I'm... not really sure they've reduced their dependence? If TSMC starts falling behind Intel--which doesn't seem likely, but what happened to Intel didn't seem likely two decades ago--Apple will be stuck.

[GeekyBear]

TSMC does have competitors.

Intel is already being evaluated to fab Apple's entry level chips, if they can meet performance, energy efficiency, and production targets.

[newsclues]

A series is their own chip design, not Power PC or Intel designs.

It's the CPUs they have built for their purposes, which is next level hardware independence.

[Cthulhu_]

It's one of the biggest and wealthiest companies in the world, but your comment seems to imply they have to pick and choose what they pursue. They really don't, especially if it's hard- vs software.

[footydude]

> seems to imply they have to pick and choose what they pursue. They really don't, especially if it's hard- vs software.

Money can often just be one part of the equation.

To do things well you also need - available & capable technical resource, suitable facilities, available & capable leadership and management (with engaging at the right level in the business) and a clear vision of what you're trying to achieve/working towards.

Given how Apple appears to operate, I wonder if a strong desire for senior management control/oversight over major developments means they (artificially) limit how many concurrent large-scale things they can work on at any given time?

Maybe not, but that'd be my guess.

[gchamonlive]

> It's a trillion dollar company, and has been for a while. Sure it could have tackled both, but what do I know.

I didn't imply, it's explicit in my comment. it's what their actions show. Their updates make their systems worse and worse, Tim Cook is out and Siri is in shambles. It might have been something else, but I'm willing to give it the benefit of the doubt, because the alternative is just sheer stupidity.

[realusername]

I always found that Apple had pretty mediocre software qualify, it's always been a very strong hardware company first and foremost.

They have great kernel, drivers and low level engineering but the stack above that has a lot of questionable stuff.

[rkapsoro]

I only partly agree with this. The answer is maddeningly more complicated.

Some parts of their software stack -- higher up than the kernel -- are actually pretty great. There's a lot of realy brilliant stuff in their system frameworks, and in SwiftUI, Cocoa, and UIKit. I've been using Linux at home recently, and I find myself missing some of it.

But, on the flip side, suddenly you hid maddening bugs, crashes, or terrible developer-experience papercuts. And, of course, there's the App Store, which is just evil. For my next app I'm just going to go Notarization only, and see how that goes...

[pfisherman]

The comment above is on to something. I find CarPlay to much more valuable and much more of a lock in to the iPhone than Siri. I do not think I could ever go back to using the infotainment systems that ship with cars. So makes sense why they might prioritize over Siri. And in the context of CarPlay, the simplicity of Siri is nice. I really only need it to execute a few simple commands like looking up directions, making calls, reading / sending texts, playing a podcast, etc.

[BoingBoomTschak]

>They have great kernel, drivers and low level engineering but the stack above that has a lot of questionable stuff.

Do they? Is Linus' rant about porting git to OSX now obsolete? At least, unlike MS with ReFS, they managed their HFS+ -> APFS migration.

[gchamonlive]

I don't dispute that, but Apple made its business on the premise of being the best in the business in terms of UX. Note though that you can have great UX powered by mediocre software, so those aren't mutually exclusive.

[robotresearcher]

Apple’s Mac software in the 90s had great UX and very underwhelming and old fashioned kernel software which they struggled to replace. Jobs knew this and did the work externally with NeXT.

[HumblyTossed]

They're valued at $4T, they have hundreds of billions hoarded. They could run 50 billion dollar startup projects and not feel it. Imagine a startup getting handed a billion dollars ... and the vast knowledge that Apple has access to already.

There's no way they couldn't do a better Siri. For some reason, they just ... won't.

[gnerd00]

here is a clue delivered -- money does not make software better, and lots of money often results in worse.. it makes no sense? actual experience begs to differ.

Classical homework assignment -- the Mythical Man Month and related essays

[ux266478]

Money is a means to an end, so that's true. Just because you have a screwdriver does not mean you will drive screws. You have to have someone who can use the screwdriver, knows righty-tighty, wants to drive the screws, etc. Stupidly throwing money at a problem can get you places, but the efficiency can also drop to near zero. The problem is we're talking about a quantity of money where you don't need to be highly efficient. Savants can do pioneering work in a cave with a box of scraps, but you don't have to strive for that kind of austere efficiency. Nobody is expecting that.

If Apple can't harness the potential of the currently overfilled labor pool, that indicates a systemic issue within Apple. The entire raison d'etre of management structures within a business is to increase efficiency of capital to drive productive forces. If they cannot do that, then that would indicate an extremely problematic competency crisis within Apple's management organ.

This kind of failure when you are a company with the valuation of a first world country's GDP should be raising alarm bells in any rational person's mind.

[HumblyTossed]

I’ve read all that. I know all that. None of that changes my point.

[foobiekr]

Heavily funded startups have terrible track records in reality. The only cases where it seems to have worked is when the money was used to undermine the market dynamics by nuking competition via severe underpricing.

[api]

This is how they usually roll. They innovate sometimes in hardware but tend to fast follow or even slow follow in software and services.

Apple Intelligence is a placeholder and a toe in the water.

[xnorswap]

They fast-follow then market so aggressively with just enough proprietary tweaks so they can trademark it that people think that Apple invented the technology.

[evilduck]

People end up thinking Apple invented something because they tend to make the first usable version of something that could appeal to the general population.

[gambiting]

I literally can't think of an example. Care to share one?

[stouset]

I’ll give you a hint: you may very well have replied on this category of device.

[dzhiurgis]

Wonder how much of it is raw compute constrain. Adding additional 1billion LLM users is near impossible. I spend a lot on tokens and still get throttled.

[mathisfun123]

I love these comments - have you ever heard of Occam's razor?

[iLoveOncall]

Not participating in the war is the only true way to win the war, nothing new.

And in this particular war, it's even worse, the "winner" will actually just be the "biggest loser", contrarily to a traditional war.

[dylan604]

It seems to be Blu-ray vs HD-DVD again. Luckily for me, I made the right decision and got out of the shiny round disc business as that battle was raging all around me having been in the DVD programming business for 8 years or so. This battle of LLMs is interesting to watch from the sidelines as I have nothing to do with them. Not sure this will end with one LLM to rule them all while the others fade away. People can use the one they prefer and not really impact others.

[joe_mamba]

>Not participating in the war is the only true way to win the war, nothing new.

Really not true both in real wars and in tech wars. There's no evidence to support this claim.

Android only exists as the dominant mobile platform because it went to full scale war with Apple when the iPhone launched. Those that didn't take part and came after the battle have like <1% market share and Apple and Google are printing money from the cut to their app stores.

Apple doesn't take part in the AI race because whichever AI wins the war in the end, they'll have to be on their Appstore to reach the users, so Apple wins regardless due to their Appstore monopoly. AIs are no threat to their phone, laptops and Appstore business.

But Google can't afford not to take part in this race because AIs are a threat to their search and ads business.

Same with real wars, US is the world superpower because it got involved in WW2 even though it didn't have to be. Same with Russia and Ukraine, provided they don't wipe each other out scorched earth, their militaries will be the most advanced on the planet on modern drone warfare they invented after the war is over, and every other military on the planet will be paying them for their gear and expertise, which they already are.

[piker]

I'm suspicious of that take from Mark Gurman. That's a lot of detail around pricing and "holding Apple over a barrel" as relates to the Siri deal that seems like a nice PR spin from Anthropic.

Anthropic probably couldn't give the uptime guarantees that Google can, right?

[Spooky23]

Apple is a pretty difficult company to deal with on a B2B basis.

If you have terms that conflict with theirs, they aren’t very flexible. Anthropic can be similarly difficult, and their needs from a business perspective probably don’t align with Siri. I would imagine that Google has a more flexible/long term approach to absorbing some risk in a revenue share arrangement than anthropic who generally wants cash.

Anthropic’s only purpose is to juice whatever KPI‘s are gonna increase their IPO market cap.

[curiouscats]

Good thoughts.

The last sentence doesn't make that much sense to me though. An agreement with Apple to be the lead AI partner would likely juice the IPO a great deal. The financial details wouldn't matter much for the IPO (as the initial financial commitments are going to be small but the halo effect would be real - I think it would in the market anyway).

I think Anthropic has real commitment to their way of doing things which can cause short term issues (and hurt the IPO). And they seem willing to keep those values rather than just making deals to pump the IPO. As you say Apple also sticks to their way of doing things even if it frustrates their partners.

I think not being the lead partner with Apple may well be good for Anthropic long term. But if all you cared about was the IPO just agreeing to Apple's terms likely would have been the best option.

These SpaceX, Anthropic and Open AI possible IPOs are so extreme it is hard to make judgements about them; so maybe there are Anthropic IPO issues to an Apple agreement that I don't appreciate.

[Spooky23]

It depends on what Apple wants and what Google was willing to give. Google is in many ways the weakest player in the individual-user facing space.

It's a weird market and these companies want global domination. TBH, i don't have the knowledge or context to understand how to think in that mode and what the real facts are.

I wouldn't put much stock in the deeply held principles of Anthropic (or Apple for that matter). That's an appeal to emotion. I love the product, but they're happy to randomly rug-pull the product and how it works, both in the publicly available products and other contexts. It's just another company.

[curiouscats]

I agree that Apple (and probably Anthropic) don't hold to the expectations some have. But I wouldn't say they are "just another company." I care about privacy more than Apple. But I think Apple does more to help privacy than any other huge company (and I can't think of another huge tech business to consumer company that does less to invade privacy, but I would be open to evidence I am wrong on that).

Apple is far from perfect but that doesn't mean they don't have a position (say privacy) that they care about and give a lot of weight to when making decisions. But as a huge company they also have many competing priorities.

Caring about privacy or potential abuses of LLM/AI services does not mean that a huge company is going to perform on those areas the way those that want maximum privacy or... want.

I do also believe Apple's marketing reasons to promote their focus on privacy make sense. And even if they don't do as much as I would want they do make a big difference (on privacy) it seems to me. And I believe long term there is big value to Apple building systems to stop users private data from being abused.

[sailfast]

You say that, but don’t you think at this point they actually believe some of the stuff they say about safety and the future of humanity? It’s tough in this day and age not to be overly cynical but they did draw a line in the sand at the DoD and that wasn’t for IPO numbers…

[Spooky23]

I'm too cynical to believe the kubuki theatre.

The US government in 2026 is openly and cravenly corrupt, and I don't believe anything at face value. The story about the targeting may be real and material, or backwards engineered to fit the reality. OpenAI is aligned with Larry Ellison and Oracle, and given the favor granted to them by the government, I'd look to that relationship first.

[piker]

Yeah, that makes more sense to me than "Anthropic had them over the barrel". Which seemed quite odd given the relative cash positions and installed base of each firm.

[engineer_22]

Tbh I thought their purpose was to power the war machine

[Lord-Jobo]

Gueman might be the only leaker in tech who, so far, doesn’t seem to fuck around. Low miss rate, rarely exaggerates. Of course that could change and he could always get insider info that is wrong.

[lostlogin]

A recent big miss of his was Cooks retirement.

https://daringfireball.net/linked/2025/12/01/gurman-pooh-poo...

[turtlesdown11]

Gurman is clearly Apple's preferred go to for leaking info

[blitzar]

Which only tells us that it is what Apple wants us to believe, not that it is the truth.

[zymhan]

It is most likely both: Intentional leaks of the truth.

Obviously, _what_ someone chooses to leak can still benefit them, even if it's true. You can be selective about what information you share.

[turtlesdown11]

absolutely, a careful eye can discern Apple's goals by examining leaker comments on topics, not that they're consistently truthful

[signatoremo]

Did you verify that? What is the miss rate?

[danpalmer]

The reporting says it's running on their own hardware.

[piker]

Internal dev tools, but the point I'm making relates to the discussion about choosing Gemini over Claude for their consumer-facing products.

[jedisct1]

> They have custom versions of Claude running on their own servers internally.

This is the important point.

Sending their internal code, documentation, secret tokens, etc. to Anthropic would be completely irresponsible.

But if they are running the models on their own servers, why not!

[JeremyNT]

Was it even publicly known that Anthropic offered this capability? I wasn't aware on-prem Claude was a thing.

[sheiyei]

If you're Apple (or even Apple-sized), you can get a bunch of things others can't.

[conception]

Bedrock? If you’ve got the cash they’ll deploy it.

[halJordan]

Yes it was known. The usg is also running their own copies on fedramp data centers (for now)