I wonder how well Apple has deployed these tools internally for security research.
Since mid-April Chrome showed 302 vulnerabilities patched, 225 of them found by Google. Same period last year was 19 vulnerabilities. They've also become more transparent recently, disclosing vulnerabilities found internally, not just externally (which Apple still doesn't appear to do). From the outside, it's hard to tell if Apple has deployed this tooling as much as Google.
I am part of Apple's SEAR (Security Engineering and Architecture) organization and can’t attest that we have been using Anthropic models, including, but not limited to, Mythos, as part of our participation in Project Glassing and previous private partnerships with different frontier AI labs for years. We simply don’t talk about it because there’s no benefit to talk about it, and also NDA’s, but mostly because there’s no benefit to talk about it other than to satiate people’s curiosity about what we do or don’t do internally.
The heavily ironic implication is that they're under NDA, so they can't attest to it, while more or less attesting it. Senator, I cannot confirm or deny that we definitely do this.
This could also be an unofficial-official way for Apple to "leak" that yes, they do this--which is on brand for how Apple handles "rumors" etc.
When the CIA representative says "I can neither confirm nor deny" it generally means the atrocities of which the agency has been accused did, in fact, take place.
When I worked in the civil service we were trained to use that phrase to any query, no matter how innocuous (unless we had permission to give more info).
You may think that not issuing a categorical denial is suspicious, but generally speaking you cannot infer any information from that response. If it was only used when really bad things might have happened, maybe you could infer more.
I think Apple became much better at security in recent years. One example which I think is indicative of their approach to security - they bothered to add a hardware microphone disconnect when a macbook is closed. Source: https://support.apple.com/en-gb/guide/security/secbbd20b00b/...
What's your thinking on this? From my perspective Apple security go pretty hard. They have a strong track record of being able to ship architectural mitigations like PACs / MIE / Exclaves first. I guess because Apple control the stack from silicon to userspace.
My thinking was in a historical context, and for their desktop OS's. I know they've been pretty on top of things with iPhones, and MacOS has become a lot better, but for the longest time MacOS was pretty lacking, coasting very much on promoting how much PCs have viruses and macs didn't, which was a marketshare thing more than a security thing. I don't think they got ASLR until later than pretty much everyone else, for example.
They've improved a lot, especially their phones, but I'd still never consider them a company that has a really strong focus on security.
They were not "coasting" on anything. Everything about OS X has always been designed to protect users from the stuff Apple hasn't caught yet, because they know they can't always catch it first - and Apple has led the pack in nearly every major OS security feature of the last 25 years.
That includes "don't give the user root, and ask the user for their password before doing dangerous things" - four years before Linux distros started moving to a similar model.
Didn’t Microsoft pioneer the privilege escalation prompts in Vista in 2007? It was a joke at the time how little things would hijack the entire screen to allow seemingly mundane things. I didn’t ever use Vista personally or professionally, but macOS has become pretty bad with basically the same model.
IMHO, both are a mode of progressively penalizing developers as a mode of API obsoletion. It doesn't feel like the opportunity to fix a degradation of user experience really motivated app developers in either case.
The difference is Apple is much more likely to progressively make these legacy feature compatibility more difficult for users to configure over time, and to remove them eventually.
Microsoft's Secure Desktop feature is actually incredibly well designed, and provides strong protect against fraudulent prompts or prompt interception attacks.
It is the default (unless they changed it in the last 2 years or so). I know for a fact that my PC and Laptop don't ask for my password and I know for a fact that I reinstalled Windows on my laptop less than 2 years ago and changed nothing regarding the UAC prompt (the closest that is even remotely close is enabling sudo in the settings).
Yeah, they were. Virus writers were not targeting them as a platform because why develop for 10% marketshare when you can target 90% for free. It just wasn't worth it to target as a platform. So there was some level of protection due to lack of interest in distributed attacks, but the OS had very little protection against targeted attacks.
> Apple has led the pack in nearly every major OS security feature of the last 25 years.
What an absurd claim. Apple trails behind, it never leads in this space. Windows 7 had numerous protections that had become standards that Apple still lacked when Windows 10 came out.
Recently there was an Anki vulnerability that gave any website access to any local files. On Windows or Linux this would be deadly. On macOS, Anki can't access my desktop or documents or Chrome storage or password manager storage. I think Apple's been smart about which security features it prioritizes.
> That includes "don't give the user root, and ask the user for their password before doing dangerous things" - four years before Linux distros started moving to a similar model.
Linux distros have always required sudo for "dangerous" things. What distros made users root by default?
That's a really strange claim given AS was a refinement of a technology other manufacturers have yet to surpass in the ten years since the T1 chip came out.
To this day nobody else ties their SMC, biometric auth, and HSM together as tightly and well as the T1 did. AS was further advancement of that.
Furthermore, Apple protects users against the legal changes that have allowed law enforcement to physically force someone to provide biometric credentials. By default MS just provides biometric auth to make it easier to log in to your system.
iOS always had a strong focus on security but if you take the time period say 2005 - 2015 it did not seem like there was much investment in macOS security at Apple. I am talking about stuff like exploit mitigations and relatively low hanging LPEs. Features like (full) ASLR / SIP / kext controls were added well after competitors.
> I guess because Apple control the stack from silicon to userspace.
People always say this but there is no real relationship there. When hardware vendors add security technologies to the hardware, the major third party operating systems add support to use it pretty much immediately, and in many cases before the hardware even ships because the hardware vendor publishes the documentation ahead of time.
Try to name something where Apple was the first to support something (by a non-trivial amount of time) not because they were the first to add hardware support but because they released the combination of hardware and software in the time between when e.g. Intel or Qualcomm added hardware support and when Linux or Windows added software support to use it.
> The affected releases include iOS 18.7.9 and iPadOS 18.7.9, macOS Sequoia 15.7.7, macOS Sonoma 14.8.7, and macOS Tahoe 26.5.
Where does this quote come from? I can't see it in https://support.apple.com/en-us/127115, the article link at time of writing. It mentions CVE-2026-28952, but we're forced to guess why. I'd take the reference to mean that this issue is fixed, but I'm just some internet rando, so what the hell do I know?
If I do a google search for "CVE-2026-28952", it points me to various pages. Here's one, for example: https://www.cve.org/CVERecord?id=CVE-2026-28952 - which is a bit more explicit, though of course this is not from the horse's mouth:
> This issue is fixed in iOS 18.7.9 and iPadOS 18.7.9, macOS Sequoia 15.7.7, macOS Sonoma 14.8.7, macOS Tahoe 26.5
I haven't been able to update my iPhone in months because it just does not have enough room available to download the update. I just checked now and it needs 13.2 GB free to be able to update to iOS 26.5 (from 26.3). On a 64gb device!
It just seems like massive software development malpractice to tie together critical operating system updates with whatever else they've bundled.
I have a 32 GB iPad, I think it's the year 2020 model. The OS alone uses 19 GB ("iPadOS" 12.3 GB + "System Data" 6.4 GB) so yeah, not much chance doing any OTA updates on that one with the requirement of 13+ GB free.
Maybe some day the fruit company with all their billions will be able to innovate a solution for deploying for example browser fixes so that they can be installed without requiring tens of gigabytes of free storage on the device. Meanwhile, we're stuck using a computer and iTunes for that.
It’s insane. It’s always an ordeal. They put so little storage on these phones that 20% of it is for iOS/system already. On top of that requiring 13-15 GB for an update is a huge pain.
For many years my go-to plan has been to stay one point release behind apple's releases, especially the .0 releases -- but, times change. Last night I pushed the button for 26.5, thinking about the Glasswing/Mythos reporting. Seems like staying on bleeding edge is going to be the name of the game.
I wonder if this will change general dynamics -- feels like LTS releases could become even more important, at the same time having reduced maintenance costs since you can have some agentic help on backporting.
Staying one point release behind is weird isn’t it? I get staying a major release behind, Apple’s x.0 releases are often pretty rough so it might be worth staying on x-1 for a while. But point releases mostly just fix the stuff they broke in the major release.. Would you really upgrade from 18.5 or whatever to 26.0 when Apple releases 26.1?
Point releases for macOS can be pretty large over the past several years - what often makes sense is waiting a few weeks to upgrade in case there's a .1 patch.
e.g. macOS 15.0, 15.1, 15.3, 15.4, 15.6 and 15.7 all had .1 patches within a few weeks of release.
CVE-2026-28952 is about an integer overflow due to lack of input validation. I wonder what makes such vulnerability difficult to discover by traditional SAST tools?
The tools are expensive. One of the major players in the market have really expensive licensing fees. Then the developers all need to be trained on how to use the tools and understand the results. It’s not something they teach effectively in schools.
Software engineering is still kind of new overall.
Apple has a massive information security organization that has pretty intense resources at their disposal.
It seems borderline impossible that there's a tool that they feel would be beneficial but that they're classed out of using by license costs or by staff proficiency.
I find this amusing as Apple were the people I had direct interactions with that didn't run stuff like fuzzers or sanitizers as a matter of course - at least not in the situations I was involved with.
Though this was ~3 years ago now, and a lot of things have changed, but these tools were very much available and well known then - they aren't new. Though perhaps as they "knew" the project was coming to a close it wasn't a priority either?
It also might have fallen through the gaps due to the Apple internal/team culture - I worked for an external vendor, and we had to work against binary built framework dumps that didn't even allow us to enable things like address sanitizer completely either, and fuzzing difficult as you'd need to trace things through their opaque binary layers before it even reached our code.
Apple did have all our code though, it was very much an asymmetrical relationship, but if they were running such things as a matter of course in CI or similar you'd see that pattern in when they reported issues it caught, and the timings from time-of-bug-caused to time-of-report. It instead suggested any such runs were piecemeal and sporadic at best.
Though, it wouldn't really surprise me if they were being run and finding issues all the time, but they never actually got back to us. This certainly wouldn't be the first time we ran into "difficulties" due to the nature of the relationship and culture.
Which tool specifically are you thinking of that might have found this but wasn't run because of it's very high licensing fees? I work in this field, I'll be familiar with it.
Question was about high licensing fees and which tools I was referring to
I’m not claiming Defensics or OpenText DAST tools are magical “find all kernel vulns” buttons
My point is more that mature fuzzing ecosystems already existed before the recent AI-driven approaches. Protocol fuzzers, syscall fuzzers, coverage-guided fuzzers, sanitizers, dynamic analysis, etc. have all historically found serious kernel bugs
We might just be talking past each other. My question, from upthread, is this: the heyday of AFL was over a decade ago. Every major platform company fuzzes at a scale that I think is difficult for lay practitioners to get their heads around. They contract, quarterly, soup-to-nuts assessments from competing software security companies, who get full source access and are measured against each other by the quality of their findings. They run bounty programs specifically to direct public researcher attention to these exact findings.
Why didn't "mature fuzzing ecosystems" find the vulnerabilities AI is now finding? It's a pretty big gap in the "fuzzing tools already do this" logic!
> Why didn't "mature fuzzing ecosystems" find the vulnerabilities AI is now finding? It's a pretty big gap in the "fuzzing tools already do this" logic!
Because they simply aren’t ran. That’s my entire argument
Claude et al have “skills” that are basically containers of tools. It’s not a huge leap to say that similar tools are in use within these containers or even same. We don’t know.
I work in this field, have done work for some of the vendors we're discussing, and talk daily to security engineers working on these problems at these vendors. You are wrong.
Where all of this is going? Will there be a dedicated servers running coding agents that iterate throught codebases for each company to find vulnerabilities 24/7?
It's just another tool in the belt. Someone will say that's cheaper than rewriting in safe rust or whatever. (Apple must have a bunch of 1980s code written to 1980s standards. But that is their moneymaker.)
Claude and Anthropic is mentioned, but not Mythos, I'm guessing this would mean then this was found outside of the whole Mythos thing, or would there be any reason for them not to mention it, if it was involved?
Yeah I’m honestly not sure why macOS updates seem to be so huge. Often gigabytes. Do they actually have thousands of changes, so they basically ship out new versions of almost all system libraries? Or is it that they don’t have good diffing in place? Or is it a BSD thing where you basically ship everyone at once since it’s all sort of “one version” of the base system?
> But aren't they able to do incremental builds and separated x64/arm64?
During the PowerPC to Intel transition, they did stuff like that; perhaps at their current scale, there's reasons why they don't.
Supporting both architectures enables a macOS install to boot an Intel Mac or an Apple Silicon Mac, which is useful in a dual-architecture environment.
It's easy to check for dual architecture support; just use the file command:
It's funny how in the past a server uptime used to be a kind of badge of honor, while now a computer running for more than a week is a massive security risk.
I've had to be on top of updating everything constantly lately.
when multiple independent parties are simultaneously tripping over different holes in the same kernel, that's not bad luck, that's a systemic attack surface problem
For the record, this bug has nothing to do with our recent MIE attack [1] [2], which exploited two different kernel bugs. Our bugs are not fixed yet.
[1] https://blog.calif.io/p/first-public-kernel-memory-corruptio...
[2] https://news.ycombinator.com/item?id=48139219