Rayan

tracking AI news, model releases & breakthroughs worth knowing

[Rayan @RayanSMX]

🚨 Attention: A small security firm just got root on Apple’s newest iPhone. And it survived a reboot.

https://t.co/EYCTWi5Ezp did it live at Hexacon in Paris on October 16. iPhone 18 Pro. iOS 27.0.1. Apple’s current release.

It is not a leaked government tool. It is not a spyware vendor. It is a team that says it sells no bugs and no exploits.

They rebooted the phone, unlocked it, and two minutes later were logged in over SSH as root. No computer attached.

Apple pays up to $2,000,000 for a spyware-grade exploit chain. With bonuses, the ceiling passes $5,000,000. Apple has paid $35,000,000 to 800+ researchers since 2020.

Calif reported nearly 100 bugs to Apple this year.

Now the AI part. Apple’s iOS 27 security notes credit https://t.co/EYCTWi5Ezp together with Claude and Anthropic Research for two bugs. In May, Calif said Claude Mythos Preview helped build a macOS kernel exploit against Apple’s M5 protections in under a week.

Nobody has confirmed AI built this exact jailbreak. But Apple itself says chains like this cost millions to develop.

iOS 27.2 shipped October 26 and fixed most of the bugs. Calif never named them. No public tool exists, and Calif plans no end-user release. Update anyway.

The demo video is public and free. Most people will scroll past it.

[Rayan @RayanSMX]

@AIronAgent everyone is fighting over video models and meanwhile a browser tab does the job

[Rayan @RayanSMX]

🚩Warning. One AI prompt just turned an old Pixel into a viral horror story. 214,856 files.

A user named Cooler says Claude Code found them in DCIM/.thumbnails on his Android. Photos, videos and screenshots he says he had permanently deleted. A passport photo, bank screenshots and a crypto seed phrase, he claims.

It is not a hack. It is not malware. It is a hidden folder of previews that Android keeps so galleries load fast.

Now the part the viral version skips.

A preview is not a recovered original. The count mixes duplicates, database files and cached content. Cyber Security News calls the claim unverified.

“Any app can read it” is too broad. Apps targeting Android 11 or higher run under scoped storage. Android 14 added selected photo access. Even the special all-files permission does not unlock other apps’ private folders.

The real story is the AI. One prompt to an agent did the digging. In the same week, Apple’s iOS 27.2 notes credited Anthropic Research on three fixes and OpenAI on four. AI is becoming the auditor of your own phone.

Check yours in two minutes. Settings. Apps. Permissions. Photos and All files access. Remove what a flashlight app has no reason to ask for.

The clip is below. Almost nobody audits their own permissions.

[Rayan @RayanSMX]

🚨Attention. Apple spent 5 years building a chip shield. A team using Claude broke it in 5 days.

It is not a leaked government tool. It is not a spyware vendor. It is a small firm called Calif.

May 14: Calif used Claude Mythos Preview to find the bugs and build a macOS kernel exploit past Apple’s Memory Integrity Enforcement on the M5. Under a week.

That model is not public. Only vetted partners in Anthropic’s Project Glasswing can use it.

Calif’s CEO says it is excellent at known attack techniques but cannot invent new ones. The bugs fell fast because they were known classes. Humans still had to beat the shield.

October 16, Paris: an iPhone 18 Pro rebooted on stage, unlocked with a normal passcode, root shell, no computer attached. Same family of shield, bypassed again. Calif has not detailed AI’s role in that chain.

Apple pays up to $2,000,000 for a zero-click chain. Over $5,000,000 with bonuses. It has paid $35,000,000 to 800+ researchers.

iOS 27.2 landed October 26 with 163 fixes. Six credit Calif. Apple’s notes also name Anthropic Research on three fixes and OpenAI on four.

No public tool exists. But phones still on 27.0.1 or 27.1 keep the old attack surface. Settings. General. Software Update. Ten seconds.

The full breakdown is in the article. Almost nobody will read it.

[Rayan @RayanSMX]

@AI_roni_X Agreed the cache is mostly previews but permissions are something you actually control.

[Rayan @RayanSMX]

🚩Warning. One AI prompt just turned an old Pixel into a viral horror story. 214,856 files.

A user named Cooler says Claude Code found them in DCIM/.thumbnails on his Android. Photos, videos and screenshots he says he had permanently deleted. A passport photo, bank screenshots and a crypto seed phrase, he claims.

It is not a hack. It is not malware. It is a hidden folder of previews that Android keeps so galleries load fast.

Now the part the viral version skips.

A preview is not a recovered original. The count mixes duplicates, database files and cached content. Cyber Security News calls the claim unverified.

“Any app can read it” is too broad. Apps targeting Android 11 or higher run under scoped storage. Android 14 added selected photo access. Even the special all-files permission does not unlock other apps’ private folders.

The real story is the AI. One prompt to an agent did the digging. In the same week, Apple’s iOS 27.2 notes credited Anthropic Research on three fixes and OpenAI on four. AI is becoming the auditor of your own phone.

Check yours in two minutes. Settings. Apps. Permissions. Photos and All files access. Remove what a flashlight app has no reason to ask for.

The clip is below. Almost nobody audits their own permissions.

[Rayan @RayanSMX]

@Thetechguy62893 at least be careful with links and downloads on that version

[Rayan @RayanSMX]

🚨Attention. One small team jailbroke two new iPhones in 7 days. Check your phone now.

It is not a rumor. It is not a leaked tool. It is a team with no public tool and no plan to release one.

October 16: iPhone 18 Pro, iOS 27.0.1, a jailbreak that survived a reboot. October 23: the new iPhone Duo, iOS 27.1, jailbroken on launch day.

The 18 Pro chain got past Apple’s Memory Integrity Enforcement, the hardware shield built against spyware. Calif has published no technical details.

Android is not safe either. In August, the same team showed an app with zero permissions getting root on a Galaxy S26, a Xiaomi 17 and an Oppo Find X9 Ultra. Verified boot still said green.

Apple pays up to $2,000,000 for a zero-click chain. Over $5,000,000 with bonuses. It has paid $35,000,000 to 800+ researchers.

iOS 27.2 landed October 26 with 163 fixes. Six credit Calif, and most bugs in its chain are now fixed. Anthropic Research is credited on three, OpenAI on four. The patch count keeps climbing as AI gets better at finding bugs.

Apple says none of the 163 was known to be exploited. Phones still on 27.0.1 or 27.1 keep the old attack surface. Settings. General. Software Update. It takes ten seconds.

The raw device log is below. Almost nobody will ever read one.

[Rayan @RayanSMX]

🚨 Attention: A small security firm just got root on Apple’s newest iPhone. And it survived a reboot.

https://t.co/EYCTWi5Ezp did it live at Hexacon in Paris on October 16. iPhone 18 Pro. iOS 27.0.1. Apple’s current release.

It is not a leaked government tool. It is not a spyware vendor. It is a team that says it sells no bugs and no exploits.

They rebooted the phone, unlocked it, and two minutes later were logged in over SSH as root. No computer attached.

Apple pays up to $2,000,000 for a spyware-grade exploit chain. With bonuses, the ceiling passes $5,000,000. Apple has paid $35,000,000 to 800+ researchers since 2020.

Calif reported nearly 100 bugs to Apple this year.

Now the AI part. Apple’s iOS 27 security notes credit https://t.co/EYCTWi5Ezp together with Claude and Anthropic Research for two bugs. In May, Calif said Claude Mythos Preview helped build a macOS kernel exploit against Apple’s M5 protections in under a week.

Nobody has confirmed AI built this exact jailbreak. But Apple itself says chains like this cost millions to develop.

iOS 27.2 shipped October 26 and fixed most of the bugs. Calif never named them. No public tool exists, and Calif plans no end-user release. Update anyway.

The demo video is public and free. Most people will scroll past it.

[Rayan @RayanSMX]

@SamuelSull67332 26.6.2 is not the version calif showed so this one isnt for you

[Rayan @RayanSMX]

@rajnish_builds you might be waiting a long time calif said they wont release it

[Rayan @RayanSMX]

@moabbikr It was shown live on stage at hexacon and apple patched the bugs afterward so its not fake

[Rayan @RayanSMX]

@r1VeN2k Same here and its a quiet sign that ai labs are now doing security research not just building models.

[Rayan @RayanSMX]

@AI_roni_X the scary part isnt the ai its how long unpatched phones stay out there

[Rayan @RayanSMX]

🚨Attention. Apple spent 5 years building a chip shield. A team using Claude broke it in 5 days.

It is not a leaked government tool. It is not a spyware vendor. It is a small firm called Calif.

May 14: Calif used Claude Mythos Preview to find the bugs and build a macOS kernel exploit past Apple’s Memory Integrity Enforcement on the M5. Under a week.

That model is not public. Only vetted partners in Anthropic’s Project Glasswing can use it.

Calif’s CEO says it is excellent at known attack techniques but cannot invent new ones. The bugs fell fast because they were known classes. Humans still had to beat the shield.

October 16, Paris: an iPhone 18 Pro rebooted on stage, unlocked with a normal passcode, root shell, no computer attached. Same family of shield, bypassed again. Calif has not detailed AI’s role in that chain.

Apple pays up to $2,000,000 for a zero-click chain. Over $5,000,000 with bonuses. It has paid $35,000,000 to 800+ researchers.

iOS 27.2 landed October 26 with 163 fixes. Six credit Calif. Apple’s notes also name Anthropic Research on three fixes and OpenAI on four.

No public tool exists. But phones still on 27.0.1 or 27.1 keep the old attack surface. Settings. General. Software Update. Ten seconds.

The full breakdown is in the article. Almost nobody will read it.

[Rayan @RayanSMX]

https://t.co/T4h3iIo5jg

[Rayan @RayanSMX]

@PawelPliku nice thought but calif said they wont release the jailbreak and the bugs are mostly patched now

[Rayan @RayanSMX]

🚨 Attention: A small security firm just got root on Apple’s newest iPhone. And it survived a reboot.

https://t.co/EYCTWi5Ezp did it live at Hexacon in Paris on October 16. iPhone 18 Pro. iOS 27.0.1. Apple’s current release.

It is not a leaked government tool. It is not a spyware vendor. It is a team that says it sells no bugs and no exploits.

They rebooted the phone, unlocked it, and two minutes later were logged in over SSH as root. No computer attached.

Apple pays up to $2,000,000 for a spyware-grade exploit chain. With bonuses, the ceiling passes $5,000,000. Apple has paid $35,000,000 to 800+ researchers since 2020.

Calif reported nearly 100 bugs to Apple this year.

Now the AI part. Apple’s iOS 27 security notes credit https://t.co/EYCTWi5Ezp together with Claude and Anthropic Research for two bugs. In May, Calif said Claude Mythos Preview helped build a macOS kernel exploit against Apple’s M5 protections in under a week.

Nobody has confirmed AI built this exact jailbreak. But Apple itself says chains like this cost millions to develop.

iOS 27.2 shipped October 26 and fixed most of the bugs. Calif never named them. No public tool exists, and Calif plans no end-user release. Update anyway.

The demo video is public and free. Most people will scroll past it.

[unusual_whales @unusual_whales]

BREAKING: Anthropic could get a junk credit rating despite its planned mega-IPO, per FT

Blackstone has warned investors that Anthropic may receive a below-investment-grade credit rating after its IPO, reversing earlier expectations.

[Rayan @RayanSMX]

@CryptoNobler look at the screenshot the biggest drop on it is 2.45 percent not 45

[Rayan @RayanSMX]

@shakhzod1001 max effort burns several times more usage and the app literally warns you about it

[Rayan @RayanSMX]

@chasmmmmmmmmmmm Building a game in the style of mario is fine but building Mario with a new name is not thats the whole rule(

[Rayan @RayanSMX]

https://t.co/T4h3iIo5jg

[Rayan @RayanSMX]

🚨Attention. One small team jailbroke two new iPhones in 7 days. Check your phone now.

It is not a rumor. It is not a leaked tool. It is a team with no public tool and no plan to release one.

October 16: iPhone 18 Pro, iOS 27.0.1, a jailbreak that survived a reboot. October 23: the new iPhone Duo, iOS 27.1, jailbroken on launch day.

The 18 Pro chain got past Apple’s Memory Integrity Enforcement, the hardware shield built against spyware. Calif has published no technical details.

Android is not safe either. In August, the same team showed an app with zero permissions getting root on a Galaxy S26, a Xiaomi 17 and an Oppo Find X9 Ultra. Verified boot still said green.

Apple pays up to $2,000,000 for a zero-click chain. Over $5,000,000 with bonuses. It has paid $35,000,000 to 800+ researchers.

iOS 27.2 landed October 26 with 163 fixes. Six credit Calif, and most bugs in its chain are now fixed. Anthropic Research is credited on three, OpenAI on four. The patch count keeps climbing as AI gets better at finding bugs.

Apple says none of the 163 was known to be exploited. Phones still on 27.0.1 or 27.1 keep the old attack surface. Settings. General. Software Update. It takes ten seconds.

The raw device log is below. Almost nobody will ever read one.