Good Morning! Apple's in hot water as developers call working with the company "abusive," citing payment delays and communication breakdowns. Meanwhile, OpenAI is spicing things up with an advanced Voice Mode for ChatGPT Plus, promising real-time AI conversations with emotion detection. In a bit of cybersecurity drama, a new Mandrake spyware variant managed to dodge Google Play's defenses for two years, reminding us that even official app stores aren't bulletproof.

— Forrest Knight & Meghanadh Vasireddy

Apple Employees Are Calling Working at the Company “Abusive”

Context: Apple Arcade, once a promising haven for mobile game devs, is facing some serious heat. Initially lucrative, the service has reportedly been tightening its purse strings, leaving smaller studios in a financial pinch.

What's New:

1. Payment Delays: Some devs report waiting up to six months for payments. Ouch!

2. Communication Breakdown: Developers cite weeks of radio silence and glacial email response times from Apple.

3. Vision Pro Woes: The new headset is apparently struggling with complex games, and Apple's engineers seem just as confused as the devs.

Apparently the QA process is so frustrating that some devs are avoiding updates altogether. And don't get them started on the tech support – it's been called "miserable" and the "worst they'd seen anywhere."

Also, Apple expects indie devs to create Vision Pro games without compensation or marketing promises. Talk about a tough sell!

While some developers acknowledge Apple Arcade's past benefits, others are less charitable. One even likened working with Apple to an "abusive relationship."

As mobile gaming continues to evolve, it'll be interesting to see how Apple addresses these concerns. Will they level up their dev support, or is this game over for Apple Arcade?

Read More Here

OpenAI Kicks Off Advanced Voice Mode for ChatGPT Plus

OpenAI is shaking things up again with the alpha rollout of its advanced Voice Mode for ChatGPT Plus users. This new feature promises real-time conversations with AI-generated voices, the ability to interrupt and redirect the chat on the fly, and even emotion detection and response capabilities.

On the technical side, OpenAI didn't cut any corners. They put the system through rigorous testing with over 100 external red teamers across 45 languages. Safety was a top priority, with beefed-up content filters and improved model capabilities to detect and refuse certain requests. They've also added new filters to block requests for generating music or copyrighted audio.

Voice Variety:

• Four preset AI voices available (down from the original five)

• Custom voice impersonation strictly prohibited

The road to release has been interesting. After the initial demo at the GPT-4 launch event in May 2024, OpenAI delayed the planned June alpha release to fine-tune the system. Now, in July, we're seeing the start of the limited alpha rollout to select ChatGPT Plus users.

What's next?: OpenAI plans to make advanced Voice Mode available to all Plus subscribers by fall 2024.

Read More Here

Mandrake Spyware Variant Evades Google Play Security

Context: Remember Mandrake, that spyware that's been around since 2016? Well, it's back and worse than ever. Kaspersky's sharp-eyed researchers just uncovered a new variant that's been chilling on Google Play for two whole years.

This latest Mandrake iteration is like a chameleon on steroids. It has some seriously advanced obfuscation techniques:

1. Shifting malicious functions to obfuscated native libraries (hello, OLLVM!)

2. Using certificate pinning for C2 communication

3. Extensive checks for rooted devices or emulators

Five seemingly legit apps (think Wi-Fi file sharing, astronomy, and even a Genshin Impact helper) managed to rack up over 32,000 downloads before getting the boot.

This campaign hit users across the globe, with Canada, Germany, Italy, Mexico, Spain, Peru, and the UK taking the brunt of it. As of July 2024, these apps were still flying under the radar on VirusTotal.

The Takeaway: This Mandrake saga is a reminder that even official app stores aren't impenetrable fortresses. As security measures tighten, malware authors are upping their game.

Read More Here

🔥 More Notes

• Websites are Blocking the Wrong AI Scrapers (Because AI Companies Keep Making New Ones)

• AI diary: Testing Apple Intelligence Writing Tools, and finding myself impressed

• Taco Bell will add voice AI ordering to hundreds of drive-thrus this year

Youtube Spotlight

How AI could help us talk to animals

Was this forwarded to you? Sign Up Here