😺 The truth about Apple Intelligence

Welcome, humans.

Last month, Google's AI was spotted suggesting some wild ideas, like slathering pizza with glue. It’s just a unique culinary journey—you wouldn’t get it!

Google promised to do better, but merely three weeks later, its AI is once again proposing we sprinkle 1/8 cup of Elmer’s glue into our tomato sauce!

The kicker? It references a Business Insider piece that initially reported on the fact that Google suggested glue on pizza. Is this the way AI ends us all? By force-feeding us glue on DIY pizza night?!

Here’s what you need to know about AI today:

• Elon Musk threatened to ban Apple devices from his companies.
• We're breaking down how Siri with AI really works.
• Elon Musk withdrew his lawsuit against OpenAI.
• AI startups snagged 40% of all VC funds in May.

On the podcast: Apple Intelligence, what it all means, and the real winners and losers from this announcement (Apple Podcasts, Spotify, YouTube).

Debunking myths re: Apple Intelligence…

Yesterday, Apple did an impressive job demonstrating the real-world utility of its new AI tools.

Still, there was a lot of confusion on two fronts:

1. When is this stuff coming?
2. How the heck does it work (specifically, is it really private)?

The frustrating thing about Apple Intelligence is that no one can use it yet—you likely won't see (A)iOS 18 until September, and it'll require an iPhone 15 Pro, iPhone 15 Pro Max, or newer iPads/Macs with M1 chips.

To its credit, Apple is investing heavily in ensuring all this AI stuff stays private, meaning the AI keeps what it learns about you from your apps strictly between you and your device.

So don’t listen to Elon Musk, who’s spreading significant FUD (Fear, Uncertainty, Doubt) by threatening to prohibit Apple devices at his companies.

BTW, getting duped by the Community Notes on the platform you own is like a magician being fooled by their own magic trick: embarrassing.

Still, Apple could’ve done a better job explaining how all this really works behind the scenes. Here’s what will really go down when you ask Siri with AI a question:

1. For almost all questions, Siri uses AI that lives on the device, aka it won’t need to hit up the cloud or ChatGPT, aka your question won’t ever leave the phone.
   1. These on-device models are decent (they’re built on top of open-source models) and outperform Google’s on-device model, Gemma-7B, 70% of the time.
2. For more complex questions like “Find the photo I took at the beach last summer,” Siri will consult a smarter AI model that runs on Apple’s servers.
   1. When Siri sends your question to Apple’s servers, your data is anonymized and not stored there forever.
3. Now, for longer questions like “Can you help me create a weekly meal plan?” or "Rewrite this email using a more casual tone,” Siri will use ChatGPT only if you give it permission to.
   1. Even if you opt for "Use ChatGPT," OpenAI isn’t allowed to store your data.

Apple Intelligence

Minute 14:48 of Marques Brownlee’s Apple Intelligence video gives a good summary of these points.

Why it matters: Even though using ChatGPT will stay private on iOS, it's not ideal to have to approve using ChatGPT every time you ask a complex question.

We suspect Apple is developing a GPT-4o-level AI model to embed directly into its devices so it doesn’t need any third-party AI…

FROM OUR PARTNERS

Hire a world-class AI team trusted by Fortune 500 companies and leading startups.

Building AI software in your company is hard. You’ve got three options:

1. Track down technical engineers and pay them a fortune.
2. Try to DIY, but building AI products is highly technical.
3. Use AE Studio.

Their stacked team of AI experts builds the perfect AI solution for your business. That’s why they’re the go-to for leading startups and Fortune 500 companies.

Their development, design, and data science teams work closely with your team to make sure your AI does exactly what you need it to.

Book a FREE consultation today!

Around the Horn.

• Elon Musk dropped his lawsuit against OpenAI in California.
• Brazil’s government is paying OpenAI to speed up the screening of thousands of lawsuits to reduce costs.
• Microsoft is ditching its GPT Builder feature in Copilot Pro, which previously let users create custom versions of ChatGPT.
• 40% of all global venture capital funding in May went to AI companies.

Treats To Try.

1. *The AI Conference brings together the brightest minds in AI. Get $200 off your ticket for the two-day event in SF on Sept. 10-11 using code “Neuron24”.
2. Mistral builds powerful open-source models and just raised a giant $640M funding round at a $6B valuation.
3. Cognigy develops AI agents to automate customer support (raised $100M).
4. Zeliq uses AI to craft personalized outreach for securing qualified sales leads (raised $15.4M).
5. Particle is a news-reading app that offers different angles on the same stories (raised $10.9M).
6. Brightwave is an AI research tool that generates financial analysis on any topic (raised $6M).

*This is sponsored content. Advertise in The Neuron here.

A Cat's Commentary.