Welcome, humans.

Apple finally unveiled its AI rebrand, and they’re calling it Apple Intelligence. Believe it or not, that's the brainchild of a $700,000-salary Stanford alum with the title "Chief Innovation Wizard and Naming Guru" who took 18 months of ayahuasca trips after ChatGPT's debut.

We're just kidding—we actually dig the name and its slogan: “AI for the rest of us.” At their WWDC conference yesterday, they took everything we knew was possible with AI and turned it into something practical and cool for the Average Joe…

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

• Apple will integrate ChatGPT into its operating systems later this year.

• Siri is getting an AI overhaul and will be able to take actions in apps.

• You’ll be able to type to the new Siri, not just talk to it.

• Mail will have new AI tools that can help write/summarize emails.

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

Apple Intelligence is here.

The world has (im)patiently awaited Apple’s AI plan for over 18 months, and it’s finally here.

Apple Intelligence is a new suite of actually useful AI tools that will be coming to Apple devices later this year.

The biggest AI upgrade is, yup, you guessed it: Siri just got much more useful. As a result, we’re ditching the old nickname, Sir-what-did-you-say…

First off, iOS 18 will indeed integrate with ChatGPT later this year, transforming Siri from a hard-of-hearing assistant to a know-it-all conversationalist. This = more natural convos with Siri (and you can now type to it too).

Siri will continue to locate info across your apps, like “Where is tonight’s dinner reservation?” or “What’s the address for my doctor’s appointment?”

But here’s the real kicker: Siri can now act on your behalf in your apps. This is new. Like very new.

This is what tech like Humane and Rabbit have been chasing: allowing an AI to see everything on your iPhone (dubbed “onscreen awareness”) so it can handle tasks you’d normally tap around for. Some are calling this a LAM—Large Action Model.

For example…

• You can have Siri pull up a photo of your driver’s license, snag the number automatically, and paste it into a form.

• Snap a photo of whatever’s on your screen and send it off to a friend.

• If a buddy texts you their new address, just say, “Add this address to their contact card,” and Siri will take care of the rest.

Beyond Siri, another juicy update is an AI writing assistant (think Gemini or Copilot) that can:

• Summarize unread emails or long group chats.

• Elevate time-sensitive messages to the top of your inbox.

• Draft or rewrite email responses.

• Transcribe and summarize phone calls.

Important: All these AI tools will only be available on iPhone 15 Pro, iPhone 15 Pro Max, and iPads and Macs with M1 chips or later, meaning you’ll need to shell out $999 to get a taste of the new Apple.

Why it matters: We’re impressed by Apple. Turns out, slow and steady > move fast and break things when you’re a $3T company (looking at you, Google).

The real deal here is that Apple is literally putting AI into the hands of >1B people, most of whom will probably be using AI for the 1st time. And it’s delivering AI that’s actually useful (forget those Genmojis, we’re talking about implanting ChatGPT-4o’s brain into Apple devices).

Catch Pete’s extended analysis on why this matters and what it means for you in today’s podcast episode (Apple Podcasts, Spotify, YouTube).

FROM OUR PARTNERS

This AI tool delivers 5x superior results when handling your sensitive data. 

Here’s an insider tip for businesses using open-source models like Llama 3 and Mixtral to classify, summarize, and augment sensitive data.

What doesn’t work is a DIY approach (do it yourself) using assorted optimization techniques. 

What does work is using OctoStack from OctoAI—take your data to the next level with significantly quicker processing and responses.

Faster responses + fewer GPUs = more $$$!

Want to stay ahead of the game?

Check out OctoAI’s blog to discover how you can achieve 5x or better performance gains compared to vLLM!

Treats To Try.

1. *Get high-quality human data to make your AI models more effective. Instantly connect with 150k+ participants and domain specialists. Start with Prolific here!

2. Jump is an AI tool for wealth advisors that converts convos into tasks, notes, and compliance records (raised $4.6M).

3. Tomato.ai adjusts the accents of offshore agent voices to reduce churn (raised >$2.1).

4. Portal.ai uses AI to draw insights from an SMB’s data from Shopify, Amazon, Google, and Meta (raised $5M).

5. Writer’s new AI Studio lets companies build no-code AI apps connected to their data.

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

Tuesday Ticker.

Here are the results from last week’s poll:

• S.H. chose Less: “I use it for cross-checking AI's answers but since Wikipedia style sometimes is hard to read and varies a lot in scope, AI (rather Perplexity than ChatGPT) has become my first choice. Perplexity usually does refer to Wikipedia content if helpful.”

• V.S. chose More: “It seems that people want "summaries" more and more. I like them occasionally, but the coolest, most useful or differentiator information might be in the details left out.”

• Y. chose The same: “I haven't really noticed a difference - I still go to the Wikipedia often to fact check ChatGPT. Not so much with Perplexity.”

We feel that ChatGPT-4o’s quality has dipped just a tad since its launch, and we even find ourselves reverting to GPT-4 in some instances.

A Cat's Commentary.