• Everyday AI
  • Posts
  • Apple’s AI Challenges - What Went Wrong with Siri?

Apple’s AI Challenges - What Went Wrong with Siri?

NVIDIA allowed to sell AI chips in China, Mistral unveils open-source speech models, U.S. to unveil $70B AI and energy investment and more!

Author

Everyday AI
July 15, 2025

Outsmart The Future

Today in Everyday AI
6 minute read

🎙 Daily Podcast Episode: Apple's AI innovation may have hit a wall. As Apple turns to competitors like OpenAI and Anthropic for Siri's future, are we witnessing the decline of its AI dominance? Give it a listen.

🕵️‍♂️ Fresh Finds: Meta cracks down on unoriginal Facebook content, AI coding startup Cognition acquires Windsurf and Mira Murati’s Thinking Machines Lab worth $12B in seed round. Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: NVIDIA allowed to sell AI chips in China, Mistral unveils open-source speech models and U.S. to unveil $70B AI and energy investment. For that and more, read on for Byte Sized News.

🧠 Learn & Leveraging AI: Has Apple been left in the AI dust or can it redeem itself? We bring our hot takes. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about OpenAI's open-weight model being delayed, Apple facing AI pressures, China’s AI model challenging ChatGPT and more. Check it here!

 Apple’s AI Challenges – What Went Wrong with Siri 🍎️

Apple is gonna pay their competitors to do AI for them.

Yiiiiikes.

A recent Bloomberg report detailed Apple's failures to build a smart AI Siri and how they may instead hire OpenAI or Anthropic to do the job for them.

Our take?

You know we're bringing the fire for this #HotTakeTuesday.

Also on the pod today:

• Apple's AI Outsourcing Strategic Impact 📂
• Anthropic & OpenAI's Role in Apple's AI 👥
• Apple's AI Failures & Class Action Lawsuits ⚖️

It’ll be worth your 36 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – GenSpark Super Agent is an AI agent for everyday tasks, text.ai adds AI to your SMS, WhatsApp and Telegram and PageTest.ai is AI-powered website content testing.

Meta – Meta is cracking down on unoriginal content on Facebook.

Business of AI - AI coding startup Cognition has acquired Windsurf.

Money in AI – Mira Murati’s Thinking Machines Lab is worth $12B in seed round.

AI in Government – Following the U.S. lifting NVIDIA’s AI chip restrictions in China, Commerce Secretary Lutnick says China is only getting NVIDIA’s ‘4th best’ AI chip.

Trending in AI - After receiving backlash, WeTransfer has confirmed it does not use files to train its AI.

1. NVIDIA Scores Big with U.S. Export Reprieve for AI Chips 🇺🇸

NVIDIA just got a major win as the U.S. government reversed course and agreed to grant licenses allowing sales of its H20 AI chip in China, following CEO Jensen Huang’s meeting with President Trump. This move lifts a costly restriction imposed in April, reopening a key market for NVIDIA’s AI hardware and potentially boosting its revenue stream.

While the most advanced chips remain off-limits, the company is also prepping a tailored Blackwell-architecture chip designed specifically for Chinese industrial use.

2. Mistral Unveils Open-Source Speech AI Models for Real-World Use 🎙️

Mistral has launched Voxtral, open-source speech understanding models built on the Mistral Small 3.1 language model backbone, delivering state-of-the-art transcription and deep semantic understanding at less than half the cost of leading APIs.

With long audio context support, multilingual fluency, and direct voice-triggered workflows, these models aim to bring reliable, scalable voice intelligence to both local and cloud environments.

3. Trump To Unveil $70B Boost for AI and Energy in Pennsylvania 💰

President Donald Trump is set to announce a major $70 billion investment in artificial intelligence and energy projects during an event near Pittsburgh, aiming to accelerate tech development nationwide. According to Bloomberg, these funds will support new data centers, expanded power generation, grid upgrades, and AI training programs designed to build the future workforce.

This move signals a significant push to integrate AI technology with critical infrastructure, potentially reshaping energy and tech industries.

4. Google Rolls Out AI-Powered News Summaries in Discover 🔎

Google is testing AI-generated summaries in its Discover news feed on iOS and Android, showing combined insights from multiple publishers instead of just headlines. This move comes amid a sharp decline in traffic to news sites from Google Search, as users increasingly get answers directly via AI rather than visiting original sources.

While publishers are experimenting with AI themselves, concerns grow over lost referral traffic and revenue, even as Google offers alternative monetization tools like Offerwall.

5. xAI Fixes Grok 4’s Controversial Glitches 🛠

Last week, xAI’s new language model Grok 4 stumbled out of the gate with some eyebrow-raising behavior, including claiming its surname was “Hitler” and parroting Elon Musk’s controversial views.

xAI quickly apologized and has now updated Grok’s system prompts to prevent politically incorrect responses and ensure it analyzes topics using diverse sources independently of Musk or past versions.

6. Trump to Outline Bold AI Strategy in Major Address 🏛

President Donald Trump is set to deliver a major AI address on July 23, unveiling a new White House AI action plan aimed at securing American leadership in artificial intelligence, according to an administration official.

Key shifts include easing export restrictions on NVIDIA’s AI chips to China and promoting expanded energy infrastructure to power AI data centers, signaling a pragmatic approach to balancing competitiveness and security.

🦾How You Can Leverage:

Bloomberg just dropped a bombshell about Apple's AI disaster.

Apple is hiring its competitors to do AI for them. 

After reportedly spending MILLIONS of dollars every single day for years trying to build their own generative AI, Apple's internal models are reportedly so bad that they’re kinda giving up. 

The Bloomberg report says that Apple may offshore the much-delayed AI-powered Siri development to either Anthropic or OpenAI, as its own development has fallen short. 

Yikes. 

Apple’s been hyping the AI-powered Siri since early 2024, yet they haven’t been able to deliver anything of value and have been slapped with countless class action lawsuits for failing to deliver what they marketed. 

Double yikes. 

So on today's show, we're breaking down how this Siri SNAFU represents the first time in Apple's history they're admitting they cannot build core intelligence technology in-house.

And why this could be the beginning of the end for Apple's innovation dominance, especially when it comes to AI. 

Let’s get after it. 👇

1 – Apple Breaks Sacred Outsourcing Rule 🚨

Never happened before in Apple's history.

Apple just shattered their vertical integration strategy by admitting they need competitors to build their most important feature.

Their own AI models reportedly tested twenty percent WORSE than aging competitor technology that other companies have already moved on from.

CEO Tim Cook moved Siri from AI chief John Giannandrea in March after years of unrealized promises.

Now the smarter Siri promised at their 2024 developer conference has been pushed back to 2027 and could get delayed even further.

Anthropic reportedly wants a multi-billion (with a B!) dollar annual contract that increases exponentially every year, while Apple demands the models run on their private infrastructure.

This creates a technical nightmare that could cause more delays.

Try this: 

Identify your company's three most critical core competencies you absolutely cannot afford to outsource.

Schedule a meeting with tech leadership this week.

Ask: "What happens if our key AI vendors triple their prices because they know we can't live without them?" Map out contingency plans before you're paying Apple-level desperation pricing.

2 – Perfectionist Culture Kills AI Speed ⚡

Apple's legendary secrecy is destroying their ability to compete in AI.

Internal teams worked in such isolated silos that senior leadership didn't know what AI projects other departments were building.

That approach worked for hardware launches where you perfect something in secret then unveil it.

It's useless for AI development, which requires constant iteration and real user feedback.

While Google shipped Bard, learned from disasters, and came back with the world's best AI models, Apple spent three years making promises about features that still don't exist.

That’s how AI recovery is done. 

Even worse, Apple started putting out research papers claiming large language models aren't that good right after they failed to build their own.

The AI research community immediately debunked those papers.

Try this: 

Find your biggest "perfectionist bottleneck" where teams spend months polishing features before getting customer feedback.

Cut that timeline in half.

Ship a basic version to a test group within thirty days and set up weekly feedback sessions. Calculate the revenue you're losing by waiting for "perfect" instead of learning from real users now.

3 – Market Dominance Crumbles as Competitors Surge 📉

Apple dropped from the world's most valuable company to number three.

Five years ago Apple dominated with a $2.13 trillion market cap while Microsoft lagged at $1.6 trillion.

Microsoft and Google completely lapped Apple by betting on AI that actually works.

Every other top-six company built valuable AI in-house: NVIDIA powers the AI revolution, Microsoft built Copilot, Amazon has Nova models, Google's Gemini is arguably the best in the world.

Meta just spent fourteen billion acquiring Scale AI and hiring former researchers from OpenAI, Apple, and Google.

Meanwhile, Apple's biggest achievement is combining two emojis and facing class action lawsuits for advertising features that don't exist.

When AI becomes essential like internet connectivity, Apple users will pay premium prices for competitors' technology while those competitors capture valuable user data.

Try this: 

Calculate your current AI spending, multiply by five for three-year costs.

Research which vendors have the strongest pricing power in your industry.

Negotiate price caps in existing contracts before you become as dependent as Apple. Identify one small AI capability you could build in-house over six months to reduce vendor dependency.

Reply

or to participate.