• Everyday AI
  • Posts
  • Google’s $1 Trillion AI Mistake: 5 Gemini missteps and if Google can recover

Google’s $1 Trillion AI Mistake: 5 Gemini missteps and if Google can recover

Google AI 2024 recap, xAI releases Grok Web, NVIDIA and Microsoft look to lead 2025, Alibaba slashes AI prices to compete and more!

Author

Everyday AI
December 31, 2024

Outsmart The Future

Today in Everyday AI
7 minute read

🎙 Daily Podcast Episode: Did Google actually...lose the AI wars in 2024? We break down Google's $1 trillion AI mistake and how they can recover. Give it a listen.

🕵️‍♂️ Fresh Finds: Hugging Face releases new agentic system, Galaxy S25 might include a Gemini subscription and did Gemini Advanced get a secret boost? Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: xAI releases Grok Web, NVIDIA and Microsoft look to lead 2025 and Alibaba slashes AI prices to compete. For that and more, read on for Byte Sized News.

🚀 AI In 5: NotebookLM’s new interactive feature is being slept on! We explain why it’s so good! See it here

🧠 Learn & Leveraging AI: Is Google an AI winner or loser in 2024? We break down their recent AI moves and where they stand among competitors. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about NVIDIA acquiring Run:ai and making it open source, OpenAI facing for-profit pushback, how ChatGPT Search is misleading users and more! Check it here!

 Google’s $1 Trillion AI Mistake: 5 Gemini missteps and if Google can recover 😬

Yeah, Google won the end of 2024 in the AI wars.

But they've dropped the bag more times than any other Big Tech or AI startup.

In our last Hot Take Tuesday of 2024, we're gonna break down the $1 trillion dollar mistake that Google made (well.... 5 of them) and examine if they can bounce back in 2025.

Join the conversation and ask Jordan questions on Google AI here.

Also on the pod today:

• Google vs. Competitors 🥊
• Google's AI Products and History 🤖
• Google's AI Strategy and Adjustments 🎯

It’ll be worth your 43 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 – Popshort turns ideas into short films, EasyLang AI helps you learn any language and Shoonya provides specialized foundation models for commerce use-cases.

Google – One of Google’s top product leaders believes a new approach could offer a quicker route to superintelligence.

Speaking of Google, we came across this while using Gemini Advanced recently… like what?!

AI Models - Hugging Face has released smolagents, a new moonshot agentic system.

Big Tech – It’s rumored that the Galaxy S25 will include a free Gemini Advanced subscription.

OpenAI – OpenAI’s new o3 model can cost more than $1,000 per query.

AI in Government – Here’s some AI laws to keep an eye on in 2025.

AI Tech – LG has unveiled four new AI hybrid laptops.

1. NVIDIA and Microsoft Could Win the 2025 AI Race 🏃

A new study from Wedbush Securities highlights Microsoft and NVIDIA as potential winners in the rapidly growing artificial intelligence landscape, with predictions of a 25% rise in tech company valuations by 2025. Analysts anticipate around $2 trillion in AI-related capital expenditures over the next three years, positioning software firms like Salesforce and Palantir to benefit significantly from large language models and generative AI advancements.

NVIDIA's processors are deemed crucial infrastructure for various industries, while Microsoft's Azure and other cloud services are set to thrive amid rising demand for AI solutions.

2. xAI Launches Grok Web 🚀

In a new development for users of Grok, xAI has started to roll out its web version, allowing easy access directly from mobile devices and browsers without needing to go through X.

This update offers users more flexibility, with chat history now syncing between the mobile app and the web, although a few hiccups remain as the feature is still in beta. As Grok expands its accessibility, it promises to enhance productivity for professionals eager to leverage AI tools in their careers.

3. Alibaba Slashes AI Prices by Up to 85% 📉

Alibaba has announced significant price cuts on its large language models, with reductions reaching as high as 85%. This decision, revealed through a WeChat post by Alibaba Cloud, highlights the fierce rivalry among Chinese tech giants like Tencent and Baidu to dominate the burgeoning AI market.

While Alibaba’s share price saw minimal movement, the company’s strategy of reducing costs—following earlier cuts of up to 97%—aims to entice over 90,000 enterprise users to adopt its innovative Qwen models.

4. ByteDance's Bold Move to Acquire NVIDIA Chips 🇨🇳

In a striking maneuver amidst U.S. restrictions, ByteDance is reportedly set to invest $7 billion in NVIDIA chips by 2025, positioning itself as a major player in the AI hardware space. According to The Information, the Chinese tech giant plans to bypass direct import limitations by storing the chips in data centers outside China, such as in Southeast Asia.

This strategy not only showcases ByteDance's resilience but also indicates a growing tension in the tech landscape as companies seek cutting-edge resources despite geopolitical hurdles.

5. Quantum AI Project Faces Bumpy Takeoff 😥

In a recent development, Eric Schmidt’s ambitious Quantum AI initiative is experiencing some turbulence right out of the gate. According to insiders, challenges in aligning the project’s goals with technological expectations have led to delays and uncertainties about its long-term viability.

This ambitious endeavor aims to revolutionize AI's capabilities, but the rocky start raises questions about its potential impact in the AI space.

NotebookLM 's Join Interactive Mode. The single BEST AI feature ever?

This one new feature inside NotebookLM is getting slept on!

We think this is one of the most IMPORTANT features in AI history.

Yep. You read that right.

Why? We dive in and show you.

🦾How You Can Leverage:

Yeah, Google crushed December 2024. 

Clap for em. 👏

But the previous two years BEFORE that? 

Straight up disastrous shorties. 

We think it could eventually be a main case study in business school in how NOT to bring a product to market. 

On today’s show, we detailed 5 of Google’s biggest missteps when it comes to its Gemini rollout, how we think it’ll cost them billions in market cap in the long run, and if they can recover. 

Ready for our last spicy #HotTakeTuesday of 2024? 

Yee-friggin,haw. 

1 – The Fabricated Demo That Shattered Trust 💔

The AI world stopped cold on that December 2023 afternoon. 

Google's Gemini launch video showed their AI playing rock-paper-scissors in real-time. Engineers were stunned. Competitors were shook. Too bad it was complete fiction.

Google staged the entire thing. Those "real-time" interactions? Just static image prompts stitched together with actors reading scripts. The smoking gun came from Google's own blog post confession.

What it means: 

Enterprise clients today STILL bring up this video during implementation meetings. The damage runs deep.

Google had to unlist their own announcement video - a marketing asset that cost millions to produce. Fortune 500 CTOs literally reference this moment as their reason for choosing other AI vendors.

Here's the brutal math: One staged demo + one forced confession = years of lost enterprise trust. In the AI arms race, trust is the only currency that matters.

2 – The Workspace Identity Crisis 😵‍💫

Pull up gemini.google.com on your personal Gmail. (Assuming you have the paid plan.)

Model selector? Check. Version info? Clear as day. 

Now try your Google Workspace account. (Paid) 

Different universe. Zero info about what you're using. Just a blank canvas of confusion.

The kicker? Business customers dropping $20 monthly per seat STILL can't tell which version of Gemini they're using. 

We tested this across multiple enterprise accounts. The results? Pure chaos.

What it means: 

While ChatGPT and Claude users get crystal-clear model information, Google created an enterprise-grade mystery box. 

Actual Fortune 100 CEOs are testing these platforms themselves before green-lighting million-dollar implementations.

When your C-suite can't even tell which model they're evaluating? Game over. They're calling Microsoft before lunch.

Remember that trust is currency, Google. You’re losing it. 

3 – When Google’s AI Couldn’t… Google 🤦

For MONTHS after launch, Google's Gemini couldn't access... Google. The search giant's AI chatbot couldn't use the world's largest search engine. The same search engine THEY OWNED.

Videos from early 2024 show Gemini failing to answer questions about events from three months prior. Meanwhile, ChatGPT had Bing. Claude had its databases. Even Meta's Llama could Google.

What it means: 

Enterprise teams needing real-time data analysis and market intelligence blacklisted Gemini immediately. The math was simple: No real-time data = no enterprise adoption.

Google's competitors turned this into their best sales pitch: "Want an AI that can actually search the internet? Don't ask Google."

4 – The Non-Functional Integration Epidemic 🏚

For 3-9 months straight, Google plastered Gemini across Gmail, Docs, Sheets, and Slides. Premium real estate for an AI that couldn't... do anything.

Ask it to analyze your spreadsheet? "Can't help with that." Summarize an email thread? "Not possible." Draft a document? Silence. We've got the receipts - videos showing Gemini failing basic tasks across every Google Workspace app.

What it means: While Microsoft's Copilot was revolutionizing office workflows, Google's AI played digital paperweight. Actual enterprise examples show teams testing basic email summaries, getting zero value, and immediately jumping to competitors.

The damage? Thousands of companies made their 2024 AI investment decisions based on these non-functional integrations. Those multi-year contracts aren't coming back.

It was hilariously bad. 

5 – The Hidden Model Fiasco 🫣

Until December 2024, Google's best AI models never touched the Gemini front end at gemini.google.com.

While they crushed benchmarks in AI testing arenas and were readily available for devs inside Google’s AI Studio, regular users who didn’t know better got stuck with models (presumably) from May 2023.

The real power? Buried in Google AI Studio and Vertex - platforms most business users never knew existed. Even VPs had to hit up Twitter just to figure out which model they were actually using.

What it means: 

Non-technical executives making million-dollar AI decisions never saw Google's A-game. We're talking Fortune 500 companies choosing AI platforms based on outdated models while Google's best tech collected dust in developer tools.

Real talk: Multiple Fortune 100 companies made their entire 2024 AI strategy based on testing front end Gemini 

They never knew Google's latest models even existed.

6 – The December Redemption Arc 💪

December 2024: Google finally woke up. 

No flashy videos. 

No staged demos. 

Just pure product heat: Gemini 2.0 experimental models crushing every benchmark. Deep research features making Perplexity look dated. NotebookLM redefining AI analysis. And Veo 2 has OpenAI's Sora looking over its shoulder.

The sleeping giant remembered how to throw hands. 

But enterprise AI contracts run long. Those multi-million dollar deals with Microsoft or OpenAI? They're not up for renewal until 2025 or 2026 or even later. 

Here's the trillion-dollar question: Will companies who've already committed millions (in money and time) to competitors give Google another shot? 

We might be watching either tech's greatest comeback story or its most expensive "what if."

⌚

Numbers to watch

428%

The C-Suite saw a 428% increase in AI-related roles over the past two years.

Now This …

Let us know your thoughts!

Vote to see live results

How many AI newsletters do you read?

Login or Subscribe to participate in polls.

Reply

or to participate.