- Everyday AI
- Posts
- Real-Time AI Search Battle: ChatGPT Search vs. Perplexity vs. Google vs. Copilot vs. Grok
Real-Time AI Search Battle: ChatGPT Search vs. Perplexity vs. Google vs. Copilot vs. Grok
Real-time AI search LLM comparison, Amazon unveils new AI models and AI supercomputer plans, Apple taps Amazon for AI chips, Meta faces content moderation backlash and more!
Everyday AI
December 03, 2024
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Sponsored by
“We’re not being as creative as we can be when we think about AI,” says Microsoft Chief Scientist and Technical Fellow Jaime Teevan.
She joins the WorkLab podcast to discuss how leaders can unlock the full potential of AI for better brainstorming, ideation, and collaboration.
Learn more on WorkLab, available wherever you get your podcasts.
Outsmart The Future
Today in Everyday AI
6 minute read
🎙 Daily Podcast Episode: The way we search is being redefined by AI. LLMs are taking over traditional search. We break down what to expect and which LLM does it best. Give it a listen.
🕵️♂️ Fresh Finds: OpenAI considers adding advertising, Ted Cruz demands probe into EU’s AI policy and Playstation’s CEO speaks on AI. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Amazon unveils new AI models and AI supercomputer plans, Apple taps Amazon for AI chips and Meta faces content moderation backlash. For that and more, read on for Byte Sized News.
🚀 AI In 5: Is Anthropic Claude’s Projects feature better than ChatGPT’s custom GPT? We put them to the test and you might be surprised by the (unscientific) results! See it here
🧠 Learn & Leveraging AI: AI is taking over the way we search. Here’s what you need to know about the future of AI search and which LLMs perform best. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about Amazon unveiling a new AI model, OpenAI being sued by Canadian media, US tightening chip restrictions on China and more. Check it here!
Real-Time AI Search Battle - ChatGPT Search vs. Perplexity vs. Google vs. Copilot vs. Grok 🔍
AI is taking over search. Whether you love em or hate em, LLM-powered searches are coming for your devices.
↳ ChatGPT Search and its Chrome extension.
↳ Google's AI overviews.
↳ Perplexity.
How we interface with the web is quickly changing, so you need to know how each platform works and the pros and cons of each.
Join the conversation and ask Jordan questions on AI search here.
Also on the pod today:
• LLM AI Search Evaluations 🥊
• Big Tech Developments in AI search 🛠️
• Significance of AI in Search Engines 💡
It’ll be worth your 58 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Upcoming Everyday AI Livestreams
Wednesday, December 4th at 7:30 am CST ⬇️
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – Motiff is an AI-powered UI design tool, Summer is an AI summary for your blog and WebSim is an AI web editor and simulator.
OpenAI – OpenAI’s CFO mentioned that it has considered adding advertising to its products but has no current plans for it.
Amazon – Amazon has rolled out new capabilities for independent software vendors (ISVs) within its Q Business platform.
AI Governance – U.S. Senator Ted Cruz is demanding a probe into the EU’s influence on AI policy.
AI Models - A new breakthrough in detecting AI hallucinations reveals key tips for writing prompts that prevent malfunctioning.
AI in Media – Playstation’s CEO believes AI can help revolutionize gaming but will not replace human touch.
AI Tech – A new photonic processor could allow extremely fast AI calculations while using very little energy.
1. Amazon Unveils Major AI Supercomputer Plans 🦸
Amazon Web Services (AWS) has announced its ambitious plan for an "Ultracluster" supercomputer, set to be one of the largest globally for training AI models, utilizing hundreds of thousands of its proprietary Trainium chips. This initiative, called Project Rainier, is part of a broader strategy to provide alternatives to Nvidia's dominant GPUs, as AWS invests an additional $4 billion in AI startup Anthropic.
The newly unveiled Ultraserver, powered by 64 interconnected chips, aims to enhance AI capabilities while promising significant cost savings for businesses eager to adopt AI technologies.
2. AWS Unveils Nova: A Multimodal Generative AI Family 👀️
At its re:Invent conference, Amazon Web Services (AWS) launched the Nova family of multimodal generative AI models, including four text models—Micro, Lite, Pro, and Premier—alongside image and video generators, Nova Canvas and Nova Reel. With capabilities ranging from processing text to generating videos, these models promise to streamline various tasks and improve efficiency, catering to both individual users and businesses.
Notably, AWS CEO Andy Jassy highlighted the models' impressive speed and cost-effectiveness, setting a new standard in the competitive AI landscape.
3. Apple Taps Amazon for AI Chips 🍎
Apple announced that the tech giant is leveraging Amazon’s custom AI chips, including the newly available Trainium2, to enhance its machine-learning capabilities. With a claimed 40% efficiency gain from Amazon’s existing chips and potential improvements of up to 50% with the Trainium2, this partnership signals a shift toward more cost-effective AI processing alternatives—especially as Apple integrates these chips into its services like Siri and Apple Maps.
This endorsement not only highlights the reliability of AWS in the competitive cloud landscape but also suggests that companies can explore non-Nvidia options for AI training.
4. NVIDIA and AWS Unveil AI Innovations at re:Invent 2024 🚀
At AWS re:Invent 2024, NVIDIA and Amazon Web Services have launched groundbreaking solutions aimed at enhancing AI capabilities, robotics, and quantum computing development. The highlight includes the introduction of NVIDIA DGX Cloud on AWS, providing enterprises with a powerful AI training platform, while advancements in liquid-cooled data centers promise improved performance for high-density AI workloads.
Additionally, NVIDIA's expanded Isaac Sim now allows developers to simulate AI-driven robots more effectively in virtual environments, paving the way for smarter industrial automation.
5. Meta Faces Content Moderation Backlash 🫢
Meta's president of global affairs, Nick Clegg, acknowledged that the company's content moderation practices are still plagued by high error rates, leading to the unjust removal of harmless posts. He expressed regret over the aggressive content takedowns during the COVID-19 pandemic, highlighting a need for improved accuracy as users voice their frustrations.
With billions spent on moderation, Clegg's comments suggest that Meta's automated systems may be too heavy-handed, risking overreach in political speech ahead of the upcoming U.S. presidential election.
Claude Projects vs ChatGPT's custom GPTs
We recently checked out Anthropic’s new projects feature and we have to say..It’s pretty dang good!
The feature essentially allows you to train your chats.
So we were curious to see, how does this new Projects feature stack up against ChatGPT’s GPTs?
We put the two to the test.
Check out today's AI in 5.
🦾How You Can Leverage:
Nothing like a good AI fight to get the Holiday season off on the right foot, ey?
(Keep the tokens, ya filthy animal!)
In the past few weeks, we’ve seen a huge tilt from Big Tech and Big AI to search.
Like no one in tech got any Thanksgiving dessert, suddenly everyone wants a slice of Google’s search pie.
But it goes beyond the search window.
The AI giants are not just shifting how their LLMs connect to the internet in real-time, but they’re also laying the groundwork for how we interact with our devices in the future.
Whether you realize it or not — it starts with search. But the definition of search is changing.
Our hot take?
Aside from traditional SEO being dead, the lines between traditional search, LLM chatbots and “smart” voice assistants is gonna blur.
Pretty soon, they’ll be indistinguishable.
How we interact with the world around us is changing before our very noses, so we brought you a front row seat to the action today.
On today’s show, we brought you a LIVE search battle.
Which LLM is poised to win at search?
Our very unofficial findings might surprise you. Make sure you first go check out today’s episode, then let’s dive in together on the 1-2-3 of what ya need to know.
Let’s gooooooooooo.
1 – It's an all-out AI Search Sprint 🏃
Everyone wants Google's lunch money.
Errr…. Search advertising money.
And in the past few weeks, all of the big players are making moves to encroach on Google’s territory. (And even Google is making big AI search plays.)
In case you’ve been sleeping since late October, we’ve seen:
- OpenAI finally unleashed ChatGPT Search to the public after keeping it under wraps. They've been secretly running a Perplexity clone for 8 months (surprise!). Plus, they dropped a Chrome extension that lets you completely bypass Google.
- Google expanded their AI overviews to 100+ countries. And Gemini’s ability to tap into big brother Google for search …… actually works now? Weird. It’s a HUGE improvement from months ago when it couldn't really talk to Google's own search engine.
- Meta said "hold my eggnog” and announced they're building their own search index. No more relying on Bing or Google. They want full control.
- Microsoft’s Bing? They're not sleeping either. They just launched their generative search experience, adding AI-powered results right in traditional Bing search.
Even the underdogs are making moves:
- Cohere released Reranked: an AI search model that handles 100+ languages
- Perplexity launched shopping features and teased hardware plans
- The Browser Company previewed Dia, an AI-focused browser dropping in early 2024
What it means:
The battle isn't just for search — it's for your device's starting point. If you're running a business or making decisions on where to spend time and money, start testing these AI search engines NOW.
Each one has different strengths. ChatGPT crushes visual data, Meta nails accuracy, and Gemini's catching up fast. Pick your workflow tools based on real testing, not hype.
2 – The devil's in the search details 🔍
Our live testing revealed some WILD performance gaps.
And some weird idiosyncrasies.
ChatGPT Search (formerly called Search GPT) dominated with interactive maps, real-time stock graphs, and perfect weather data. When we asked for Chicago pizza spots, it didn't just list names — it showed an interactive map with every location.
Meta AI shocked everyone. It nailed the Bears schedule, current stock prices, and even got our podcast episode count (412) exactly right. Nobody saw that coming.
Gemini's grown up. Remember when it couldn't even connect to real-time data? Now it's delivering accurate stock prices and weather updates with slick visuals.
Perplexity? Big yikes.
It gave us 8-month-old data for the Bears schedule and completely whiffed on current AI news. Their pro features aren't saving them from accuracy issues. We wanna have a huge AI crush on Perplexity, but we’ve noticed it slowly trending downward and recently focusing more on shoving products down our throats vs. answering user intent.
What it means:
Stop treating all AI search engines as equal. They're not.
For real-time data like stocks or sports? Use ChatGPT Search or Meta AI.
Need location data with maps? ChatGPT Search or Perplexity.
But double-check EVERYTHING from Grok or you'll end up with Twitter-sourced hallucinations from bot farms.
3 – Winners, losers and surprises 🏆
Hold your sour grapes — this ain’t scientific shorties.
Don’t take our fun lil live search battle experiment as grounds for your business’s decision making.
You GOTTA run these AI search bots through the wringer like we did, but with all of YOUR common use-cases.
But for us, the (subjective) scores kinda shocked us.
- ChatGPT Search: 65
- Meta AI: 58 (Wait, what?)
- Google Gemini: 57
- Microsoft Copilot: 55
- Perplexity: 53 (Whoa!)
- Grok (X/Twitter): 47 (Zero shock value)
ChatGPT crushed visual presentation and accuracy. They're not just winning — they're changing how we expect search results to look.
Meta AI came from NOWHERE to take second place. Their free tool beat most paid competitors. We’ve kinda been sleeping on Meta. Maybe we should pay more attention?
The biggest loser? Perplexity. We think they've lost their edge recently. We’ve seen hallucinations and inaccuracies become more common place.
What it means:
There’s a clear path forward for businesses:
1. Get your internal brain trust together, and find out how you’re wasting time the old school way. (Manually browsing, compiling info and creating docs.)
2. Run a similar use-case study as we did today, live. Keep your own tabs.
3. Map the best use-cases to the AI-powered search tool that is reliably the most accurate.
Easy Yeezy.
What do you think?
Hit us with a reply, and make sure to join us tomorrow as we dive deeper into this topic!
⌚
Numbers to watch
$4 Million
Ex-Microsoft employees have received $4M from Accel to build an AI tool for product presentations.
Now This …
Let us know your thoughts!
Vote to see live results
Do you attend our livestreams?Every weekday, we bring you fresh AI insights, exclusive interviews, and breaking news with our Everyday AI livestream. |
Reply