- Everyday AI
- Posts
- Sam Altman and OpenAI wants $7 trillion for WHAT? 🤯
Sam Altman and OpenAI wants $7 trillion for WHAT? 🤯
🤖 Gemini Advanced Overview, OpenAI's AI agents, FCC bans fake AI robocalls, Sam Altman seeking $7 trillion and more!
Everyday AI
February 09, 2024
Outsmart The Future
Today in Everyday AI
7 minute read
🎙 Daily Podcast Episode: Did Google just release a ChatGPT killer? Gemini Advanced is here and we’re breaking it down. Give it a listen.
🕵️♂️ Fresh Finds: Creating no edit videos from anything, Midjourney launches alpha site, and How Disney uses AI. Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: Midjourney might ban political images, FCC bans fake AI robocalls, and OpenAI working on AI agents and a $7 trillion investment. For that and more, read on for Byte Sized News.
🚀 AI In 5: Is Google’s new Gemini Advanced really…advanced? Here’s what you should know! See it here
🧠 Learn & Leveraging AI: looking to get the lowdown on Google Gemini Advanced? We’re breaking down 7 things you should know. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We talked about how AI helps your biz speak all languages, Google says bye to Bard, and Microsoft Copilot gets a redesign. Check it here!
Google Gemini Advanced - 7 things you need to know 🤖
Did Google just release a ChatGPT killer?
Google is going all in on Gemini, first rebranding Bard to Gemini and now releasing its new Gemini Advanced, a paid offering to the free Gemini
Should ChatGPT and other AI chatbots be worried?
We're diving in and taking a look at Gemini Advanced and comparing it to ChatGPT.
Join the conversation and ask Jordan questions on Google Gemini here.
Also on the pod today:
• Features of Google Gemini ⚙️
• Performance and Comparisons 🥊
• Issues with Google Gemini 😬
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 – Rizzle lets you create no edit videos from anything, Heyday is an AI copilot for all your docs, notes, and ideas, and Capitol AI is AI content creation blending storytelling, design & research.
Trending in AI – Midjourney has announced access to its alpha site for those who have generated more than 1,000 images.
Big Tech - Microsoft is hinting at some AI upgrades to Windows 11 as it has an upcoming OS update.
Pop Culture – Check out how Disney is using AI for its streaming ads.
AI in Sports – This is how AI is helping the NFL prevent injuries.
1. Midjourney Considers Banning Political Images 🚫
Midjourney's CEO hints at banning political images like those of Biden and Trump due to rising concerns over deepfakes. The company's move aims to curb the spread of fake images and misinformation ahead of the upcoming US presidential election. Despite existing rules against misleading content, Midjourney faces challenges in preventing the misuse of its AI tools for political disinformation campaigns.
2. FFC Bans Fake AI-generated Robocalls 📞
The FCC is cracking down on scammers using AI-generated voices to impersonate people like Taylor Swift and President Joe Biden in robocalls. These fraudsters are now on notice as states gain new tools to protect the public from misinformation and scams.
3. OpenAI Working on New AI 'Agent' Software 👀
OpenAI, with Microsoft's backing, is developing software agents to automate complex tasks directly on users' devices. These agents will handle various web-based activities, from collecting company data to booking flights, promising to tackle tasks independently. The upcoming tech could revolutionize how we manage personal and work-related responsibilities.
4. Baidu Teams Up with Lenovo to Boost AI Smartphones 📱
China's Baidu joins forces with Lenovo to amp up AI on smartphones and take on global tech giants like Google and Apple. With Lenovo integrating Baidu's Ernie large language model into its devices, the competition in the AI-capable smartphone market is heating up!
5. OpenAI CEO Sam Altman Seeks $7 Trillion for AI Chips 🤑
Sam Altman is on a mission to revolutionize the global semiconductor industry with trillions of dollars in investments, aiming to address the supply-and-demand issue for AI chips. Altman’s ambitious project could require raising $5 trillion to $7 trillion, attracting interest from various investors including the government of the United Arab Emirates.
Google Gemini Advanced: 7 things to know!
Is Google’s Gemini Advanced really…advanced?
Is it better than ChatGPT?
We’re diving into Gemini Advanced to tell you 7 things you should know about Google’s new paid offering.
Check out today's AI in 5.
Or check out this related video:
🦾How You Can Leverage:
If you were going to buy a new car, would you test drive every ride on the lot?
Nah.
But, you got us.
With new GenAI tools popping up faster than little vermin in Whack A Mole, we get that you don’t have time to dork out on every new LLM.
So we did. 🤓
Today, we took an under-the-hood look at Google Gemini Advanced.
Here’s the tl;dr overview
Google Bard was first powered by the LaMDA model.
Then Google Bard was powered by the PaLM model.
Then Google Bard was powered by PaLM-2.
Then, Google Bard was powered by Gemini Pro.
Now Google Bard is dead.
And Gemini is powered by Gemini Pro. (Free)
And Gemini Advanced is powered by Gemini Ultra 1.0 (Now $20/mo after a 2-month free trial)
Not confusing at all, right?
Now that we’re kinda up to speed, let’s jump behind the wheel of this new Gemini Advanced ride and see if it’ll win some awards or if it’s a lemon.
Will you be able to use Gemini Advanced today to grow your biz? Is it ChatGPT challenger?
Here’s 7 things you need to know about Google Gemini Advanced.
Let’s ride shorties. 🚗
1 – Not Available for Workspace Users 🤨
The entire upside of an internet-connected LLM is the ability to use it in a work setting, right?
Maybe not? 🤷
Many Google Workspace accounts can’t sign up for Gemini Advanced.
Want to connect all your docs, emails, calendar, etc to Gemini Ultra?
Too bad, apparently.
What it means:
If you use Google for your main work accounts, you may only be able to use Gemini Advanced with your personal Gmail account.
Kinda useless if you ask us.
2 – Struggles With Its Identity 😵💫
Ball = dropped.
Any large language model should have AT LEAST a system prompt that tells it what it is.
What happens when you ask Gemini Ultra the difference between Gemini Pro and Gemini Advanced?
For us, we got told it doesn’t exist and instead got pointed to a crypto product. 🥴
What it means:
In our opinion, Google skipped over some of the basics.
People need to be able to interact with a model and ask it questions about its capabilities. You may get better or different results, as Google is undoubtedly in scramble mode squashing bugs.
At least we hope.
3 – Struggles with Real-Time Events 🗞
Isn’t the big advantage Google has here over others, real-time search?
Welp, not for Gemini Advanced apparently.
It doesn’t know about the upcoming Super Bowl, of its own existence, or even who won the recent New Hampshire primary elections.
What it means:
Can we just put a facepalm emoji here?
Oh, it’s our newsletter so we can do what we want?
Bet.
🤦♂️
4 – Reasoning and Logic Lacks 🧠
In fairness, reasoning and logic are a bit tricky for many large language models.
And keep in mind, you’ll get different results each time you run a similar prompt.
But, here’s what we got when we asked both Gemini Advanced and ChatGPT Plus the same prompt:
“please write me 3 short jokes that start with the word 'What' and end with the word 'blue’”
Google:
OpenAI:
What it means:
As we talked about in a previous episode, the next version of the GPT tech is supposed to make great strides in reasoning and logic.
So we assumed that Gemini Ultra would be light years ahead of GPT4.
Not so much.
5 – Image Generation and Watermarks 🖼
Big win here for Google. ✅
(Did you think we were gonna be ALL negative? Nah!)
Using the Imagen model inside of Gemini, Google will now include invisible watermarks for AI-generated images it produces.
What it means:
Even though Imagen images from Gemini and Dall-E images from ChatGPT don’t look very real (at least compared to Midjourney) we’re happy to see this responsible move from Team Google.
6 – Mobile App Available 📱
So… now you can use an iffy model on mobile too?
(Seems kinda like a consolation prize to us.)
Alright, we’re maybe being harsh, but we’re actually fans of Google rolling out the dedicated Gemini app on Android.
We’d like to see the same on iOS, but right now they just included Gemini capabilities in the Google app.
What it means:
Once Gemini improves on points 1-4 above, having that kinda power on mobile is needed.
It took OpenAI a while to roll out its dedicated mobile app, so we’re excited to see Google lead with mobile.
7 – Gemini Ultra outperforms GPT-4? 😳
Soooooo about that.
We went on a rant on Google initial Gemini rollout in December. So we won’t beat that dead horse.
(Down Bullseye!)
But, we’re excited to see updated benchmarking now that Gemini Ultra is out in the wild.
What it means:
Initial benchmarks showed Gemini Ultra to win like every benchmark, but many pointed out that the benchmarks might have been flawed.
Keep an eye out soon for apples-to-apples benchmarks to see just how well Gemini Ultra 1.0 stacks up to GPT4.
⌚
Numbers to watch
200 AI companies
The U.S. AI Safety Institute Consortium (AISIC) was announced, uniting over 200 member companies and organizations in support of AI safety and trustworthiness.
Now This …
What do you think?
Vote to see live results
Will you be using Google Gemini Advanced? |
Reply