Welcome, humans.
Despite earlier reports of a January launch, Google pulled a big surprise and dropped Gemini yesterday.
About time they learned the whole “underpromise, overdeliver” thing! Let’s dive in…
Here’s what you need to know about AI today:
- Gemini Pro makes Bard finally usable and competitive with free ChatGPT.
- Gemini Ultra should rival ChatGPT 4, but we can’t say for certain.
- Gemini Nano is a small step towards making AI in phones real.
- Meta announced a big batch of AI features on its products.
ChatGPT’s main competitor has arrived.
Google really spent all of 2023 saying “don’t call it a comeback”, and they finally have the new AI models to show for it.
After much anticipation, Google is finally giving us a look at Gemini Pro and Ultra. They’re competitive with ChatGPT 3.5 and 4, respectively, and they’re multimodal (meaning they can understand images).
Since Google Bard is already upgraded to use Gemini Pro, we decided to take Bard for another spin since we sorta wrote it off back in September.
Our take: Bard is good now! Especially for personal use cases that tap into other Google products like search, Gmail, Flights, Maps, YouTube, etc.:
- Daily digests for your email + drafting responses
- Travel planning
- Summarizing a YouTube video
Unlike the September version, today’s Bard is coherent, thoughtful and way more accurate. The best part might be that it embeds maps, emails, webpages and anything else that’s helpful.
If you’re not using the Google integrations or you need long prompts, you’re still better off using ChatGPT 3.5 or Claude (for those extra long PDFs).
As for beating ChatGPT 4 (which we recommend for work use cases), Google says Gemini Ultra is the one to watch, and we’ll have to wait for it. Until then, the community is staying snarky about its high benchmark scores.
After all, everyone knows that one guy who aced the SAT math section but can’t split a bill at a restaurant…
If you haven’t tried Bard, now’s a good time. Check it out for free here (must be on personal Gmail, not Google Workspace). Note: Bard is not using Gemini Pro if you’re in Canada or Europe.
FROM OUR PARTNERS
Learn to Build AI Applications With Zero Code.
AI is full of confusing buzzwords.
But behind all the tech-y terms (GPT, Claude, prompt, retrieval; what’s it all mean?) is a simple truth: it’s easier than ever to build an AI app.
The MindStudio team is hosting a free webinar on 12/14 at 12 PM PT / 3 ET to help you navigate the noise and build your own AI apps, zero code required.
Whether you're a seasoned AI enthusiast or a novice, this webinar will give you the skills to launch your own AI applications ASAP.
Google’s taking first shot at this AI-powered boost to phones.
The hype around Gemini Pro and Ultra buries a hidden gem: the third child called Nano.
Nano is the smallest version of Gemini, designed to live on phones. But it’s an extremely limited rollout right now. Gemini Nano is only available when you’re 1) using the “Smart Reply” feature 2) on WhatsApp 3) on a Google Pixel 8.
1 feature on 1 app on 1 phone is about as much commitment as La Croix gives to flavors.
Still, Google and Apple putting these models on our phones means soon, you can:
- write anything on any app using AI
- have an AI understand what you’re pointing your camera at
- voice chat with an AI on the fly.
All without an Internet connection, special apps or secret websites. Phew - now I can use AI instead of faking a phone call to leave an awkward convo.
While Google has to move slow in AI search because they have a lot to lose, they have the opposite with their phones. Let’s see how much they take advantage of it in 2024.
Around the Horn.
- Meta announces AI features for images, comment drafting and searching for Reels.
- AMD announces new chips, with Meta and Microsoft as customers.
- Andreessen Horowitz and LinkedIn release 2024 predictions, both with plenty of AI mentions.
- TIME names OpenAI CEO Sam Altman as “CEO of the Year”, which it has never named before this year.
- We previously wrote that the latest Claude update has a huge caveat. They fixed it by adding one line to the prompt.
Treat To Try.
Imagine is the new AI model that is powering Meta’s image generation features in Messenger and Instagram.
Using Imagine in Messenger and Instagram (by typing “/imagine [prompt]”) will also let you remix AI images that your friends generate. It’s multiplayer AI!
We curated the top 27 tools you need for work in our “Top Tools For Business”!
Thursday Trivia.
AI used to completely butcher generating fingers in AI art—they looked worse than a 3rd-grade art class. Now they’re pretty damn realistic.
One glance, everyone knows the rules: one cartoon is AI-generated and one is human-generated.
Which is which?
A/
B/
A Cat's Commentary.
Trivia Answer: A is real and B is AI!
That’s all for today, for more AI treats, check out our website.
Get your brand in front of 220,000+ professionals here.
See you cool cats on Twitter: @nonmayorpete & @noahedelman02
