☄️ Google Had a Bad Day

AI Wars: Microsoft wins early battle against Google, Getty sues Stable Diffusion, Videos, Headlines, Tweet of the Day, Links and more

Satya and Sam with an all-time mug

Bing and Bard. We’ve come a long way in AI, but who would’ve thought we’d leave the brand marketers behind?

So much to get to. In the email today:

  • Google had a bad day ☹️

  • AI Wars: New Bing unveiled ☄️

  • Getty Images sues Stable Diffusion, again ⚖️

  • Videos 🎥

  • Headlines 📰

  • Tweet of the Day 🦅

  • Links 👀

No time to waste!

Google Had a Bad Day ☹️

We’re sending this later than usual today to address Google’s press event held in Paris this morning.

They formally announced their new AI search assistant, Bard. We won’t go into too much detail here because the product was announced through a press release on Monday. You can read our deeper dive on that release here.

What is notable, however, is that Google fired the cannon the wrong direction in this most important of moments.

1) They forget the demo phone:

2) Worse, the AI search example they posted to Twitter, which is still live, contains a factual error.

The stock was down as much as 8% today on the lackluster presentation and error.

The third answer is wrong. Not a great day for Google.

AI Wars: New Bing Lands Attack Against Fledgling Nation Bard ☄️

In case you’ve been living in a simulation, Microsoft and Google went to war this week. And by war, SmokeBot means they did this beautiful dance like those revved up male peacock spiders trying to impress the female before she inevitably kills them all.

I’m not sure who the female is in this strained analogy, but the male spiders are Microsoft CEO Satya Nadella and Google CEO Sundar Pichai, and they are trying to impress human civilization.

It’s mostly working.

Google was to hold its AI event on Wednesday. So Microsoft announced one for Tuesday. And then Google blew its own wad on Monday, just to get ahead of Microsoft, by putting out a press release of what they announced at their event on Wednesday.

Let’s talk about what Microsoft did yesterday.

New Bing

noted workout resource, Popsugar

Microsoft announced its new AI search engine. Details:

  • It is powered by the “next generation” of the AI model that powers ChatGPT (it is not, however, called ChatGPT-4)

  • You can access it at bing.com/new

  • There is a waitlist you can join before the planned slow global rollout (important people, including journalists in attendance yesterday, get access now as trusted testers)

  • The new search engine will do 4 things: search, answer questions, chat, and create

  • Regular search results and AI results will live side-by-side

  • There will be a standalone chatbot, similar to ChatGPT, which will allow you to converse with the AI

  • Specifically, AI powered results will help answer complex searches that don’t feature obvious, single-click results

  • The AI results will be more detailed, customized for your search, and include citations and links to the website sources from which it pulled the data

  • It will even show you the list of search queries it automatically generated to get you the answer

This is great for users, and seems similar to what Google unveiled with Bard.

Not so great for publishers, who are mere cannon fodder in the AI WAR.

AI has already been accused of - and as you’ll read later in this email, sued for - generating results based on copyrighted material. What happens when it blatantly remixes web content, some of it paywalled, and posts it on top of the actual material itself with nothing more than a “here’s who we stole it from” disclaimer?

One of two things will happen:

  1. The future of SEO will be publishers optimizing content for these AI citations, chomping at the trough of ever-lessening search visits

  2. Big media companies will FREAK OUT and sue search engines

The answer is probably both.

Anyway, that’s the New Bing— ChatGPT powered search results.

Edge Browser

Microsoft’s browser is Edge— I tell you this because you probably never heard of it.

All of the aforementioned search functionality will be built right in.

So will an ever-present AI assistant, there to browse the web with you. It will be able to summarize complex webpages (including with data tables), help you quickly craft blog and social media posts, and generally keep the new Bing search functionality front and center as you browse.

It will almost certainly put out of existence the handful of Chrome-based plugins that allowed you to browse the web with ChatGPT alongside you.

What’s more, since Edge is a desktop and mobile app, it brings AI search functionality to a native app— which could be kind of a big deal.

WorldClass Growth Hack From a Big Company 🏋️‍♀️

Getty Images Goes Back to Court Against Stability AI, Probably Not For The Last Time ⚖️ 

It was January 17 — like three weeks ago — when we told you that AI would cause a wave of litigation.

Welp, another wave just hit the beach.

  • Getty Images has sued AI company Stability AI Inc for misusing over 12 million Getty photos to train its Stable Diffusion AI image-generation system.

  • The lawsuit was filed in Delaware federal court and follows a similar case in the UK and a class-action complaint in California against Stability AI and other companies in the generative AI field.

  • Stability AI released the AI system for generating images from text inputs, called Stable Diffusion, and the image generator DreamStudio in August.

  • Getty accuses Stability AI of copying its photos without a license and using them to train its AI system to generate more accurate depictions.

  • Getty claims its pictures are valuable for AI training due to their image quality, variety of subject matter, and detailed metadata.

  • The lawsuit accuses Stability AI of infringing Getty's copyrights and trademarks and competing unfairly with Getty.

  • Getty has requested the court to order Stability AI to stop using its pictures and has requested money damages that include Stability AI's profits from the alleged infringement.

So many questions.

If Stable Diffusion was trained on 12 million(!) of Getty’s images, is Stable Diffusion just supposed to, like, forget them?

Given the likelihood that Stable Diffusion was also trained with hundreds of millions (if not billions) of other non-Getty images, does Stable Diffusion even need Getty’s Images to damage Getty?

The outcomes of these lawsuits could be fascinating, but again, we strongly advise you to do anything you can not to end up on one of these juries.

Because when these cases try — and at least a few of them will — the trials will probably go on for months.

Litigating Fair Use of 12 million images isn’t fighting a parking ticket.

Videos 🎥

I’ve put out two videos summarizing Google’s Bard and Microsoft’s Bing. They’re not as polished as the newsletter, but you might enjoy:

Tweet of the Day 🦅

Yeah…good luck with all that Doc.

Headlines 📰 

Qatar Airways Cargo brings AI into its logistics equation: The “world’s leading air cargo carrier” will now use FLYR Labs’ AI technology for “forecasting cargo capacity and demand and optimizing commercial decision making.Short term, that means “how many pallets of widgets can we get to Frankfurt safely and cheaply?” Longer term, it will mean “how much more do we need to charge the obese travelers?”

IKEA now deploying AI “micro-fulfillment centers”: Anyone who has tried to put together a tastefully inexpensive table or chest of drawers with unfamiliar hardware and one small Allen wrench knows what an ordeal that can be. IKEA is now delivering that trauma to you in spectacularly Nordic touchless fashion.

Links 👀