The AI Daily Brief: Artificial Intelligence News and Analysis - Canva Acquires Leonardo - Is a Wave of AI M&A Coming?

Episode Date: July 31, 2024

Canva has acquired AI image generation company Leonardo, marking its second major acquisition this year. What does this mean for the AI startup space and the future of M&A in the industry? Explore the implications of this deal and the potential for increased consolidation in the AI market. Stay updated with the latest in AI by following the AI Daily Brief. Concerned about being spied on? Tired of censored responses? AI Daily Brief listeners receive a 20% discount on Venice Pro. Visit ⁠⁠⁠⁠⁠https://venice.ai/nlw ⁠⁠⁠⁠and enter the discount code NLWDAILYBRIEF. Learn how to use AI with the world's biggest library of fun and useful tutorials: https://besuper.ai/ Use code 'podcast' for 50% off your first month. The AI Daily Brief helps you understand the most important news and discussions in AI. Subscribe to the podcast version of The AI Daily Brief wherever you listen: https://pod.link/1680633614 Subscribe to the newsletter: https://aidailybrief.beehiiv.com/ Join our Discord: https://bit.ly/aibreakdown

Today on the AI Daily Brief, Canva has acquired leonardo.aI. Is this the beginning of a new M&A wave in AI? Before that in the headlines, Meta releases AI Studio and gives Instagram users the ability to create their own custom AIs. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show notes. Welcome back to the AI Daily Brief Headlines edition, all the AI Daily News you need in around five minutes. We kick off today with some new product releases. from Meta. Obviously, meta was big in the news last week with their model releases for

Lama 3.1, which included for the first time their large model, Lama 3.105B. But now we are back to a more consumer-focused announcement where people can create their own AI avatars. The company writes that they're rolling out AI studio, which they call a place for people to create, share, and discover AIs to chat with. Anyone can create their own AI, designed to make you laugh, generate memes, give travel advice, and so much more. They're also saying that creators can make AI versions of themselves, or or as extensions of themselves to use their phrase, to answer common DM questions and story replies. AI Studio is the first product that Meta is releasing that's built on Lama 3.1.

AI Studio is available on the web or in the Instagram app. When you create an AI, they say it can be just for you or you can share it with your followers and friends. With the studio, you can customize an AI character's name, personality, tone, avatar, and tagline. They've even created an 18-page handbook to help you actually do this. Although very different, there's some similarities with custom GPTs in that these AIs can be just for you or they can be shareable more broadly. Some of the examples they point to are a chef-created AI called Eat Like You Live There, which offers personalized tips for embracing local dining customs while traveling, What Lens Bro, which is exactly what it sounds like, an AI from a videographer and photographer

that offers tips for finding the perfect lens, as well as some more positive affirmation and meme type AIs as well. Now, I think one of the more interesting things about this is the focus on giving creators the ability to create an AI as an extension of themselves to interact with their fans through DMs and story replies. Meta writes creators can customize their AI based on things like their Instagram content, topics to avoid, and links they want it to share. They can do this through the professional dashboard in the Instagram app. They also point out that responses from creator AIs are clearly labeled so there's full transparency for fans. One of the things that makes meta an interesting company to follow when it comes to AI is the fact that they're competing both on the the model and ecosystem side of things as well as on the consumer product side, so I will certainly be watching to see what type of uptake, if any, there is for these new custom AIs.

Yesterday, we discussed how Apple was reportedly pushing back the release of its Apple intelligence features to October, whereas iOS 18 is coming out in September, but as a little advanced morsel, Apple did officially debut its AI features for developers as part of their iOS beta. Investopedia writes that the beta version's release could boost confidence in Apple's AI potential, following these reports that it might not release its AI update until after the planned iOS 18 rollout in September. The technical team at Apple also used it as an occasion to share more about what they had been building. Roming Pang from Apple writes, as Apple intelligence is rolling out to our beta users today, we are proud to present a technical report on our foundation language models

that power these features on devices in cloud. The report describes the design and evaluations of our LLMs in detail, including architecture, data curation, pre-training, and post-training recipes, optimization, feature adaptation, and evaluation results. While these LMs are not chatbots, we train them to have general purpose capabilities so that they can power a wide range of features including summarization, writing assistance, tool use, and coding. Now, there's been a lot of discussion of these particular models, but I think the interesting thing from my standpoint is just that it further shows how much more Apple is engaging with the larger developer community and the AI community in general when it comes to their model development. One other interesting nugget,

Apple said in that the technical paper that the two AI models that it used in Apple intelligence were pre-trained on Google design chips in the clouds. CNBC and other media outlets are trying to characterize it as Apple trying to find alternatives to Nvidia, but there's nothing necessarily in that note to actually indicate that. For example, later in the article CNBC writes, Apple doesn't name Google or Nvidia in its 47-page paper, but did note its Apple Foundation model or AFM and AFM server are trained on cloud TPU clusters. Speaking of Google, unfortunately, once again, people are talking not about their technology, but about the discourse around them, summed up in this New York Intellectalibuble.

The Intelligencer article, everyone hates that Google AI Olympics commercial. The Intelligencer writes, If you've been watching the Olympics, you have inevitably been exposed to an advertisement from Google called Dear Sydney. The premise involves a father talking about his daughter, a grade school track and field athlete who would like to write a fan letter to Olympian and 400 meter world record holder, Sidney McLaughlin-Liveron, says the narrator dad, she wants to show Sidney some love and I'm pretty good with words, but this has to be just right.

What does this made-up man to do? To share this sweet moment with his daughter, he decides to phone it in. The dad tells Gemini Google's artificial intelligence model to, quote, help my daughter write a letter telling Sidney how inspiring she is and be sure to mention that my daughter plans on breaking her world record one day. Continues the article, what? Why would a dad, who is pretty good with words, need an AI model to help his daughter write a heartfelt message to her favorite athlete? Aren't these moments what parenthood is all about? What sort of lesson is this? Not only does it imply to your kid that it's okay to offload writing assignments to AI, it also suggests a good idea to let the computer express feelings for you.

Now, regular listeners know that I am going to be the first to jump up and down all over people who I think are just bringing some pre-existing bias and hatred of AI to their critique. But I do think that the accusation that this feels a little tone-deaf to be kind of accurate here. I saw a lot of people even who really like AI, pointing out that if this is the best use case that Google can think of, it's not a particularly good look for the field. Indeed, one might argue that the whole point of AI is to create more space and time for exactly this type of human interaction, rather than outsourcing these incredibly special tasks. Lastly, today, a little prediction from Silicon Valley around the advance of AGI. At an event last week, Kora and Poe founder, as well as OpenAI board member,

Adam DeAngelo, was asked when his best guess was, in terms of when AI would be able to do the majority of work humans do today, and he said within five to 15 years. Said DiAngelo, I think that at some point there's going to be a threshold we hit where the AI is able to do the work of machine learning researchers that are currently creating the AI itself. After AI models can create a feedback loop known as recursive self-improvement, DeAngelo said, we will have achieved, quote, something that's fundamentally very different than the kind of AI we have today. So what do you think? Is 5 to 15 years a good guess? Is it too short, too long? Let me know in the comments here on YouTube or on Spotify. But for now, that is going to do it for the headlines. Up next,

the main episode. Today's episode is brought to you by Super Intelligent. As you guys know, super intelligent is a platform we are building to help everyone, individuals and teams, maximize their use of AI. We help you figure out how to use AI tools, as well as what to use AI for. And this is really important. The whole goal of superintelligent is not just to give you tutorials and lessons, but to show you how other people like you are actually getting value from AI right now. For those of you who are still out there working, learning, and grinding deep in the summer, I'm excited to share our best offer ever. If you sign up with code year 50 right now, you will get 50% off the already reduced annual price. That means you'll get access to Super Intelligent for a full year for less than $100. bucks. Again, that code is Year 50 for 50% off the already reduced annual fee. The Year 50 code will

expire at the end of July. So if you are interested in this, head on over to B-supert.a.I. And can't wait to see you on the platform. Today's episode is brought to you by Venice. The leading AI company store your entire conversation history and attach it to your identity forever. That's every question you asked, every answer you receive, every image you generate, every thought you share with the machine it's all being spied on. If you trust all the company's hackers and NSA board members that will ever have access to your AI conversations, then rejoice, for you are well served. For the rest of us, Venice is an alternative. Venice is a powerful AI app for text, image, and code generation that respects you as a sovereign individual and believes privacy

and free speech are not only human rights, but necessary for civilizational advancement. Private, permissionless, and uncensored, you can try it for free without an account. AI Daily Brief listeners receive a 20% discount on Venice Pro. Benis.A.I.S.NLW. Daily Brief. That's NLW Daily Brief. All one word. Welcome back to the AI Daily Brief. First of all, an acquisition that happened. Canva, the online image editing suite, has acquired Leonardo, an AI image generation company. Today we're going to talk about both this deal, but also the state of M&A in general in the AI space and how it might end up affecting how the startup space around AI develops. So first, let's start with the news itself.

Canva tweeted this morning, exciting news. Welcome Leonardo AI. Since 2013, Canva's mission has been to empower everyone to design. Today, we're thrilled to join forces with Leonardo AI, a leader in generative AI. Leonardo's team in tech will boost our AI capabilities, enhancing our products and driving innovation. With 120 experts, we're set to unlock new horizons in design AI. Together we'll lead AI-driven creativity, helping our 190-plus million users bring their ideas to life. So a couple things about this. First of all, this is Canva's second big acquisition this year. The first was back in March when Canva acquired Affinity, which was a professional design software that was at the time used by more than 3 million creative professionals.

Leonardo will be similarly acquired in full, meaning that the entire company of 120 employees will come over as part of the acquisition, and Leonardo will continue to operate independently. Canva writes Leonardo AI will continue to develop its web platform, for its millions of users, including its business customers, now accelerated with Canva's financial resources, expertise, and licensed content from our Canva creators program. Now, Canva has been making a major push into AI. It's one of the platforms we cover most frequently over at Superintelligent, and they've long been thinking about how to integrate AI into their suite of tools. They launched a background remover back in 2019, a first iteration of a text to image tool in 2022, then in

2003 came their full Magic Studio suite. In their announcement post, Canva said that Magic Studio has been used more than 7 billion times. And they say in 2024, they've seen a 40% increase in monthly users who are using their AI tools. So what is the goal of this Leonardo acquisition? Well, one part of it is that Leonardo has been very successful in a short period of time. According to this post, they have more than 19 million users just two years after launching. But it's clear that there is something deeper going on here as well. You have to think that if you're Canva, one of your questions, is how much to invest in your own model development versus just using things that are available

either via open source or off the shelf. It appears that this acquisition is part and parcel of trying to get a little bit deeper on their own model development. They write, as well as using this acquisition to supercharge Leonardo's platform and user growth, we also plan to rapidly integrate Leonardo's leading technology and their Phoenix foundational model

into our existing suite of Magic Studio products, such as our AI image and video generator Magic Media. Deeper investment into foundational model research and development will be pivotal to both Canva and Leonardo, and we're enthusiastic to be also investing in supporting Leonardo's B2B adoption. Obviously, I have no idea about either of these businesses, except in so far as I have used Leonardo before, and we featured tutorials about it on Super, and I am an everyday user of Canva. But I think hold aside all of the Leonardo business stuff.

This acquisition for Canva, to me, is probably about increasing their technology chops when it comes to actual foundational model research and development. Exactly what they feel they need out of that, and what will give Canva as a product suite an edge compared to what it can get now, I'm not exactly sure. But if I had to guess, it feels like that would be the motivation to actually go all the way and do this deal. Now, the other piece, of course, could be Canva moving to be a little bit more aggressive about specific domain opportunities as well. They point to a recent report they did, the Marketing and AI report, and hone in on the marketing team use case for generative AI, discussing new workflows, new efficiency

opportunities. And it feels like potentially from a business standpoint that is the other side of this, trying to go after that particular market with a more robust product suite. Leonardo founder Ethan in his tweet about the acquisition certainly adds some evidence that part of their goal together is around the model model. Phoenix, which is the name of their underlying model, Ethan writes was just the start. The future of creating will be mind-blowing, you have my word. Leonardo's users, while excited for them, are a bit trepidacious. PZFAI writes,

Wow, this is unexpected news and I'm not entirely sure how I feel about it, to be honest. I hope this will help Leo thrive long-term. as it's an outstanding creative tool set that I feel I can't do without. I also hope it won't lose its amazing approach to innovation and will keep the great culture it's created for its user community. Now, I think that part of why this is interesting to me outside of the specifics of the Canva Leonardo deal, is that M&A has been very quiet in the AI space,

despite the fact that you would think that we would be heading into a consolidation period. It's been a couple of years since ChatGPT was launched, a couple of years since a lot of AI startups got funding, and this tends to be kind of the moment in the development of a new technology cycle, where lots of those promising early ideas either fail or get acquired. So far, we haven't seen a ton of that.

Back in March, CrunchPace News wrote, AI venture funding may be hot, but M&A remains slow. CrunchPace writes, Last year, despite all the hoopla and headlines concerning AI, actually saw M&A dealmaking in the sector declined 31% from 2022, with only 190 deals consummated compared to 276 the year before. In fact, in Q4, they said, dealmaking hit its slowest pace since the first quarter of 2019.

Only 39 deals were announced in the last quarter of 2023. This is annualized data crunched three months later, which is now five months out of date. But this has broadly been the perception for a little while now, that M&A was kind of quiet. Still, there is a sense that things might be changing. Back in May, Bloomberg wrote, AI's smallest firms brace for consolidation with tech giants on the prowl. They write, there's a wave of consolidation coming for artificial intelligence's smallest startups, as the weakest seek help and bigger names see Opportunity. That was the view of AI startup founders, CEOs, and investors.

who spoke at last week's Bloomberg Tech Summit. Said Hugging Face co-founder, Clem DeLang, in general, in AI, you're going to see more and more M&A because a lot of companies took a lot of very risky bets, and a lot of them are running out of money. Bloomberg also pointed out that at the time of writing back at the beginning of May, deal volume had been picking up. In 2024, according to Pitchbook,

there had already been 105 deals at that time worth a total of $35 billion. They also pointed out that the dollar value in 2024 had already outpaced the total for all the deals in 2023. And so the question is, is this the beginning of that anticipated consolidation wave? Right now, it's just a single acquisition. But my guess is we're going to see a lot more of this in the months to come.

For now, though, that is going to do it for today's AI Daily Brief. Until next time, peace.