The AI Daily Brief: Artificial Intelligence News and Analysis - The Most Important AI Product Launches This Week

Episode Date: June 28, 2024

The productization era of AI is in full effect as companies compete not only for the most innovative models but to build the best AI products. Learn how to use AI with the world's biggest library of fun and useful tutorials: https://besuper.ai/ Use code 'youtube' 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, we look at all the various product announcements that happened this week. 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. Hello, friends, last day of travel here. And so this might be a slightly clipped, a little bit of a shorter episode, but we are looking at something that really characterized this week, appropriate for the week that I was at the AI Engineer World's Fair. basically we are going to look at a set of specific product announcements. So these are new features, new tools, new models,

basically things that allow users to do more with their AI platforms. We kick off with Anthropic, which has followed up their announcement of Claude 3.5 Sonnet and artifacts from last week with a new tool that they are calling projects. They write, Claude AI Pro and team users can now organize their chats into projects, bringing together curated sets of knowledge and chat activity in one place. with the ability to make their best chats with Claude viewable by teammates.

With this new functionality, Claude can enable idea generation, more strategic decision-making, and exceptional results. So what might you use this for? Well, let's take, for example, the idea of team onboarding. With the project knowledge base, you could add all sorts of documents that relate to company policies, standard operating procedures, things like that, and that becomes the basis for the project. Anthropic Rights, this added context enables Claude to provide expert assistance across tasks.

What's more, on top of just adding different documents, you can also define custom instructions that help tailor Claude's responses. For example, one project might ask for a more formal tone of voice, given whatever the outputs are supposed to be. What I had said about Claude Artifax is that what it represented wasn't so much a big model shift, but a user interface shift, a user experience shift, if you will. Projects is another version of that. It's starting to put a U.S. around a particular type of use, in this case, team use, that's

just going to make it much more intuitive for people who have that particular use case. I think this and many of the rest of the announcements in this episode will show that we are firmly in the user interface or user experience era of generative AI. Speaking of user experiences, chat GPT's Mac app is now available to everyone. One of the things that makes this a really interesting experience is that by pressing option and spacebar, you can bring up chat GPT on Mac from anywhere. The goal for OpenAI is, of course, to integrate chat GPT into your experience in a much more fluent way. They want chat Chapti to be a companion, a co-pilot for basically anything you're doing

on the computer. We've seen in a number of demos over the past couple months, the advanced versions of how they imagine that going, how voice integrates with it, how video becomes a part of it, and how they might give chat Chbt access to look at your desktop. So I think that you can think of this as step one to start to get people habituated with calling up and asking chat Chapti for things as they're interacting with their computer and other contexts. Speaking of voice, however, earlier this week, we also got an official delay of the voice assistant. Now, the cries from the community to actually get this feature have been getting louder and louder.

People are really frustrated that OpenAI seems to have become one of those companies that announces the thing way in advance of actually having the thing. Last year, it was Google who was doing that, and people were none too pleased about that either. The justification that OpenAI gave was that they, quote, need one more month to reach our bar to launch. They write, we're improving the model's ability to detect and refuse certain content. We're also working on improving the user experience

in preparing our infrastructure to scale to millions while maintaining real-time responses. They do say that as part of their iterative deployment, they are starting with an alpha group of users to get more feedback, but they say all plus users won't have access until the fall. Yesterday, I posted on Twitter a video of Roman Hewitt from OpenAI demoing. Some of these new features all come together to create a documentary in the span of just about three minutes. First, he takes a video generated by Sora, which is amazing, and then he had ChatGBTBT watch that video to come up with a script, And then he recorded a sample of his own voice to create a new voice model of himself that could narrate using that script, which was then overlaid onto the video, and which could then be translated into other languages.

It's an incredibly impressive demo. I suggest you go check it out on my Twitter, X.com slash NLW. Of course, the critique for many is this would be great if we actually had access to any of these tools. Still, it clearly shows the potential in where things are going, and I think even though people want it right now, they're still pretty excited about that future. Still more news in OpenAI world. They have announced another licensing deal this time with Time magazine. Axios writes that the deal gives OpenAI access to Times archives from the last 100 years or so to train its models as well as to cite time in its responses. We've talked about this before.

These deals to me are really not so much about training data, although that's a nice bonus. Instead, they are about OpenAI moving ChatGBTT into a competitive slot with Google Search, where they're trying to get everyone to just start whatever question they have with chat GPT instead of with Google. Time also partnered with another AI company 11 Labs. Basically, the company has added automated voiceovers to Time.com content. The idea is to make it easier for people to consume time content in whatever fashion makes sense to them. If people are on the go, if they're listening on the subway or something like that, this sort of audio narration can be extremely useful and of course done at scale with the help of a partner like 11 Labs.

11 Labs this week also announced a new iPhone app that basically brings this type of partnership to the entire web. The idea is that with this new app, you can effectively turn any website into a podcast. The app gives you the choice of multiple different voices, or even synthesizing your own voice, to read basically anything on the web. One example they gave, for example, in the promotional video, is a woman walking around a grocery store listening to a particular recipe being read out. Today's episode is, of course, brought to you by Superintelligent, and today I want to tell you about Super Intelligence Summer Challenge. We have a couple of low-key months. key, at least in the professional sense, and that makes them a perfect time to build some new skill.

If you want to come back to the fall, to your work, to your school, to wherever you are, with a whole new set of talents and capacities when it comes to artificial intelligence, you should come join the Super Summer Challenge. Each week will share a new challenge that gets you actually using the most important tools in AI and figuring out what these things are going to be good for in practice in your job and in your career. As you learn, you'll also become part of a community that discusses the projects and use cases that we're all sharing. If you have wanted to dig deeper into AI and have just been waiting for the right excuse, this is it. Use code summer fun to join the summer challenge and get 50% off your first couple of months. You can get all the

information you need at B-supert.a. That's B-super.a. See you there. Over in the land of Google, Gemini has increased its context window to two million tokens. They had announced this back at I.O. But at that point, it was behind a wait list. This week, they wrote, we're opening up access to the 2 million token context window on Gemini 1.5 Pro for all developers. We are still barely scratching the surface of what that type of long context window can enable for people, but I'm sure we're going to start to see examples of it coming soon. Another little bit of Google News, Google Translate has added 110 languages. That's its largest expansion ever, and of course it all comes down to AI. A lot of these languages include

regional dialects and really reinforce just how valuable AI is going to be in closing linguistic barriers between different cultures. Moving over into a different sphere of AI, character AI had some updates. This is one of those platforms that many of the business listeners of this community might not be as familiar with, but is extremely popular. Character AI is kind of exactly what it sounds like. It allows people to interact with AI avatars. So far, that's been entirely in text, but now they're allowing users to actually talk with them over calls as well. TechCrunch reported that during characters test of this, more than 3 million users had made over 20 million calls. Trying to suggest that there's some utility other than just fun, they said that those calls can be useful for things

like practicing language skills, mock interviews, or adding them to role-playing games. If you want a sense of how valuable people think that this type of interaction is, the information reported earlier this week, that Google is starting to work on a challenger for character AI and the meta chatbots that are also kind of a competitor with character AI. Writes the information. Staff have discussed launching the new chatbots as soon as this year. The plans which haven't been previously reported speak to how tech giants are looking for ways to turn breathtaking advances in generative

artificial intelligence into apps that can keep consumers hooked. Speaking of meta, it has also made moves in this space, finally starting to test user-created AI chatbots on Instagram. In a post on Zuckerberg's channel, he said, rolling out an early test in the U.S. of our AI studio so you might start seeing AIs from your favorite creators and interest-based AIs in the coming weeks on Instagram. These will primarily show up in messaging for now and will be clearly labeled as AI. It's early days in the first beta version of these AIs, so we'll keep working on improving

them and make them available to more people soon. So basically, it sounds like step one is allowing existing creators to build these types of AI avatars and then roll it out more slowly to everyone. Zuckerberg has been fairly open about the fact that they're just not sure how this is all going to play out, said Zuck, I don't think we know going into this what is going to be the most engaging and entertaining and trust-building formula. So we want to give people tools so that you can experiment with this and see what ends up working. Last one today, closing on a much more creative and business-oriented launch, Figma is holding their big annual conference in San Francisco this week,

and as part of that, they announced a slew of new AI tools. The company writes, our main goal is to give you the tools you need to do your best work, to look beyond the hype and find real solutions to real user problems. Through that lens, we're excited to introduce Figma AI, a collection of features designed to help you work more efficiently and creatively. So, some of those tools include enhanced search. Visual search lets you find and reuse designs by uploading an image, selecting an area on your canvas or entering a text query. An asset search uses AI to understand the semantic meaning and context behind your search queries.

They're integrating AI-powered content generation tools around both copy and images. They're doing something called quick-click prototyping. By clicking Make Prototype, you can rapidly turn static mocks into interactive prototypes. And it just goes on from there. You can generate designs from text prompts. There's automatic layer renaming. All in all so far from what I've seen, designers are incredibly excited about these tools. And since our team is a big Figma user, I am sure we will have more direct feedback on them to share soon.

For now, though, that is going to do it for today's AI Daily Brief. Like I said, tons and tons of product announcements this week. Really exciting stuff to actually dig into and get your hands dirty with. Appreciate you listening as always. And until next time, peace.