The AI Daily Brief: Artificial Intelligence News and Analysis - 7 Observations From the AI Engineer World's Fair

Episode Date: June 28, 2024

Dive into the latest insights from the AI Engineer World’s Fair in San Francisco. This event, touted as the biggest technical AI conference in the city, brought together over 100 speakers and countless developers. Discover seven key observations that highlight the current state and future of AI development, from the focus on practical, production-specific solutions to the emergence of AI engineers as a distinct category. Learn about the innovative conversations happening around AI agents and the unique dynamics of this rapidly evolving field. 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're looking at what the AI Engineer World's Fair tells us about the state of AI development. 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. In this category of novelty slash human interests slash AI moving into pop culture, NBC has announced that they will offer a a customized daily highlight reel during the upcoming Paris Olympics, with AI-narrated generation that sounds like Al Michaels. Al-Michaels has been covering the Olympics for decades now,

and rather than him being, quote, hold up in a broadcast booth each night, briefly summarizing dozens of Olympic events, instead, NBC has trained an AI model on Al-Michel's voice, with, of course, Al-Micha's approval. Said Michaels, when I was approached about this, I was skeptical, but obviously curious. Then I saw a demonstration detailing what they had in mind. As the New York Times points out, AI generated Al Michaels is almost sure to draw interest given its novelty. But there has been no shortage of tales of embarrassing errors, face plants, and slightly alarming hallucinations as AI has burst into widespread use over the past 18 months.

So what's interesting about this is that where voice technology doesn't make sense is if it's just replacing a thing that it would be easy for humans to do. Where AI voice cloning gets more interesting is when it's actually applied to a type of volume or customization that would be extremely difficult to, not impossible for an individual to do on their own. This looks to be the second type of approach. Writes the Times, subscribers who want the Daily Peacock highlight reel will be able to choose the Olympic events that interest them most, and the types of highlights they want to see,

such as viral clips, gold medalists, and elimination events. From there, the Times writes Peacock's AI machines will get to work each evening cranking out the most notable moments and putting them together in a tidy customized package. Mr. Michael's recreated voice will be piped over the reels. Now, the Times also says that humans will make quality control checks on the AI highlight reels, but presumably that's in the aggregate, not humans checking each reel, as it wouldn't really be particularly scalable if that was the approach.

And here's where we get to the volume and customization that I said would be impossible without AI. NBC officials anticipate 7 million different variations of customized highlights. Obviously, that is a volume of work that Al Michaels, even as energetic as the 79-year-old is, simply could not do. So we'll see. I think they're right that this is likely to cause a lot of controversy and focus, but the concept could ultimately be pretty interesting. Next up, seemingly never not fundraising, Perplexity AI, has secured more money at a new $3 billion valuation, according to Bloomberg. SoftBank is chipping in $10 to $20 million as part of a larger $250 million funding round

at a $3 billion valuation. It sounds like the deal has not yet been finalized, given that no one was willing to comment on it, but there have been a ton of reports around this, and so I think it's fairly safe to say that some version of it is likely to happen. This would be around 6x growth in the company's valuation over the last six months. Over in the annals of how companies are starting to actually deploy generative AI, Amazon is apparently using Gen. AI in its finance organization, writes the Wall Street Journal, Amazon's finance teams are turning to generative AI in areas such as fraud detection, contract review, financial

forecasting, personal productivity, interpretation of rules and regulations and tax-related work. The company says that these are at a combination of experimental and implementation stages, said George, VP of finance technology at Amazon, while experimentation and getting to know the technology are things that we really want to speed up, actually deploying this into production and making sure that we are in a well-controlled situation is very, very important for us. The reason that this type of story matters is that it shows that we are entering a different phase with AI, especially when it comes to its integration into the enterprise. More and more you're going to see articles like this that aren't just talking about the little proof of concepts that people are doing, but instead the full-scale

organization-wide implementations, where companies are actually trying to wrestle to the ground the productivity gains that could come from AI. Speaking of, another story in the journal, Goldman Sachs has started deploying its first generative AI tool across the firm. This first tool is actually for code generation, and it will roll out to thousands of developers across Goldman Sachs by the end of the month. Wrights WS.J, Chief Information Officer Marco Argenti said the company's approach to generative AI involves centralizing all proprietary uses of the technology on an internal platform

and restricting them elsewhere. Said Argenti, it might have slowed us down initially, but then we definitely gained a lot of velocity afterwards. I will be interested to see how this plays out. I have a lot of skepticism around the approach of trying to tightly control things in quite the way that GS is doing. I think financial institutions obviously have a totally different level of risk assessment and compliance and restrictions around that might lead it to make sense to have this kind of approach when it comes to the sort of vertical AI that everyone in the organization is going to use. However, doing that to the exclusion of allowing employees to also experiment with other simpler, safer, productivity-winning AIs along the way, I think will end up being seen as a mistake.

Finally, a conversation that I think is probably worth a little bit more time even than just the highlights. Bill Gates has said that we shouldn't be overly concerned about AI's impact on energy use. Although Gates said that data centers would drive a rise in global electricity use of between 2 and 6%. Gates also said the question is, will AI accelerate a more than 6% reduction? And basically his argument comes down to the same thing that Sam Altman has been talking about, that in fact the demand for energy for AI is likely to produce energy breakthroughs as a second order consequence. Super interesting stuff, like I said, this is deserving of a much larger conversation,

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 fractional AI, my go-to AI Devshop. When we wanted to build a specific AI product feature, what we called our AI Toolfinder bot for super intelligent, we knew we needed incredible AI engineers that could build something fast, and so we turned to Fractional because they've got some of the best AI engineers on the planet. The team is a group of senior engineers in San Francisco working on some of the most exciting projects in applied AI. They work with everyone from startups all the way through the Fortune 500.

To request a free consultation, head to Fractional.aI. And again, if you want help identifying and building AI projects for your business, I highly recommend you check them out. Again, that URL is fractional.aI to request a free consultation. Today's episode is of course brought to you by Super Intelligent, and today I want to tell you about Super Intelligence Summer Challenge. We have a couple of low-key months, low-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 Bsuper.a.i. That's Bsupert.a.i. See you there.

Welcome back to the AI Daily Brief. Today I get to do something a little bit different and really fun. I am currently at the AI Engineer World's Fair. It's happening in San Francisco, and they're calling it the biggest technical AI conference in San Francisco. This is an event that I believe started out with a much smaller set of ambitions, but when they saw so much interest it just kept expanding. When all was said and done, they had over 100 different speakers,

delivering talks, workshops, et cetera, across nine different tracks. And there is just a ton of developer energy here. So what I want to do is in no particular order, share around seven different observations from my time here over the last couple days. The first observation is that this is very inward facing. What I mean by that is that the people who are here are not trying to impress tech crunch and they're not trying to impress investors. Although if that happens, that's a nice side benefit. Instead, this is a bunch of AI teams building for other AI teams. The types of things that these developers are working on are often infrastructure and solving,

problems of AI development and implementation, which I think says a lot about the state of where things are, which gets me to the second related point. These folks are super focused on production specifics, what makes this technology actually work. It is not forward-looking futuristic stuff meant to open people's eyes and make them dream of possible futures. It's really solving problems in the here and now. And part of what's shaping that, and maybe this is the third observation is that this event finds these people with now more than a year of learnings under their belt in this current stage of generative AI development. Basically, if this event had been held 12 months ago or 15 months ago, the type of conversations would have been very different. This time

around, they're informed by what people have actually learned are the problems, the challenges, the holes that need to get fixed for this technology to live up to its full potential. And as you can see, all three of these initial observations, the team's building for teams, the focus on how to make these technologies actually work, the year of learnings under their belt, all add up to a super applied feeling event. It is hyper-practical. And that brings me to my fourth observation. It also feels very separated from the hype.

It's not that somehow these companies and participants are blithely unaware of how significant the AI conversation is or what Wall Street is doing around AI. It's just not where they're focused. Again, these are engineers building for other engineers. And I think that for anyone who is concerned that AI is getting out over at skis, spending some time in this place, which is so focused on the practical and what comes next and what needs to get done now, would maybe make you think a little bit different about things. Next observation is that for as much time as we spend on the AI Daily Brief talking about AI in the enterprise and AI coming to companies,

there is much less Fortune 500 participation at this event at least. They do have an AI in the Fortune 500 track and there are some really interesting people here. And notably, the people that they have brought from big companies, are extremely technical. I think it could be a mistake to overly infer from the participation here exactly what the state of technical AI development is within the Fortune 500. In that I wouldn't want to extrapolate that just because there isn't all that much of that action here doesn't mean it's not happening elsewhere. It could have just not been as much of a focus for these organizers. But it is notable that at this event at least, the Fortune 500 is only very, very barely intersecting

with this whole world of AI engineering. It may be as simple as the natural place for Fortune 500 intersection with this audience being a little bit later in the development journey. But whatever the reason it is notable to me. A sixth observation is that there really is something to this idea of a new category called AI engineers. On the one hand, it could be tempting to just view that as marketing speak, to ask how different they are than other types of software engineers, or perhaps on the other end of the spectrum to point out that there have been AIML engineers for a ton of time, but this really does feel like a different type of middle category. I was talking with the CEO of a big agent company that was at this event, and they were talking about

how interesting it was that the AI Engineer World's Fair had carved out this middle space between, on the one hand, the super technical conferences made up of folks who have been working on these types of problems for 20, 25 years, and on the full other end of the spectrum, the absolute glut of business-y-type AI conferences, which are focused on use cases and things like that. This category of applied day-in, day-out, AI engineers who are focused on engineering in this specific discipline, but who aren't necessarily doing the same type of deep research that some of those others have in the past does feel like this interesting novel space. And I think the fact that this event grew so organically speaks to the resonance of this as an actual category. Lastly,

when it comes to specific topics, one that is notable even for a non-technical semi-outider, is that agents continue to be a major, major focus. Indeed, in their description of the event, they call AI agent applications the holy grail of AI, full self-driving autonomous execution. There are a ton of companies here working on agent-related projects and a significant number of conversations happening around it as well. However, and this is really interesting to note,

these agent conversations are so much more specific than the type of very generic, hypey things that we were seeing call it a year ago at this time. The companies that are working on, agent problems here aren't just a group of people trying to create a generic AI agent from scratch. They are working on discrete challenges within a larger agentic framework trying to solve them one by one as we evolve towards an overall ecosystem, powered by, enabled by, and primed for AI agents.

So once again, even in the most noticeably hypey and exciting and experimental piece of this whole field, still in practice, the conversations in the focus are much more practical and specific at this particular event. I have had a ton of fun at this event. I've had a great time talking to all the people here. Big congrats to Small and the team that have put this on. I think it's a smashing success. And I'm excited to come back in a year again or whenever you do the next event to see once more just how fast things are moving. For now though, that is going to do it for today's AI Daily Brief. Until next time, peace.