The AI Daily Brief: Artificial Intelligence News and Analysis - AI's Future is Here Now

Episode Date: December 22, 2024

A reading and discussion inspired by https://www.oneusefulthing.org/p/the-present-future-ais-impact-long Brought to you by: Vanta - Simplify compliance - ⁠⁠⁠⁠⁠⁠⁠https://vanta.com/nlw 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, why AI's future is right here in the present. 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 long read episode here before the holiday. And today, once again, we are basing our discussion on a piece by Professor Ethan Mollock. This piece is called the present future AI's impact long before superintelligence. The TLDR of this piece is something we've discussed.

quite a bit on this show, that even if we are facing currently a major plateau in the performance of generative AI, the impact of the current crop of tools and capabilities is still going to be absolutely transformative for work in ways that leave it nigh unrecognizable from where it is today. Let's read Ethan's take on this, or rather turn it over to an 11 Labs version of me to read Ethan's take on this, and then I'll come back for a bit of a discussion. The present future, AI's impact long before superintelligence. The AI labs are absolutely confident that larger, more powerful AI models are coming soon, ones that will enable autonomous agents and systems smarter than human PhDs.

You can see this confidence in two separate essays by the CEOs of two of the leading AI labs, Sam Altman of OpenAI and Dario Amadei of Anthropic, that discuss the coming age of superintelligent machines. But these are not uncontroversial assertions, and we do not know if they are right. Yet, in many ways, we do not need super-powerful AIs for the transformation of work. We already have more capabilities inherent in today's Gen 2-G-T-4-class systems than we have fully absorbed. Even if AI developments stopped today, we would have years of change ahead of us integrating these systems into our world. Today's AI models are already multimodal,

able to process and generate various types of media, like text, images, and sound. They can write code, operate computers, access the internet, and more. The pieces are all there, and we are starting to see them come together. They do not do any of this flawlessly and remain inconsistent and prone to hallucination. But there are many fields where AI abilities, flawed as they are, are already useful. Areas where perfect accuracy is not expected, or where having a second opinion is helpful, or where there would otherwise be no one to help, or where the best available human performs worse than the best available AI. Consider, for example, the combination of the ability to AI to both process images and reason over them. It means that you can add intelligence to any video feed by

just giving it to an AI, doing what was previously impossible. For example, I gave Claude a YouTube video of a construction site and prompted, you can see a video of a construction site, please monitor the site and look for issues with safety, things that could be improved, and opportunities for coaching. There is no special training here, just the native ability of Claude 3.5 Sonnet with computer use, taking screenshots every few seconds, and studying them. You can see the sped-up video of the system at work below. In the video, Claude analyzes various aspects of the construction site.

workers' protective equipment usage, placement of materials, work patterns, and potential hazards. These observations are interesting, but the system can go further. I then asked, What did you conclude? Write up your observations as a punch list. The AI created a spreadsheet summarizing what it observed in a few seconds, something that would have taken humans far longer. Note how it took all of the many issues it spotted across the video and applied reasoning to them, breaking them down by priority order, making logical inferences about how to address them, and more.

Then Claude asked me a question, would you like to create a tracking system for completion verification? That seemed like a good idea. So, I agreed and it made one, purposefully including fake names as an example of the data I had to fill in. The results seemed good from reviewing the video, but I am not an expert, and I would be surprised if there were not serious hallucinations mixed in. For this and many other reasons, I would never want this system to be used to punish or reward people. Yet consider a case where there would otherwise be no one monitoring a potentially dangerous environment or where mentorship or advice is lacking. Then an AI who could flag a human to dig into a potential issue or opportunity could be a useful asset. I improvised this system with a couple of prompts. With more work, the error rates and costs of AI monitoring will drop, even if no new models are

released. These systems will get better. Organizations will be tempted to deploy AI observers everywhere, governments may follow suit. What could be a mentor and safety check could become a panopticon where everyone is watched and judged by AI. The choices companies make, and the rules put in place by governments will determine whether AI is used to help or to monitor us. One of many complex adjustments we will need to make to an AI-filled world. But observation is only one area where AI is already showing high levels of capability. The digital world in which most knowledge work is done involves using a computer, navigating websites, filling forms, and completing transactions. Modern AI systems can now perform these same tasks, effectively automating what was

previously human-only work. This capability extends beyond simple automation to include qualitative assessment and problem identification. Here I ask Claude, go to the Walmart webpage and test it like a naive user trying to buy something, then go to Amazon and do the same thing. Write up your findings in a report, in a document. Again, you can see in the sped-up video that the AI goes to each website and roleplays a user searching for and buying products. It then wrote up two reports, a narrative and a testing report. There were no hallucinations I spotted, and while they are not the most insightful reports I have ever seen, they were quite solid. The AI is already a reasonable intern that, when given an assignment, executes it quickly and well, using judgment to solve problems along the way.

As models get better and these systems get less complicated to use, it is easy to imagine managers using teams of AI agents to do analysis and repetitive tasks in the near future. We saw how multimodal inputs and tool use transform how AIs interact with the world, but it gets stranger still when we add multimodal outputs. Here I invited an AI avatar made by Hagen into a Zoom call. The avatar is completely AI powered from the voice to the image to the behavior. In fact, I prompted the avatar to act in the most stereotypical and corporate possible way for a Zoom meeting. While the uncanny valley, that unsettling feeling we get from almost but not quite human representations, is obvious in the slightly unnatural voice and visual glitches like the changing shirt, the interaction fundamentally

mirrors a typical Zoom call. This is a first-generation tool and it actually works. I would not be surprised if many people are fooled by virtual avatars in the very near future. These capabilities demand immediate attention to both policy and practice. Even as imperfect as they are, current AI systems are already reshaping fundamental aspects of work, from how we monitor safety to how we conduct meetings. The choices organizations make today about AI deployment will set precedence that could echo for a long time. Will AI-powered monitoring be used to mentor and protect workers or to impose algorithmic control? Will AI assistance augment human capability or gradually replace human judgment? Organizations need to move beyond viewing AI deployment as purely a technical challenge.

Instead, they must consider the human impact of these technologies. Long before AI's achieve human-level performance, their impact on work in society will be profound, and far-reaching. The examples I showed from construction site monitoring to virtual avatars are just the beginning. The urgent task before us is ensuring these transformations enhance rather than diminish human potential, creating workplaces where technology serves to elevate human capability rather than replace it. The decisions we make now in these early days of AI integration will shape not just the future of work, but the future of human agency in an AI augmented world. Today's episode is brought to you by Vanta. Whether you're starting or scaling your company's

security program demonstrating top-notch security practices and establishing trust is more important than ever. Vanta automates compliance for ISO-2701, SOC2, GDPR, and leading AI frameworks like ISO-42,1, and NIST AI Risk Management Framework, saving you time and money while helping you build customer trust. Plus, you can streamline security reviews by automating questionnaires and demonstrating your security posture with a customer-facing trust center, all powered by Vanta AI. Over 8,000 global companies like Langchain, Lila AI, and factory AI use Vanta to demonstrate AI trust and prove security in real time. Learn more at Vanta.com slash NLW. That's vanta.com slash NLW. If there is one thing that's clear about AI in 2025, it's that the agents are coming.

Vertical agents by industry, horizontal agent platforms, agents per function. If you are running a large enterprise, you will be experimenting with agents. next year. And given how new this is, all of us are going to be back in pilot mode. That's why Super Intelligence is offering a new product for the beginning of this year. It's an agent readiness and opportunity audit. Over the course of a couple quick weeks, we dig in with your team to understand what type of agents make sense for you to test, what type of infrastructure support you need to be ready, and to ultimately come away with a set of actionable recommendations that get you prepared to figure out how agents can

transform your business. If you are interested in the agent readiness and opportunity audit, reach out directly to me, NLW at B-Super.a.I. Put the word agent in the subject line so I know what you're talking about. And let's have you be a leader in the most dynamic part of the AI market. All right, back to Real NLW here. Ethan really sums it up strongly in the last couple of paragraphs. Even as imperfect as they are, he writes, current AI systems are already reshaping fundamental aspects of work. And organizations need to move beyond viewing AI deployment as purely a technical challenge. Instead, they must consider the human impact of these technologies. five quick things I want to talk about following up from this.

In terms of ways that I have conversations with big enterprises about this particular issue, this idea that even if things stop now, there would still be an enormous amount of ketchup and transformation work to be done. The first thing I want to flag is use cases. We are still in the process of discovering use cases for these tools. Even things that seem like obvious one-to-one replacements for work that happens right now are in practice in many cases not that obvious. A great example of this, I think, is with meeting note takers. There are, of course, infinite versions of these tools. If you are anything like me, probably a third to a half of the

participants at any given Zoom or Google Meet are these note takers. But what are they actually useful for? I'm sure that for some folks, they do go back and refer to their notes, using them as a personal summarizer, something akin effectively to Dumbledore's Pensive. But I actually think it is the other use case of the Pensive, which is sharing memories with people, where these tools really thrive. At Superintelligent, what we do is we have certain key meetings set up to automatically push those notes into Slack channels for the people who weren't in the meeting. The real value of the summarization becomes not for the participants, but in bringing people who weren't in that meeting up to speed without relying on those participants rehashing the whole thing. It increases collective

knowledge, it increases our ability to be in sync, and it does so at a lower time cost. The point is, we spend months using these tools before we really landed on that being the best use case for them for us, and we're a company who exclusively does AI and focuses on helping enterprises figure out how to use AI. The point being that I think we are still just barely scratching the surface of really figuring out the use cases that are most valuable, even with today's current tools. Second theme that I want to discuss is user experience. it has really only been in the last half of this year that we've started to really see focus on user experience as a key vector of the AI product competition.

Up to that point, it was just all about model innovation and who had greater capabilities. But then you started to see things like Claude's artifacts, and all of a sudden, everyone is thinking about the UX of these tools. As that continues, it is going to open up tons and tons of new usage, as people unlock either more familiar or simply better user experiences for interacting with generative, of AI that get them using it more deeply. Next theme, innovation. In our discussion of use cases, I was just saying that we mostly think about one-to-one

replacements for today's activity with an AI-enabled version of today's activity. In marketing, I'm producing more content, and I'm producing it faster and cheaper, but I'm still producing content. I think, however, that the really interesting stuff is going to happen when we're not just one-to-one replacing current activities, but actually fundamentally thinking differently about what we can do because of Gen A.I. Tools. To use that marketing example, if tools like Cursor and Devon become usable enough so that non-coders can actually build software applications, all of a sudden, software becomes a new type of quote-unquote content that marketers could create.

When you think about what marketers care about, eyeballs and impressions, engagement and time spent, creating custom software or games, especially things that can be spun up quickly, deployed via social media and other channels, perhaps paired with cultural moments, and you have this entirely new category of things that marketers can do. And this is, of course, just one example. Some of this innovation will come from enhanced capabilities, but a lot of it's going to come from just people figuring out how to use the current crop of tools better. For theme agents, agents are one of the things that is most likely to benefit significantly

from expanded future capabilities. However, what we're seeing with vertical agents right now is that they're all. are some applications, some functions, that even right now, agents are really good at. I think figuring out how to integrate agents alongside AI-enabled workers is going to be a totally new and difficult discipline. Once again, even if nothing progressed from where we are right this moment. Lastly, fifth theme following up from this mindset, the organizations that are and will be most successful when it comes to AI transformation are those that are going to embrace a particular their mindset. They're going to embrace a mindset of continual learning, where there is never a done point,

where they are never, quote unquote, fully transitioned, but are always looking out for what's next. They are going to have a mindset of understanding their core values. They're going to know what their business stands for and how they operate, regardless of what technology they're using. And lastly, the mindset of not waiting is going to be so essential. Organizations that think they can just wait till someone else figures it out are going to be so woe. fully behind that they just don't stand a chance. And so again, ultimately my point here is that I agree wholeheartedly with Ethan. Even if things stopped now, we would still be at the beginning of a decades-long transition and transformation. There is no time to wait, and there is no substitute

for experience. Dive in, and of course, if you need help, reach out. For now, that's going to do it for today's AI Daily Brief. Until next time, peace.