Tech Digest ~ AI, weaponization and regulation
oh, maybe a few other things too...
Hello hello everyone, and welcome to my first edition of Tech Digest! I'm Isabela - but I usually go by Bela, a venture capitalist with a passion for tech (no surprise there), sports, art, film, and sciences. I'm thrilled to share my insights with you as we explore this ever-evolving theme. ⚡
In this newsletter, we'll be digesting some of the most thought-provoking and intriguing articles in the tech industry. My aim is to share and digest some of my favorite reads of the week. But of course, if you have any recommendations, feel free to share them with me.
So, sit back, relax, and let's dive into the world of tech together - oh, a non-tech surprise bit by the end. 🙃
~ Time to start digesting… 🍩
It's not surprising that AI is a topic of interest to many people, and there are many different perspectives on the theme. I'd like to share some of my favorite sources from this past week, along with a brief summary of each.
// Chamath Palihapitiya Joins Dan Nathan at iConnections Global Alts20
I once heard someone compare AI to the beginning of the App Store. When the App Store was first created, a lot of useless apps were developed, like the candle light app (my fellow millennials will remember this one 🙃), but no one could have predicted that a very useful app for connecting drivers with passengers would later emerge. Along similar lines, Chamath has made an interesting analogy. Just as the person who invented refrigerators profited from their creation, the development of large language models has opened up new opportunities for businesses to monetize data. However, just as Coca-Cola built an empire on the back of refrigerators, the true beneficiaries of large language models have yet to emerge.
He goes on to elaborate that while everyone is currently mining the open web, the future belongs to those who can develop private models. He predicts that the next wave of mergers and acquisitions may involve big tech companies acquiring inputs from smaller players to enhance their large language models. For early-stage investors (like myself⚡), the opportunity lies in investing in companies that are creating a unique data repository to feed these bigger players.
Chamath also discusses how these advancements could have practical day-to-day applications. Taking Coca Cola as an example again, companies could find new ways to streamline their operations and reduce costs. For instance, industries may replace their bulky refrigerators with cameras that capture images of their coolers every 10 minutes, allowing for predictive inventory management. This would revolutionize the value chain, potentially decarbonizing it and having a significant impact on both COGS and their carbon footprint.
In summary, Chamath and Dan Nathan delve into various topics, including the decline of risk assets, issues with growth investing, identifying indicators of a potential bottom in private tech, and the potential benefits for American companies due to increasing geopolitical tensions between the U.S. and China. If these topics interest you, I highly recommend giving their discussion a listen.
// [Research Paper] X-Risk Analysis for AI Research
When Elon Musk, Steve Wozniak, Andrew Yang, and many other influential personalities called for a temporary pause on the development of AI, I admit I was a bit surprised. However, this reflects the seriousness with which some experts view the potential risks of AI and the need for caution in its development. So, not only did I read the letter itself (which, let's face it, was widely read), but I was also curious to read some of the references cited in the letter. The X-Risk Analysis, in particular, struck me the most - it's a 20 minute read, but worth it. 🧠
Well, to start off, the potential weaponization of AI poses a clear risk to society as a whole. This concern stems from the fact that advanced AI systems can outperform humans in various tasks, including aerial combat, chemical weapon discovery, and automated cyberattacks. Furthermore, military leaders have discussed the possibility of AI systems having decisive control over nuclear silos, and superpowers of the world have declined to sign agreements banning autonomous weapons.
While the weaponization of AI may sound scary, the paper proposes a potential silver lining to it (thankfully). The paper suggests practical ways to make AI safer by focusing not only on its intelligence but also on its ethical and moral considerations. Researchers can improve the safety of their AI methods by taking two steps. First, they can measure how their methods improve safety goals or metrics, such as anomaly detection or robustness accuracy. Second, they can measure the impact of their methods on general capabilities, such as accuracy on ImageNet, to avoid any unintended negative consequences on capabilities. By doing so, these steps help ensure that AI is developed in a responsible and ethical manner, taking into account both moral and intellectual values.
// [Podcast] AutoGPT's massive potential and risk, AI regulation
In this podcast, Jason Calacanis, David Sacks, David Friedberg, and Chamath discuss the impact of generative AI on various fields such as art, videos, and industries like film (I have to say, I find it poetic how tech & art are so symbiotic👾). The speakers provide an example of how the movie "Everything, Everywhere, All at Once" was made with the help of only seven visual effects professionals and software. In just two years, we might see visual art professionals become visual directors, using AI to fully create characters in motion.
An interesting part of the discussion is about regulation. The speakers have differing opinions, which leads to a proactive debate on whether we are ready to regulate such technologies and the implications for foreign affairs. On one hand, the human brain supposedly doesn't evolve as fast as AI, which may lead to unforeseen problems in the future. On the other hand, restricting certain technologies may lead to a new geopolitical order (let's just say that the race of AI in the 2020's is the new space race in the 60s 🛰️).
// [Book] AI 2041 by Kai Fu Lee and Chen Qiufan
Well, it seems like I covered all sorts of media types (from podcast to videos) - but this was definitely not on purpose. But as a final AI-related recommendation, if you're looking for a thought-provoking and optimistic read on the topic, I highly recommend checking out "AI 2041" by Kai Fu Lee and Chen Qiufan. The book explores the potential impact of AI on various aspects of human life in the year 2041, and each chapter is divided into two parts. The first part is a science fiction section that presents a vision of the future in which AI has transformed society and human relations, while the second part examines the implications of AI on topics such as healthcare, finance, and privacy. Whether you're seeking positive insights or looking to think critically about how we can shape the future of this technology, "AI 2041" is a must-read.
At the end of the day, I agree with Sundar Pichai, Google's CEO, that AI could have a profound impact on human existence, surpassing that of fire and wheels. However, the full potential applications of AI are yet to be discovered, as Chamath Palihapitiya has also stated. Therefore, it may be too early to understand or measure the true impacts of AI. Although we should be thinking about regulating it in the future, for now, we should keep an open mind towards technologies that can potentially disrupt and improve society. But I'll say one thing, my fellow readers, the more I read about it, the more I want to buy a house immersed in nature to disconnect from time to time. 👩🏻🌾
& I'm also reading... 🔗
Grimes Tells Fans To Deepfake Her Music, Will Split 50% Royalties With AI
Berkeley researcher deploys robots and AI to increase pace of research by 100 times
~ & maybe a few other things… 🎨
Although this newsletter is titled Tech Digest, I'll also be including a non-tech-related article or fun fact as a bonus. As I mentioned my love for the intersection of tech and art earlier, I was reminded of an article that I'm particularly fond of and would like to share with you all. It's about how Caravaggio's paintings influenced the aesthetics of modern cinematography, including the work of Martin Scorsese. 👽
// Caravaggio: A Pioneer of Modern Cinematography?
The article discusses how the dramatic lighting and composition in the paintings of Italian Baroque artist Caravaggio influenced the aesthetics of modern cinematography. Caravaggio's contrast between light and dark, sacred and profane, created a sense of depth and emotion in his paintings, similar to the way lighting and camera angles are used in film. The author also highlights specific examples of how Caravaggio's works have been directly referenced in film, such as in Scorsese's "Taxi Driver" and Ridley Scott's "Gladiator." The article concludes by noting how Caravaggio's contributions to art continue to influence other fields, including film, centuries after his death. Cool, right? 👽
~ & that's all folks… 🥕
Well, as we come to and end, I'd like to thank everyone who stuck around and read it until now. I'll keep improving this format and I'd love to hear some feedback from all of you. But if you liked Tech Digest so far, make sure to subscribe & share it with your friends. ⚡
Bisous,
Isabela Mac
Very good