» Die Themen der Folge 231:
Heute mit automatischer Zusammenfassung via AI:
Generative AI: Hype, Security, and the Race for AI Technology
Generative AI is a complex technology that is subject to hype cycles and is potentially difficult to deploy in production. Companies investing in AI must consider data availability, privacy, and security, while Generative AI provides opportunities for working with unstructured data. AI-based technologies have become a strategic resource, with NVIDIA monopolizing the market for the chips required for training, and countries and tech giants racing to acquire them. This situation has driven Nvidia’s stock to astronomical heights, and regulation is being debated in order to control the use of AI.
- NVIDIA has monopolized the market for the chips required for training AI technologies.
- AI-based technologies have become a strategic resource that countries and tech giants are racing to acquire.
- Regulations are being debated in order to control the use of AI, leading to significant impacts on the stock market.
Last week, various topics were discussed related to Generative AI, Gartner Hype Cycle, S&P Global Survey on AI use, Entropic’s custom LLM for telecoms, US law on copyrighting AI-generated art, NVIDIA’s GPUs, states‘ rights over AI, Digital Services and Digital Market Acts, Alien’s stock crash, Twitter’s uncertain future, and Elon Musk’s comments. Mercedes is further ahead than Tesla in self-driving car technology, according to the Gartner Hype Cycle. This cycle consists of five phases, from a technology trigger to peak of inflated expectations, leading to disillusionment, enlightenment, and productivity
Bill Gates famously said that the short-term effects of technology are often overestimated and the long-term underestimated. AI is often subject to hype cycles, where expectations are initially high and then disappointments lead to market corrections. AI is often seen as the technology that computers cannot do – when it works, it just becomes software
Chatbots have been around since 1966, and AI has gone through cycles, where big leaps forward are followed by periods of stagnation. Despite this, AI is still making progress, and GPT is a fascinating example of how quickly technology can progress – although people should not rely solely on it. 70% of companies surveyed are using AI, but only 28% have it in production
This is due to the complexity of AI and Generative AI, and the potential for mistakes when deploying it without proper testing. AI projects are defined as tasks that computers cannot yet do, but a major hurdle is data management. Other challenges include company workflows, privacy, and second-order effects such as CO2 footprints
These issues have caused some disillusionment as companies expected a quick impact. Companies investing in AI must consider data availability, privacy, and security, while Generative AI provides opportunities for working with unstructured data. A South Korean telecom invested in a startup to build a specialized language model, but it’s still unclear which layers of AI will provide a competitive advantage
Generative AI and intellectual property is still a debated topic, as demonstrated by a recent court decision that artwork created by AI is not copyrightable. This raises questions about how art created using AI will be protected and whether it is unique enough to be copyrightable. Companies using AI must also consider the rights to the training data used in the process
AI is raising questions about copyright protection as courts and companies grapple with who owns the rights to works created with generative AI, and artists such as Stephen King and the New York Times have raised legal challenges. AI is also used in music and medicine, and even photos taken on an iPhone are impacted by AI, leading to more questions about who owns the rights to works created with AI. AI-based technologies have become a strategic resource, with NVIDIA monopolizing the market for the chips required for training
Many countries and tech giants are racing to acquire these chips, with 95% of venture capital raised by startup Inflection AI going towards purchasing them. Embargoes and resource scarcity are making it difficult for smaller startups to acquire these chips. Taiwan is a key global chip producer, providing 80% of the world’s chips
In response, Germany is funding a TSMC factory in Dresden, which poses a risk to Taiwan as it reduces the world’s dependence on Taiwanese chips. This situation has driven Nvidia’s stock to astronomical heights due to their GPUs being key to the development of generative AI. Nvidia is expected to have a great quarterly result which could shape the entire tech industry
There is a debate about whether AI should be under state control, with some suggesting a similar model to that used for nuclear energy, however, this poses difficulty due to its international nature and the fact that AI is a black box which is not easily understood. Regulation could have unintended consequences, as seen with GDPR, which has not been effective for smaller tech companies. EU regulations are creating a deep level of compliance costs for tech companies that can be difficult to manage
The Digital Services Act and Digital Markets Act are set to be implemented in early September, and will have a deep impact on tech companies‘ business models and architectures. Examples include Google having to choose which browser to set as default, Apple having to allow users to install apps, and TikTok having to offer non-personalized recommendations. The EU is introducing strong regulations on Big Tech companies to limit their ability to personalize data and to prevent the spread of Chinese government propaganda via ads
These regulations carry heavy penalties and could even lead to breakups if companies don’t comply, though some view them as a positive step towards global regulation. Big Techs face new times, with one EU unicorn, Adyen, having difficulty due to slower growth and competition from Stripe and PayPal. Adyen was once a first mover, making it easier for developers to integrate payment options, and had a 900 million dollar exit
Alien, the second largest company listed on the Eurostox 50, has seen its stock price drop 40% after poor results. Twitter also has its own problems, with its „Free Speech“ platform introducing penalties for linking to certain media outlets, which could lead to a high abandonment rate. Elom Musk recently announced that Twitter will no longer be allowing blocking, which could have a negative impact on those who experience bullying and harassment on the platform
Additionally, an analysis of his followers revealed that 40% of them have no followers or tweets, suggesting that many are fake or inactive accounts. These changes have also had consequences for app store rankings, as X is ranked at #52 on Android, while Threads is ranked at #2. In this episode of the podcast, the speakers discuss the dramatic developments of X, the reality of bots solving captchas better with AI and the deflation of the Venture Space
They also recommend the book „Infinite Powers“ by Steven Strogatz which is a great read that covers the history of mathematics and its fundamental principles. In this podcast, they discuss the book „Infinite Powers: How Calculus Reveals the Secrets of the Universe“ by Steven Strogatz. They discuss how calculus can help understand the world of technology and help overcome the trauma of math from school
A link to the book and other resources are in the Show Notes.