Artificial Intelligence: AGI or just AI? In the realm of artificial intelligence (AI), two terms often arise: Artificial General Intelligence (AGI) and specialized AI tools. AGI refers to the pursuit of creating highly autonomous systems that possess human-like cognitive abilities, capable of performing a wide range of intellectual tasks. As of today, true AGI doesn’t exist and there are various predictions when this could happen, some as soon as end of 2023 or as late as 2059. In contrast, specialized AI tools are designed for specific functions, such as image recognition or recommendation engines, without aiming for general intelligence. Chat GPT, autonomous vehicles, medical diagnosis, and smart home devices are some examples of this. I use ChatGPT in my daily routine. In fact, it helps me with writing projects. I also produce videos using AI-powered video apps. Everything can be done faster than ever before. It’s like having a very smart assistant. The end composition however takes a human to bring it all together to put it into context, add some personal flavor, and draw upon your experience and knowledge. AI affects energy use, the economy and the environment Using AI requires energy, most of which is consumed by computing. Compute refers to the processing power required for AI algorithms to perform tasks, while storage relates to the capacity to store and retrieve data efficiently. These two elements are fundamental in AI systems, as the availability of computational resources directly impacts the speed and complexity of AI operations. You may have noticed that most of the new AI apps have a free version, but you are lowest priority when it comes to performance and paying subscribers. I subscribe to ChatGPT3 & 4 for this reason. I tried out Midjourney but I will need to subscribe if I’m ever going to get to use the system because the free version doesn’t allow me to engage. And DALL-E has a system where you buy credits as you go in order to generate images using AI prompts. As we all start using AI, the increasing demand for computational power is increasing and it poses a challenge due to the substantial energy consumption involved. The world is in a race to see who can mobilize and concentrate energy levels the fastest since computing for AI systems is growing at such a breakneck speed. So it’s not just about energy, it’s about the economy. Who will control our future? It's also about control. Currently, most of the energy levels are in centralized data centers with major companies and the highest costs are for the servers in these locations. Control lies in the hands of the companies that are storing our data. As we all know, there are many views on how much control or authorization companies have to acquired data from individuals. These companies set the terms and price for these computational systems. What’s the solution? Decentralized computing networks offer a potential solution to be more efficient with storage and computing. It also eliminates central control. It may sound like a dream, but it’s already in use and it’s evolving. With blockchain, by distributing computing tasks across a network of decentralized nodes, the burden on individual centralized systems is alleviated, leading to enhanced efficiency and reduced energy consumption. One notable example of decentralized technology is blockchain. A blockchain is a decentralized and transparent digital ledger that records transactions across multiple computers. To participate in a blockchain network, individuals acquire tokens, which serve as the native currency of the blockchain. You can obtain digital tokens through exchanges or transactions with other users. Just like cash or credit card, you store and manage your tokens in a digital wallet. This provides you with a unique address for sending, receiving, and securely storing tokens. In financial investing for instance, you can invest in a cryptocurrency like Bitcoin that operates on a decentralized network using blockchain. A good way to learn is start small and invest in crypto. And blockchain technology isn’t only for financial transactions. It’s also used to record transactions between different members of a supply chain, providing visibility and the ability to track and trace transactions across an entire supply chain network. It can also identify the origin and authenticity of products. What about the environment? When you hear someone talking about how AI is a threat to the environment, it may not be readily obvious what they mean. They are referring to how much energy is required to run large-scale data centers for AI computing and storage. These require vast amounts of electricity that contribute to carbon emissions, use water for cooling, and also generate e-waste (outdated end of life hardware). Summing it all up 1) The distinction between AGI and specialized AI lies in the scope of capabilities, with AGI striving for human-like intelligence. Now is a good time to start learning by using AI. 2) Compute and storage are integral components of AI systems, with the race to meet energy demands driving innovation. This is important to be aware of as we elect officials and draft policies that drive innovation and yet protect the environment. 3) Decentralized computing networks, such as blockchain, offer potential solutions to the energy challenge by distributing computational tasks. This will affect future control and democratization of the internet. 4) Participating in a blockchain involves acquiring tokens and utilizing a digital wallet, enabling secure transactions and engagement within the blockchain ecosystem. It’s easy to get involved with blockchain by investing in cryptocurrencies. My approach is to start early, learn by doing, create a plan, get a pilot going for yourself or your organization (however basic), modify and evolve and get some initial wins with proof of concepts. Every business is different and so is every culture and creating a culture of innovation is what sets apart companies that are growing/changing versus a wait and see approach.
0 Comments
Leave a Reply. |
Michael Richter-authorMichael has over twenty years of experience including global marketing, strategy & executive producer roles. He is also an adjunct professor at Thunderbird School of Global Management. Categories
All
Archives
January 2025
|