Are We Safe with the Latest AI?
Are We Safe with the Latest AI?
How does the development of AI technology impact the market?
A proper understanding of AI is more like thinking about electricity than cryptocurrency. When electricity was introduced, it took time to see everyday applications like cell phones, but it never went away. People imagined a future with flying cars, but not all those dreams came true. Similarly, AI is not a fad; it’s here to stay and is improving every aspect of business. The direct impacts of AI are already evident and don't rely on new inventions. Everything mentioned so far is achievable with current technology.
However, there are also those who envision AI becoming so advanced that it will outperform humans, merge with machines, cure diseases, and extend human life. This is similar to how people in 1900 thought we’d be living on the moon by now. This aspirational view often leads to inflated expectations, similar to the Gartner Hype Cycle's "peak of inflated expectations." After this peak, we may enter the "valley of disillusionment" when people realize AI is not advancing exponentially in every direction.
AI currently feels like magic, much like electricity did when it first appeared. Some view it with skepticism, considering it almost mystical. We must distinguish where we are now—just beginning to explore the practical applications of AI—and recognize that it’s a technology with immediate and future impacts. If you don't embrace the existing AI advancements, you risk being left behind in the next few years. The current technology is incredibly powerful, but we have only scratched the surface of its potential. The key is to stay agile and ready to adapt.
How is AI improving decision-making and workflow automation?
When you learn programming or computer science, your brain rewires itself to think differently. Interestingly, Mark Zuckerberg recently announced a new version of "Go Code Llama," their code generation AI model, which they are open-sourcing. He mentioned that integrating this model into the next version of Llama, Llama 3, shows that training AI to think like coders improves its reasoning abilities. It not only helps humans make better decisions and think differently but also enhances AI models' reasoning capabilities.
Everything in life is a process or workflow. A cooking recipe is a process, for example. When you think like a coder, you see a business process and naturally think in terms of "if this, then that." Traditional automation tools use hard-coded rules, but life is unpredictable and random. AI models, for the first time, can execute workflows with intelligence that allows them to navigate these unpredictable situations. They know when to ask for help, when to use a tool, and can recognize spam emails even without specific keywords. This ability, along with other factors, is fundamentally changing what can be automated in the workplace.
What does the future of AI look like?
I won't be the one stopping our robot future from happening. However, constrained alignment ensures that the AIs you use can't go beyond their allowed workflows or capabilities. They can't operate outside their defined parameters and can only pass inputs and outputs to the next approved step. At any point, a human controls the workflow and decisions, with full transparency at each step of the AI's process.
We won't be stopping "iRobot" scenarios because the future won't look like the movies. In the movies, you can stop a rogue AI by shutting down a server or robot. Instead, I see the web evolving as a collective consciousness with intelligent orchestration. The internet itself is like Skynet, where every connected computer and device is a part of it.
One website may specialize in writing, another in editing, another in posting, and another in math or project management. A sufficiently advanced AI could know when to use each website's skills to solve parts of a problem and check if it was done correctly. This creates an autonomous AI that can leverage the entire internet. In this way, the internet becomes a collective consciousness and, in a sense, is alive.
Just as the human body removes dead cells, outdated websites disappear, and new ones emerge. Websites improve over time, becoming stronger and more capable. Web3 focuses on making websites, domains, and companies more accessible to AI models, which will become the main consumers of the internet. You will sell more of your services to AI agents than to humans browsing your site.
AI agents will read websites, purchase services, and act on behalf of users and businesses. This is Web3, the Executable Web—a constantly expanding, improving, and optimizing collective consciousness. To unlock this, we need an AI that can orchestrate that intelligence.
What steps should businesses take to integrate AI and stay competitive?
At a minimum, you need to explore how you and your organization can start an AI transformation. Even if humans are better at certain tasks today, if you haven’t explored AI's potential, you're falling behind. Things change quickly, and if you haven’t done the basics, you won’t be ready for the next step.
Many software services, like Canva and Bing, already integrate AI, often at no extra cost for basic versions, so there is no excuse not to start. For example, my child had trouble with an assignment because he didn't understand what the teacher meant. We took a picture of the assignment, fed it to AI, and asked it to interpret what the teacher might mean. It gave a coherent answer. This shows the power of AI—it's like having the world's knowledge in your pocket.
If you're already doing the basics and want to go further, consider agent engineering. We created SmythOS.com with the future in mind, where three people can manage 100 AI agents, multiplying human productivity. Agent engineering—building, operating, and maintaining AI agents to perform tasks at scale—will be a critical skill. I recommend training a few people in your organization to become agent engineers and start automating every part of your business. If you don't, you may find yourself outpaced by competitors who have done so, and suddenly, they can offer services at a fraction of your cost.
If you don’t have the capacity to create AI solutions yourself, become a consumer. We're likely to see agent stores emerge, replacing platforms like Upwork and Fiverr. Instead of hiring a physical expert, you could interact with an AI agent that has their knowledge and can handle most tasks while the expert is away. If you aren’t developing AI solutions, at least upgrade to AI-powered tools like GPT for Teams. This allows each department in your business to use AI to communicate and operate more efficiently.
Step one is for everyone, including your children, to learn how to use AI responsibly and effectively. Step two, if you're ambitious, is to learn agent engineering or partner with a service like SmythOS. Finally, if you don’t plan to build custom solutions, leverage emerging app stores and extensions to amplify productivity. At the very least, be on the front lines of consuming these innovations.
Why is understanding and embracing rapid technological progress essential for long-term business success?
Humans perceive time in a linear way—the sun rises and sets and the moon goes through its phases. We think of time as moving in a straight line, which shapes how we anticipate problems. For example, if you play basketball with friends who always pass the ball softly, you get used to it. If one suddenly throws it fast and strong, you won’t be ready to catch it.
This is like the acceleration of progress—it’s hard for us to grasp. Progress can come at us rapidly, like that unexpected pass, and we’re not ready. In one year, humans now produce as much knowledge or content as all of written history combined, and this rate is increasing exponentially, thanks to AI. Our brains aren’t prepared to understand or accept this speed. Even my predictions about the market often happen faster than I can imagine.
For software startup founders, agencies, and other businesses, it’s a challenging time. The pace of change and the need to adapt quickly can be exhausting. The companies that will survive are the ones that are most agile, lean, and adaptable. There's no point in resisting this future; you need to embrace it fully. If your business is not lean, you need to trim down. If you’re not using these technologies, you risk becoming irrelevant in a few years—not because you don’t provide a good service, but because competitors can do it at a fraction of the cost.
The key lesson is that while people can adapt and will achieve great things in the future, the transition will be tough. Those who stay adaptable, lean, and open to change will thrive.
*This interview has been edited and condensed for clarity.*
