🤝 How AI could soon be earning for you
Dear curious mind,
In our previous issue, we explored how AI enabled almost anyone to create digital solutions. Today, we're taking this concept a step further. We'll delve into an exciting frontier where AI agents don't just help you build apps — they could potentially generate income on your behalf by solving paid tasks or host a created digital solution they created for you.
In this issue:
💡 Shared Insight
From Sci-Fi to Reality: AI Agents are Becoming Economic Players
📰 AI Update
Replit Introduces an Agent Which Codes and Deploys
🌟 Media Recommendation
From Autopilot to AI Education: Andrej Karpathy on No Priors Podcast
💡 Shared Insight
From Sci-Fi to Reality: AI Agents are Becoming Economic Players
Years ago, I read a book that left an indelible mark on my imagination: "Daemon" by Daniel Suarez. There was an AI that solved tasks on platforms like Amazon's Mechanical Turk, earning money, which it then invested in additional computing resources. This idea of a self-improving AI, bootstrapping its way to greater capabilities, has stuck with me. Today, as AI rapidly advances, I think it might become reality rather sooner than later.
Let’s take a step back to where this story begins: Amazon's Mechanical Turk (MTurk), a platform where people can get paid for completing small tasks that computers still struggle with—like labeling images or answering surveys. The name "Mechanical Turk" has its roots in a famous 18th-century hoax where an automaton appeared to play chess against human opponents, when in fact a hidden human was controlling it [Wikipedia article]. Today, Amazon Mechanical Turk is a digital marketplace where workers take on microtasks to get paid.
But here’s the fascinating twist: it's no longer just humans who are working on these platforms, even if the platform owners state this as a requirement. So far, automated bots have been limited to tasks like scraping data or automating repetitive processes, but more advanced AI agents are emerging, capable of handling increasingly complex jobs.
The difference between bots and agents is subtle yet crucial. Bots are typically pre-programmed to execute narrow, repetitive tasks. In contrast, AI agents, can learn, adapt, and make decisions based on their environment. With these advances, we’re seeing AI agents being able to earn real income for people.
As AI agents continue to evolve, the concept of an AI earning money — like in the sci-fi vision described in “Daemon” by Daniel Suarez — doesn't seem so far away anymore. With more capable agents at our fingertips, the future of AI as an active, productive part of the economy might be closer than we think.
But besides large cooperations also individuals could "employ" AI agents to perform tasks or generate value. We will see the rise of AI-human partnerships, where people act as managers or supervisors for a team of AI agents working across various domains. This could democratize access to specialized skills and services, allowing individuals to offer complex services traditionally reserved for larger companies.
📰 AI Update
Replit Introduces an Agent Which Codes and Deploys
Replit, the online coding platform, has unveiled its new "Replit Agent" in Early Access. This AI-powered tool aims to revolutionize app development by handling the entire process from ideation to hosting.
CEO Amjed Madani demonstrated the agent by creating a location-based landmark app in just two minutes, showcasing its ability to rapidly prototype ideas on any device. The tool works seamlessly within Replit's existing environment, giving users full access to and control over the generated code.
Already on its release day, there are many posts on 𝕏 (thread showing seven projects) which demonstrate its capabilities.
The Replit Agent feature is currently available in early access on all paid Replit plans, which start at $10 per month.
My Take: The real novelty of Replit Agent lies in its end-to-end approach. While AI-assisted coding tools are becoming more common, Replit takes it a step further by not only generating code but also handling deployment. The ability to go from a simple prompt to a fully deployed, cloud-hosted application in minutes could be a game-changer, potentially democratizing the app creation process.
🌟 Media Recommendation
From Autopilot to AI Education: Andrej Karpathy on No Priors Podcast
In episode 80 of the No Priors podcast, Andrej Karpathy, former Tesla Autopilot leader and OpenAI founding member, shares his expertise on autonomous driving and generative AI.
Despite Waymo's apparent lead, Andrej believes Tesla will ultimately win the self-driving race. He argues that Tesla faces a software challenge, while Waymo has a hardware problem. Tesla's advantage lies in its vast fleet of vehicles collecting real-world data globally.
Andrej discusses his new venture, Eureka Labs, which aims to revolutionize education using AI. He envisions AI acting as a front-end interface for students, potentially providing personalized, adaptive learning experiences at a global scale.
For young people preparing for the future, Andrej recommends focusing on mathematics, physics, and computer science. He believes these disciplines provide the best foundation for developing critical thinking and problem-solving skills in both pre- and post-AGI worlds.
Disclaimer: This newsletter is written with the aid of AI. I use AI as an assistant to generate and optimize the text. However, the amount of AI used varies depending on the topic and the content. I always curate and edit the text myself to ensure quality and accuracy. The opinions and views expressed in this newsletter are my own and do not necessarily reflect those of the sources or the AI models.