- Tech Tales & Tactics
- Posts
- AI Just Changed. Have You Noticed?
AI Just Changed. Have You Noticed?
AI is forcing us to rethink the role of humans altogether.
Sebastian Muehl
February 13, 2025
We think of AI as a tool. How can this tool help me in my current task? Something we pull out when needed—a sidekick.
Most AI is used like C-3PO — it talks a lot and gives us advice, but it is ultimately useless in a fight. AI Agents? They’re like R2-D2 — hacking doors, flying the X-Wing, and keeping control while humans panic.
But what happens when people see AI as an end-to-end actor in our workflows? A new teammate? What if AI stops waiting for commands and starts completing entire jobs end-to-end?
This isn’t a thought experiment. AI Agents are emerging, and they aren’t waiting for us to catch up.
It’s a world full of opportunity and significant change. And it isn’t coming someday. It’s already happening.
So it’s time we look at AI Agents rather than AI Models.
💡What if we shift focus from AI Models to AI Agents?
AI Models (ChatGPT, Gemini, Claude), as we know and talk about them, are reactive. They take our input and generate an output. We decide when to activate them and when we leave them out. Most often, we do not even consider AI models in workflows (yet).
AI Agents go beyond responding. Instead of waiting for our input, they can be assigned to a field of work and act proactively. Imagine an agent monitoring the sales funnel of an e-commerce website. The agent would interpret the data and see that a green buy button performs worse than a blue buy button. While analyzing the data, the agent can make changes to the sales funnel in real-time.
AI Agents work 24/7, don’t get sick, and don’t complain about new processes or tools. And they don’t ask for pay raises or titles (yet). And that’s the reason why big companies are already making them part of their workforce.
Cultural Readiness
We are always talking about tech readiness for AI. And yes, it is a valid point, as running models that power these agents is still costly. However, is anyone addressing cultural readiness?
Are we and, are companies, ready to hand off a full role to an AI Agent? Can we trust them? What happens if things go wrong? What about liability and regulation? It's not that we slow down progress, but we must start thinking about the cultural impact and readiness.
AI Agents will revolutionize the traditional business order.
AI is not a tool; AI is completely reshaping the role of humans. Are we culturally ready for that?
Regulation
While regulators are catching up on copyright violations, AI Agents are taking on end-to-end work tasks, and nobody knows how to handle if things go wrong, how we will handle liability, who is to blame for error?
The First AI Agents are Emerging
Open AI’s Operator Mode - An Agent that can go to the web to perform tasks for you. It can use its own browser, interact, click, type, and do everything you can do on a website.
Imagine the revolutionary leap here. We now have AI that can handle all the tasks humans can do on websites. Browse, fill out forms, collect specific data, and work with complex interfaces. It opens up a world of new workflows that it can own end-to-end.
Sidenote - How does OpenAI do this?
OpenAI’s “Operator” is powered by a brand new model. It is called computer-using-agent (CUA).
The magic: It combines computer vision capabilities with advanced reasoning through reinforcement learning. So, the model is directly trained to interact with user interfaces. It understands every button, text field, and graphic we see on the screen.
OpenAI is already working with DoorDash, Instacart, Uber and others.
Operator is still in the early stages and has a fallback mode in case it gets stuck.
If you are craving full technical detail, please dive in on the OpenAI Research Blog
Adept’s AI ACT-1 and Amazon AGI
The belief in AI Agents is strong in Amazon. They just recently opened a new AI Lab here in San Francisco. This Lab will focus solely on developing foundational capabilities for AI agents.
Amazon recently hired executives from Adept. A company that has focussed on AI Agents since 2022. Bringing the experience from Adept together with Amazon’s most recent Nova model seems to promise functionality similar to that of OpenAI’s Operator.
Amazon showed some impressive Nova demos on how the model understands images and videos as prompts. Incredibly impressive.
Model output: “The video depicts a football game in progress on a green field. Players from two teams, one in yellow uniforms and the other in white, are engaged in a play. The quarterback from the yellow team throws a pass to a receiver, who catches the ball and begins running downfield. Defenders from the white team pursue him, attempting to tackle him. The play culminates in a tackle, bringing the receiver down on the field.”
Once AI learns to “see” and truly understand what it’s looking at, it can take over operational workflows—leaving creative problem-solving as one of the last frontiers for humans.
Conclusion
Over the last few years, AI has found a way into everyone’s toolbelt as a sidekick. An assistant, waiting for new instructions before they get active.
But what needs to be understood is that we are moving towards a world where AI does not just support work—it owns it.
The question isn’t just how good these agents will be but how comfortable we are with letting go.
Once AI stops waiting for commands and starts to act, we are rewriting the rules of work itself. A structure that has been built over centuries is going to change.
Are we ready for this shift?
Not that it matters. Because while we’re debating, a lot of smart people are working to make it real — and nobody is waiting for permission.
What This Means for Us
For my fellow product folks: AI agents are not just features — they’ll own workflows end-to-end. Companies must rethink the entire user journey, not just add AI on top.
For founders: Could AI make your SaaS redundant? Which parts of your business can AI take on autonomously? Optimize lead funnels? Run A/B tests? Optimize pricing?
For engineers: Be aware that the UI you might build is no longer just for humans. AI agents are the new users. Are your interfaces AI-ready?
🎧Playlist of the week
If you have some big meetings coming up, pull up this playlist. No doubt, it will help you go Kendall Roy Supersonic in your next meeting. 🤘
How do you feel about AI Assistants? Would you feel comfortable to bring into your day-to-day work? Or maybe have it book a vacation for you?
Have a great rest of the week,
How did you like this edition? |
Reply