π€ Stop using AI like a human and start thinking exponentially
Why Your Human Approach to AI is Holding You Back
Dear curious mind,
The main story in this issue challenges us to think differently about how we interact with AI tools, moving beyond our linear, human-centric approaches to fully leverage AI's parallel processing potential. As we are rapidly moving towards an era of AI agents that can perform complex tasks autonomously, it's crucial to understand how to harness these technologies to our advantage and not limit ourselves.
In this issue:
π‘ Shared Insight
Break Free from Human Limitations: The Power of Nonlinear AI Thinking
π° AI Update
Llama 3.2 Vision Usage Made Easy with Ollama Support
FLUX1.1 [pro] Now With 4x More Details and More Natural Photography
π Media Recommendation
Video: Sam Altman Says That Reasoning is OpenAI's Next Frontier
π‘ Shared Insight
Break Free from Human Limitations: The Power of Nonlinear AI Thinking
Many of us approach AI tools like ChatGPT with a mindset constrained by our human limitations. We ask for one solution, wait for the response, evaluate it, and if unsatisfied, ask for another. This linear thinking pattern is deeply embedded in how we operate, but it's time to break free from these self-imposed constraints.
I fall into this trap myself constantly. Just for the previous issue, I was writing this newsletter and needed a catchy title. I asked AI for a suggestion. When it didn't quite hit the mark, I asked for another. And another. I was stuck in this inefficient loop of requesting one title at a time, each time waiting for and evaluating a single response. However, this method significantly underutilizes AI's capabilities. Instead of requesting one title, why not ask for multiple different options, along with explanations of their potential impact on reader engagement? The AI doesn't tire or slow down when generating multiple options β it's not bound by human cognitive limitations like I am.
With the capabilities of generative AI we have access to today, we are the curators of AI output, we are modifying ideas and combining concepts. We use AI as a source of inspiration. Working effectively with AI isn't a passive process of accepting whatever it generates. At the same time, we aren't constrained by human limitations like patience or endurance when collaborating with AI. The key is to break free from the human limitations and let AI's capabilities expand our creative horizons.
All major AI companies - from OpenAI and Anthropic to Google and Meta - agree that AI agents which can perform tasks over extended periods are the next frontier. Sam Altman, CEO of OpenAI, provides an illuminating example which shows that we have to take care to not apply the same human limitations to AI agents. Instead of calling one restaurant for a reservation, an AI agent could simultaneously check hundreds of restaurants to find the perfect match based on your preferences, available times, and special offerings. While this would be overwhelming and time-consuming for a human, AI can handle these parallel tasks effortlessly.
Professor Ethan Mollick suggests in his book Co-Intelligence that you should "Invite AI to everything you do." But as his π post in the image above highlights, you should take care that you don't just invite AI as a helper bound by human constraints β embrace its ability to operate at scales and speeds beyond human capability. When using AI:
Think exponentially: Instead of requesting one solution, ask for dozens or even more suggestions
Request detailed analysis: Have the AI explain the reasoning behind each suggestion
Leverage parallel processing: Let AI explore multiple approaches simultaneously
Provide comprehensive context: Share your goals and constraints to get more targeted results
The future belongs to those who can think beyond human limitations and fully harness AI's potential for nonlinear, parallel processing. Whether you're writing, researching, or problem-solving, remember: Your AI assistant isn't constrained by human bandwidth β don't limit it to human-scale tasks.
π° AI Update
Llama 3.2 Vision Usage Made Easy with Ollama Support (Ollama blog)
The privacy-first approach of running Llama 3.2 Vision locally is undoubtedly what many of us have been waiting for! From OCR and handwriting recognition to general image understanding - having these capabilities without sending sensitive data to the cloud is a game-changer.
FLUX1.1 [pro] Now With 4x More Details and More Natural Photography (BlackForrestLabs blog)
Finally, some fresh European AI power has emerged. Black Forest Labs continues to prove that you don't need Silicon Valley to push the boundaries of artificial intelligence.
π Media Recommendation
Video: Sam Altman Says That Reasoning is OpenAI's Next Frontier
In a recent interview with Harry Stebbings of 20VC at OpenAI Dev Day, Sam Altman shared insights on OpenAI's strategy and his perspectives on AI development.
OpenAI is focused heavily on reasoning capabilities as their strategic priority, believing it will unlock major advances in science and coding.
Altman cautions startups against building products that just patch current model limitations, as future iterations will likely address these shortcomings.
He predicts trillions in new market value will be created by AI-enabled products and services that were previously impossible.
My take: This interview provides food for thought for founders and investors in AI. Altman's emphasis on reasoning capabilities and warning about short-term fixes suggests entrepreneurs should focus on building products that become more valuable as AI models improve, rather than trying to patch current limitations.
A big thanks to Manoj for making me aware of this interview in our latest π space, which by the way happens every other Tuesday (uneven weeks) at 8 PM UTC on π. Join our discussion and share your thoughts, or just listen in. All are welcome!
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.