- Brain Scriblr
- Posts
- A Dive into Qwen
10 News Items This Week
How Shell is harnessing AI for the energy sector.
Can AI help cover this ‘green blind spot’ and help ESG initiatives survive?
Anthropic fused reasoning into its models (a new one coming soon), challenging OpenAI’s split approach—now OpenAI is following suit.
Perplexity has released their own free to use Deep Research tool.
Cursor is the fastest growning SAAS product ever. See the chart. I personally at this point don’t use Cursor as I use Windsurf and AIder instead, nevertheless the growth of Cuirsor is impressive.
This is an interesting discussion from Harvard on how sales teams can and are using AI.
Likely the most interesting read this week is on how AI is affecting the way we make decisions.
Continual pre-training on Hephaestus-Forge improves the resulting model's agentic capabilities, with scaling law experiments identifying an optimal data mix ratio of approximately 1:1:1 for agent, code, and text data.
Eudia a legal AI startup has raised $105 million in series A funding.
LangChain published a study on the performance of AI and the gaps in human performance.
AI For Good
Even though commercial whaling has largely diminished, whale populations globally continue to be threatened by various human activities, especially shipping collisions. This growing concern has prompted scientists to call for improved marine management strategies to mitigate these risks.
Researchers at Rutgers University have developed an innovative tool powered by machine learning designed to predict endangered whale habitats, thereby helping ships avoid fatal collisions with these majestic creatures.
The model integrates two comprehensive datasets accumulated over more than three decades: one consisting of satellite imagery and another comprising data collected by underwater gliders—autonomous vessels dedicated to data collection.
Initially intended to assist offshore wind developers, the researchers' goal was to create a system that could support their projects. However, the tool has proven to be invaluable in informing conservation strategies and promoting responsible ocean development.
"With this program, we're correlating the position of a whale in the ocean with environmental conditions," explained Josh Kohut, a Rutgers professor of marine sciences. "This allows us to become much more informed on decision-making about where the whales might be.
We can predict the time and location that represents a higher probability for whales to be around. This will enable us to implement different mitigation strategies to protect them." This program represents a significant step forward in the efforts to protect endangered whale populations and ensure their survival for future generations.
Prompt
Cinematic image of a woman in a phone booth
Cudo Compute is a cloud-based service provider that offers high-performance computing, AI, and deep learning solutions. |
Dubsado is great for contract writing and project management. |
Folk is the number one AI powered CRM tool. |
N8N is the most powerful automation tool |
Mental Shortcuts and How Heuristics Shape Human Thinking
Human reasoning is often celebrated as a hallmark of our intelligence, but what if much of what we call “rational thought” stems from mental shortcuts? Research across cognitive psychology, philosophy, and artificial intelligence reveals that humans are fundamentally heuristic thinkers. Instead of engaging with independent, logical reasoning, we depend on patterns learned through experience to navigate decisions quickly and efficiently. This approach, while adaptive, comes with trade-offs—our reliance on heuristics often leads to biases and errors, challenging the notion of humans as purely rational agents.
Cognitive psychology provides compelling evidence for this view. Pioneering work by Tversky and Kahneman shows that people use heuristics like representativeness, availability, and anchoring to make judgments under uncertainty. These mental shortcuts allow us to process information rapidly but can result with systematic deviations from logical or statistical norms.
An example might be the representativeness heuristic leads us to judge probabilities based on similarity rather than base rates, while the availability heuristic skews our perceptions based on how easily examples come to mind. Such biases underscore that our default mode of thinking is not analytical but intuitive, shaped by prior experiences rather than abstract reasoning.
Philosophy and neuroscience further support this perspective. Thinkers like David Hume argued that reason is often a servant of passion and habit, while modern neuroscience suggests that our sense of free will and rational agency may be illusory.
Studies show that decisions are often initiated unconsciously, with conscious reasoning serving to justify actions after the fact. This aligns with the idea that much of what we consider “rational thought” is actually post-hoc storytelling, driven by ingrained patterns rather than spontaneous logic.
Artificial intelligence offers a fascinating parallel. Machine learning systems, like humans, excel by recognizing patterns within data rather than applying formal logic. AI models can outperform humans during tasks requiring consistency and unbiased analysis, but they also inherit human biases when trained on flawed data. This comparison highlights both the strengths and limitations of heuristic thinking: while it enables efficiency and adaptability, it can also perpetuate errors and blind spots.
Understanding the heuristic nature of human cognition has profound implications. It challenges us to rethink education, decision-making, and even moral responsibility. Rather than expecting individuals to overcome biases through sheer willpower, we can design systems—like “nudges” or AI-assisted tools—that guide intuitive thinking toward better outcomes. Recognizing our cognitive limitations also fosters humility, reminding us that rationality is not an innate trait but a skill honed through practice and supported by external structures.
Within a world increasingly shaped by technology, this perspective invites collaboration between humans and machines. By leveraging AI to handle tasks where human biases are problematic, we can focus on areas where our intuitive strengths—creativity, empathy, and contextual understanding—shine.
Embracing the heuristic mind is not a critique of human intelligence but a call to understand it more deeply, designing systems that complement our strengths and mitigate our weaknesses.
The heuristic mind is a testament to our adaptability, but also a reminder of our fallibility. By acknowledging how we think, we can make better decisions, build smarter systems, and navigate an increasingly complex world with greater clarity and purpose.
For more on this read Bejamin Libet’s work online and his book Mind Time
4 Key Things to Know About Qwen
Unlocking the Power of Qwen: 4 Key Things You Need to Know
The world of artificial intelligence is advancing rapidly and large language models (LLMs) are at the forefront of this transformation. Among these models Qwen developed by Alibaba Cloud stands out as a powerful tool for natural language processing. Whether you're a developer a business professional or simply curious about AI understanding Qwen can help you harness its potential.
Here are the four essential things you need to know about Qwen to get started.
What Qwen Is
Qwen is a state-of-the-art large language model designed to handle a wide range of natural language tasks. It can generate human-like text translate languages summarize documents and even power conversational AI systems.
What sets Qwen apart is its integration with Alibaba Cloud’s ecosystem. This makes it a versatile and scalable solution for businesses and developers. Whether you're building a chatbot or analyzing large datasets Qwen is built to deliver high performance and accuracy.
How to Access Qwen
Accessing Qwen is straightforward thanks to Alibaba Cloud’s API. You don’t need to be a coding expert to use it as it supports multiple programming languages and tools like Postman.
First you’ll need to create an Alibaba Cloud account and subscribe to the Qwen service. Once you have your API keys you can start integrating Qwen into your applications. The process is designed to be user-friendly even for those with limited technical experience.
Main Uses
Qwen is incredibly versatile and can be applied to a variety of real-world scenarios. One of its most popular uses is building chatbots for customer support or virtual assistants. These chatbots can handle complex conversations and provide accurate responses.
Another key use case is content creation. Qwen can generate articles summaries and even creative writing pieces. This makes it a valuable tool for marketers writers and educators.
Additionally Qwen excels at translation tasks. It can translate text between multiple languages with high accuracy. For businesses operating globally this feature can be a game-changer.
Before diving in it’s worth exploring real-world examples of Qwen in action. For instance Alibaba Cloud has shared case studies on how businesses use Qwen for customer support automation and multilingual content generation. You can read more about these use cases in their official blog post here.
While Qwen offers impressive capabilities, it’s important to be aware of its limitations. Like other large language models, Qwen may produce biased or inaccurate outputs and its environmental impact is a growing concern. Additionally, its reliance on Alibaba Cloud’s infrastructure means users must consider data privacy and platform dep.
Just don’t share your private data with any LLM no matter who built the tool. That’s a good standard of practice.
Why Qwen Matters
Qwen is more than just another AI tool. It represents the next step in making advanced language models accessible and practical for real-world applications. By understanding these four key aspects you can unlock its potential and explore how it can benefit your projects or business.
Dive in and discover what it can do for you.
