Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: robotics/AI

Natural-language AI helps chemists design molecules step by step

Designing molecules is one of chemistry’s most complex challenges. From life-saving drugs to advanced materials, each compound requires a precise sequence of reactions. Planning these steps demands both technical knowledge and strategic insight, making it a task that often relies on years of experience.

Two problems plague much of modern chemistry. The first is retrosynthesis: Chemists start from a target molecule and work backward to identify simpler building blocks and viable reaction pathways. Retrosynthesis involves countless decisions, from choosing starting materials to determining when to form rings or protect sensitive functional groups. While computers can explore vast “chemical spaces,” they often struggle to capture the strategic reasoning used by human experts.

The second problem is reaction mechanisms. These describe how chemical reactions unfold step by step through the movements of electrons. Mechanistic insight helps scientists predict new reactions, improve efficiency, and reduce costly trial and error. Existing computational methods can generate many possible pathways, but often lack the chemical intuition needed to identify the most plausible ones.

Bridging structure and function: artificial intelligence-based modelling of kidney proteins

Advances in artificial intelligence-driven algorithms and experimental technologies have revolutionized the field of protein modelling. This Review describes how these developments have provided unprecedented insights into the structure of key proteins within the kidney, improved understanding of the relationships between protein structure and stability, and enabled mechanistic interpretation of variants that underlie a variety of kidney pathologies.

Nvidia becomes first company to cross $5 trillion in market value

Nvidia has achieved a historic milestone. The chipmaker is now the world’s most valuable listed company. Its market capitalization has surpassed five trillion dollars. This surge places Nvidia ahead of tech giants like Alphabet and Apple. The company’s success is driven by its crucial role in supplying GPUs for artificial intelligence models. Nvidia’s stock performance reflects its strong market position.

Machine learning identifies catalyst ‘sweet spot’ for greener urea from waste gases

Urea is an extremely important chemical, especially for fertilizers. But, making urea is energy intensive and relies heavily on fossil fuels. However, new findings from Griffith University and the Queensland University of Technology have highlighted new ways to produce urea electrochemically, using electricity and waste gases such as carbon monoxide (CO) and nitrogen oxides (NO) instead.

The paper, “Machine Learning-Assisted Design Framework of Carbon Edge-Dominated Dual-Atom Catalysts for Urea Electrosynthesis,” has been published in ASC Nano.

“The challenge is that when CO and NO react on a catalyst, they usually don’t form urea,” said co-lead author Professor Qin Li from Griffith University.

Label-free optical imaging enables automated measurement of human white matter microstructure

White matter pathways allow distant parts of the brain to communicate, supporting memory, emotion, and language. One such pathway, the uncinate fasciculus, connects the front of the temporal lobe with regions of the frontal cortex involved in decision-making and social behavior. Despite its importance, little is known about the microscopic structure of this tract in the human brain.

Traditional techniques such as electron microscopy can reveal fine details, but they often fail when applied to postmortem human tissue, which is frequently degraded.

In a study published in Biophotonics Discovery, researchers report a new way to examine white matter structure in postmortem human brains.

AI Models Mirror Human Logic on Real-World Scenarios

“What we show is that the models actually capture that human uncertainty pretty well,” said Michael Lepori. [ https://www.labroots.com/trending/technology/30475/ai-models…cenarios-2](https://www.labroots.com/trending/technology/30475/ai-models…cenarios-2)

Can AI models distinguish fact from fiction? This is what a new study scheduled to be presented at the International Conference on Learning Representations this weekend hopes to address as a team of scientists investigated how AI models could tell the difference between facts and “fake news”. This study has the potential to help scientists, engineers, and the public better understand how AI models can evolve to meet human needs, which comes at a time when AI is becoming more integrated into our everyday lives.

For the study, the researchers analyzed how AI language models (LMs) were able to differentiate between different topics and information and judge what’s true and what’s fake. The motivation behind this study was to address a knowledge gap regarding whether large language models (LLMs) have a human-like understanding of the world or if they simply make decisions based on what’s given to them.

The goal of the study was to ascertain if the LMs could determine whether an event is real or fake, along with ascertaining when the LM makes this determination during its thought process. For example, the researchers would give the LM simple scenarios like “clean a car”, clean a road”, and “clean a cloud”, and ask the LM to figure out which was real or fake. In the end, the researchers found that large LMs were capable of differentiating between real and fake events or data.

Why faster AI isn’t always better

In the race to make AI models not just reason better but respond faster, latency—the delay before an answer appears—is often treated as a purely technical constraint, something to minimize and move past. But how is this relentless push for speed actually impacting the people using these systems every day?

There is a rich body of work in human–computer interaction linking faster response times to better usability. But AI models are fundamentally different from the deterministic systems that previous research was built on. When you wait for a file to download or a page to load, the outcome is fixed and predictable.

AI models are probabilistic—you cannot anticipate the precise response. Their conversational interface means users naturally read human social cues into the interaction. A pause might be read as the AI “thinking,” for instance. Users are increasingly asked to choose between faster models and slower, deeper-reasoning ones, without guidance on what that choice actually means for their experience.

/* */