This week Google DeepMind made waves by announcing Veo 3, its newest and most powerful video generation model. Veo 3 can create high-quality videos from simple text prompts, pushing the boundaries of what AI can do in creative storytelling. Alongside that, Google also supercharged its Gemini 2.5 Pro model, improving its ability to reason, generate text, and assist with complex tasks.

But while the tech world was celebrating faster and smarter AI, a detailed study from MIT brought a moment of pause. Researchers found that some of these cutting-edge generative models come with a hefty energy price tag. To generate just five seconds of high-quality video, an AI model could consume as much electricity as running a microwave oven for a full hour.

Advances in Artificial Intelligence

At Microsoft Build 2025, the company showcased its commitment to AI integration across its platforms. Highlights included the expansion of Copilot AI in Windows 11 and Microsoft 365, and the introduction of new tools like Discovery for scientific research.

Farmland’s Plastic Problem Is Bigger Than We Thought

A new review by Murdoch University has revealed a staggering fact — agricultural soils now contain 23 times more microplastics than the world’s oceans. While plastic pollution in oceans often makes headlines, this study brings attention to a quieter but potentially more dangerous crisis unfolding beneath our feet.

The research highlights that soil microplastics are not just leftover bits of bags or packaging. They can carry up to 10,000 different chemical additives, many of which are unregulated in farming. These compounds could pose serious risks to soil health, crop safety, and the overall integrity of the ecosystem.

Climate Change and Extreme Weather

Severe flooding in New South Wales, Australia, resulted in at least five deaths and significant property damage. Officials attributed the increasing frequency and intensity of such events to climate change, emphasizing the need for enhanced preparedness and infrastructure resilience.

A Close Encounter with a Giant Asteroid

On May 24, asteroid 2003 MH4, roughly the size of the Eiffel Tower, passed near Earth at a speed of about 30,000 kilometers per hour. While there was no danger of collision, NASA emphasized the importance of monitoring such near-Earth objects to safeguard our planet.

Solar Storms Prompt Global Preparedness

NASA issued an emergency alert due to intense solar flares from sunspot AR4087. These flares raised concerns about potential disruptions to power grids and communication systems. A global drill was conducted to assess preparedness, revealing areas needing improvement.

A New Shield Against Gonorrhoea

In a groundbreaking public health move this week NHS England began rolling out the world’s first vaccine to prevent gonorrhoea. The vaccine has shown an efficacy of about 30 to 40 percent. While that might sound modest it marks an important step in the global fight against one of the most common and increasingly antibiotic-resistant sexually transmitted infections.

Sources: Tom’s Guide Wikipedia The Guardian Life Technology

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post Solar Storms, AI Leaps, Novel Vaccine & Plastic Problem — Science Just Had a Big Week appeared first on Everyman Science.

The post Solar Storms, AI Leaps, Novel Vaccine & Plastic Problem — Science Just Had a Big Week appeared first on Everyman Science.

Building Without Boundaries

Traditional construction is constrained by the ground—workers, cranes, and ground-based robots can only reach so far. Aerial robots, however, operate within an unrestricted work envelope, soaring over terrains and scaling heights previously deemed inaccessible or hazardous. This newfound freedom promises to unlock construction possibilities in remote, difficult, or dangerous environments where conventional methods fall short.

Moreover, these aerial systems could minimize environmental impacts. By using advanced material deposition techniques, drones can potentially reduce construction waste, lower energy consumption, and improve the overall sustainability profile of large-scale projects.

What is Aerial Additive Manufacturing?

Aerial AM is essentially an airborne version of 3D printing. Drones, outfitted with precision manipulators and material dispensers, deposit construction materials layer-by-layer, shaping structures mid-flight. While early prototypes have focused on simpler tasks like on-demand repairs and modular assemblies, the vision is to scale this technology for full-fledged construction projects.

Challenges

Despite its promise, the field of Aerial AM remains in its infancy. As Dr. Basaran Bahadir Kocer from the University of Bristol’s School of Civil, Aerospace and Design Engineering notes, several hurdles must be overcome:

• Material Durability: The materials used must harden quickly yet remain strong enough to support the structure under varying environmental conditions.
• Outdoor Localisation Systems: Accurate navigation in uncontrolled, outdoor environments remains a significant technological challenge. GPS alone may not offer the precision required.
• Fleet Coordination: Large-scale projects would require swarms of drones working simultaneously, which introduces complex challenges in real-time coordination, collision avoidance, and task delegation.

Addressing these technical bottlenecks is critical for transitioning from controlled laboratory demonstrations to practical, real-world applications.

Testing the Skies

To move beyond theoretical models, researchers are now testing these systems in real-world environments. At DroneHub, based at Switzerland’s EMPA (Swiss Federal Laboratories for Materials Science and Technology), Aerial AM technologies are being trialed outside the laboratory for the first time. This platform offers researchers an invaluable opportunity to fine-tune drone performance under authentic conditions.

This initiative is further supported by a thriving collaboration between the University of Bristol and EMPA, strengthened by fully-funded PhD studentships and access to cutting-edge facilities like Bristol’s Flight Lab, Bristol Robotics Laboratory, and the Aerial Robotics Group.

Why It Matters

The implications of this research stretch far beyond traditional building sites. Imagine drones rapidly constructing temporary shelters in disaster-stricken areas, repairing critical infrastructure in inaccessible locations, or even assembling structures in extraterrestrial environments like the Moon or Mars.

By integrating aerial robotics into construction, we’re looking at a future where building is faster, safer, greener.

Source:

Research Paper Referenced: “Aerial Additive Manufacturing: Towards On-site Building Construction with Aerial Robots“
Authors: Yusuf Furkan Kaya, Lachlan Orr, Basaran Bahadir Kocer, Vijay Pawar, Robert Stuart-Smith, and Mirko Kovac
Published in: Science Robotics

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post Drones in Construction: How Aerial Robotics Could Revolutionize Building Practices appeared first on Everyman Science.

The post Drones in Construction: How Aerial Robotics Could Revolutionize Building Practices appeared first on Everyman Science.

For years, studies have indicated that AI could aid medical professionals in identifying cancers by detecting abnormalities in CT scans and spotting signs of breast cancer in mammograms. However, most studies until now have been retrospective. Many have also been limited by small sample sizes. This has left a gap in understanding AI’s efficacy in real-world clinical settings.

A Nationwide Study

The research, conducted as part of Germany’s national breast cancer screening program, analyzed data from 461,818 women aged 50–69 who participated between July 2021 and February 2023. All scans were independently reviewed by two radiologists. However, for 260,739 women, at least one radiologist used an AI tool for additional support.

The AI tool performed two critical functions. First, it labeled scans deemed unsuspicious as “normal.” Then, it issued “safety net” alerts for scans it flagged as suspicious but were considered normal by the radiologist. These alerts highlighted specific areas for further review.

Enhanced Detection with Minimal Risk

The study revealed that the AI-assisted group achieved a breast cancer detection rate that was 17.6% higher than the standard group, equating to one additional cancer diagnosis per 1,000 women screened. Importantly, this improvement did not lead to a higher rate of false positives. The recall rate for further investigation remained consistent across both groups.

Professor Alexander Katalinic from the University of Lübeck emphasized that the integration of AI enhanced detection rates without increasing harm for participants. He stated:

“This is a better result, with the same harm.”

AI as a Safety Net

The AI tool’s safety net feature played a pivotal role in preventing missed diagnoses. In the AI-assisted group, this feature was triggered 3,959 times, leading to 204 confirmed breast cancer cases. Without this safeguard, 20 cancers in the AI group could have been overlooked. This highlights the system’s value in complementing radiologists’ expertise.

Stefan Bunk, co-founder of Vara, the company behind the AI tool, noted that the technology streamlined the review process for scans labeled as “normal.” He added that even without expert review, overall detection rates would remain higher. This would result in fewer false positives and reduced workloads for radiologists.

Expert Perspectives

The findings have garnered widespread attention and praise from experts, with Stephen Duffy, emeritus professor of cancer screening at Queen Mary University of London, calling the results “credible and impressive.” He highlighted the potential for AI to replace the need for two radiologists in settings like the UK. In the UK, radiologist shortages are a pressing concern.

However, some experts raised cautionary notes. Dr. Kristina Lång of Lund University acknowledged the study’s encouraging results but warned about the increase in detected in situ cancers, which may contribute to overdiagnosis. She said:

“Long-term follow-up is essential to fully understand the clinical implications of integrating AI into mammography screening.”

Similarly, Dr. Katharine Halliday, president of the Royal College of Radiologists, stressed the importance of careful implementation. She noted the 29% shortfall of radiologists in the NHS. While AI tools can boost accuracy and productivity, she emphasized that expert oversight is crucial to mitigate potential risks.

A Promising Step Forward

This study marks a breakthrough in integrating AI into routine breast cancer screening. It boosts detection rates, eases radiologists’ workloads, and tackles resource shortages in global healthcare.

Yet, experts stress the need for rigorous research and long-term follow-up to ensure AI’s benefits translate into better patient outcomes. The journey is just beginning, but AI’s transformative potential is unmistakable.

Source: Nationwide real-world implementation of AI for cancer detection in population-based mammography screening

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post New Study Proves AI Can Improve Early Breast Cancer Detection appeared first on Everyman Science.

The post New Study Proves AI Can Improve Early Breast Cancer Detection appeared first on Everyman Science.

January 2024: A Promising Start

2 January: The Japan Meteorological Agency confirmed 2023 as the warmest year on record globally.

3 January: Scientists created the first functional semiconductor made from graphene, paving the way for faster and more efficient electronic devices.

13 January: NASA opened the Bennu asteroid sample container, offering insights into the early solar system’s formation.

17 January: A newly discovered hormone, Glycogenin-2, was linked to improved glucose metabolism, sparking hope for new diabetes treatments.

19 January: Japan achieved the fifth successful soft landing on the Moon, continuing humanity’s exploration of our closest celestial neighbor.

24 January: The Large Hadron Collider reported new evidence of tetraquarks, expanding our understanding of particle physics.

31 January: NASA discovered a super-Earth, TOI-715 b, located in the habitable zone, reigniting discussions about extraterrestrial life.

February 2024: Pioneering Solutions

2 February: Researchers proposed a potential solution to the longstanding three-body problem, a major breakthrough in astrophysics.

22 February: Intuitive Machines’ Nova-C lander became the first commercial vehicle to land on the Moon, marking a milestone in private space exploration.

26 February: Prebiotic resistant starch was shown to aid weight loss in small trials, offering new avenues for combating obesity.

March 2024: Quantum Leaps and Medical Marvels

4 March: Europa, Jupiter’s moon, may have less oxygen than previously thought, challenging its potential for life.

Biochemists create an RNA molecule capable of making accurate copies, supporting the RNA world hypothesis.

6 March: Colossal Biosciences creates the first induced pluripotent stem cells for the Asian elephant, a step toward woolly mammoth de-extinction.

13 March: The EU passes the world’s first comprehensive Artificial Intelligence Act.

14 March: SpaceX successfully launches Starship, but the rocket fails upon re-entry.

19 March: Scientists demonstrate a wireless brain-sensing network of 78 sensors.

26 March: A study proposes ecological pandemic prevention measures for policy frameworks.

28 March: LHS 3844 b confirmed as the first tidally locked super-Earth exoplanet.

April 2024: Innovations

1 April: A new class of antibiotics targeting multi-drug-resistant bacteria was discovered, potentially revolutionizing infection treatment.

4 April: A study suggests global CO2 emissions increased by only 0.1% in 2023, indicating a potential plateau.

9 April: A rare genetic variation that reduces the risk of developing Alzheimer’s by over 70% is identified.

12 April: Researchers find that tardigrades’ biochemicals, especially Dsup protein, protect them from radiation exposure.

23 April: University of Maine presents the world’s largest 3D printer capable of printing large-scale objects.

26 April: mRNA-4157/V940, the first personalized melanoma vaccine based on mRNA, enters Phase III trials.

May 2024: The Cosmos and the Human Mind

1 May: The discovery of a brain circuit potentially acting as a “master regulator” of immunity linked the brain and immune system in groundbreaking ways.

5 May: The James Webb Space Telescope identified JADES-GS-z14-0, the most distant galaxy ever discovered, formed 290 million years after the Big Bang.

17 May – A team from the University of Bristol unveiled the world’s smallest quantum light detector, built on a silicon chip. This new version is 50 times smaller than their previous model.

24 May: Researchers from the Chinese Academy of Sciences announced the successful tuning of the Casimir effect using magnetic fields.

31 May: Biologists have reported that Tmesipteris oblanceolata, a fern ally plant, contains the largest known genome.

June 2024: Space and Sustainability

2 June: China’s Chang’e 6 lunar mission successfully landed on the Moon’s far side, collecting valuable samples.

5 June: The Iberian lynx’s population surge led to its removal from the “endangered” list, showcasing conservation success.

11 June: Research suggests a link between serious kidney disease and human spaceflight.

24 June: Three potentially habitable super-Earths were discovered around the star HD 48948, advancing the search for life beyond Earth.

July 2024: Climate and Dark Matter Insights

9 July: Rising sea levels caused the extinction of the Key Largo tree cactus, highlighting the dire effects of climate change.

11 July: The Hubble Telescope uncovered 3D velocity dispersion in a dwarf galaxy, offering deeper insights into dark matter distribution.

August 2024: Mars, mRNA

12 August: Liquid water was confirmed to exist 10–20 km below Mars’ surface, raising possibilities for future exploration and the search for life.

23 August: The first mRNA lung cancer vaccine, BNT116, began clinical trials, representing a major leap in cancer treatment.

September 2024: Quantum World and Climate Warnings

4 September: The European Space Agency’s BepiColombo spacecraft performed a record-setting 165 km flyby of Mercury. (Video)

11 September: A study conducted by Osaka University reveals that the bluestreak cleaner wrasse (Labroides dimidiatus), a small tropical fish, may exhibit a form of self-awareness.

23 September: A study forecasted the near-total collapse of the West Antarctic Ice Sheet by 2300 if current trends continue.

October 2024: Technological Feats, Planetary Insights

2 October: Scientists completed the first-ever mapping of the entire brain of a fruit fly, revealing over 50 million connections between 139,000 neurons.

October 7: The Nobel Prize in Physiology or Medicine 2024 was awarded jointly to Victor Ambros and Gary Ruvkun “for the discovery of microRNA and its role in post-transcriptional gene regulation”

October 8: The Nobel Prize in Physics was awarded to John J. Hopfield and Geoffrey Hinton for “for foundational discoveries and inventions that enable machine learning with artificial neural networks”.

9 October: AI-driven protein design led to the creation of enzymes capable of breaking down harmful industrial pollutants, offering a new tool for environmental cleanup.

9 October: The Nobel Prize in Chemistry was awarded to David Baker “for computational protein design” and Demis Hassabis and John Jumper “for protein structure prediction”.

13 October: SpaceX achieved the first successful return and capture of a Super Heavy booster from Starship, setting a new benchmark in space technology.

18 October: Researchers from the Hebrew University of Jerusalem have introduced a holography-based computational technique that significantly enhances medical optical imaging.

23 October: Physicists from MIT and the Caltech have observed a “black hole triple” for the first time.

30 October: A new study on the infinite monkey theorem concludes that, even if all 200,000 chimpanzees on Earth typed one key per second, they would still fail to produce the complete works of Shakespeare before the universe’s end.

November 2024: AI and Health

15 November: Global measles cases have reportedly surged, with an estimated 10.3 million infections in 2023, marking a 20% increase compared to 2022.

21 November: The European Southern Observatory’s Very Large Telescope Interferometer has captured the first close-up image of a star beyond the Milky Way. The star, WOH G64, located in the Large Magellanic Cloud approximately 160,000 light-years away, is revealed to be encircled by a torus-shaped cloud.

December 2024: AI, Human Health

On December 5, researchers identified a single mutation, Q226L, that significantly enhances the ability of H5N1 (‘bird flu’) to infect human cells, particularly in the respiratory tract. Previously, it was believed that at least three mutations were necessary for the virus to infect and spread among humans.

On December 10, AI-based transfer learning models predicted that global warming could reach 3°C sooner than previously anticipated.

Welcoming 2025

In 2024, humanity took remarkable strides in science, whether it was unraveling the mysteries of space, advancing medical frontiers, or harnessing the transformative power of AI and technology. These breakthroughs remind us of the boundless potential of human curiosity and innovation.

At Everyman Science, we remain committed to delivering in-depth stories that decode complex topics, be it scientific, societal, or political. We hope this roundup of 2024’s major scientific achievements inspired you and deepened your appreciation for the incredible progress made this year.

Here’s to another year of discovery, wisdom, and scientific advancement. Bring on 2025!

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post The Top Scientific Breakthroughs of 2024 appeared first on Everyman Science.

The post The Top Scientific Breakthroughs of 2024 appeared first on Everyman Science.

Understanding the Winners

David Baker: Architect of New Proteins

David Baker, a professor at the University of Washington in Seattle, has made significant strides in the design of new proteins. His inventive approach involves the creation of entirely new types of proteins, which were previously unimaginable in nature. Baker’s research group has been at the forefront of this innovative field, developing imaginative protein designs that could play crucial roles in various sectors, including pharmaceuticals, vaccines, nanomaterials, and sensors.

The implications of Baker’s work are profound, as his engineered proteins may pave the way for new treatments and technologies, emphasizing the vital role of protein research in addressing global challenges as antibiotics resistant bacteria.

Demis Hassabis and John Jumper: AI and Protein Structure Prediction

On the other hand, Demis Hassabis and John Jumper, both associated with Google DeepMind in London, have made remarkable contributions through their groundbreaking AI model known as AlphaFold2. This model has transformed the landscape of protein structure prediction, enabling scientists to anticipate the shapes of nearly all known proteins—over 200 million in total.

The accuracy of AlphaFold2 has far-reaching implications for biological research, including insights into antibiotic resistance and the visualization of enzymes that can break down plastic. Such advancements are critical in our understanding of complex biological processes and the development of solutions to pressing environmental issues.

Related: Hopfield and Hinton Win 2024 Nobel Prize for Breakthroughs in Machine Learning

The Significance of Protein Research

Proteins: The Building Blocks of Life

Proteins are integral to the functioning of all living organisms. Often described as nature’s “ingenious chemical tools,” they facilitate and govern the chemical reactions that sustain life. Proteins have diverse roles, ranging from acting as hormones and signaling molecules to serving as antibodies and foundational components of tissues. The work conducted by Baker, Hassabis, and Jumper not only enhances our understanding of these vital molecules, but it also opens vast new possibilities for their applications in science and medicine.

The Nobel Prize: A Recognition of Groundbreaking Discoveries

The 2024 Nobel Prize in Chemistry comes with a total prize amount of 11 million Swedish kronor. David Baker receives half of this sum, while the other half is equally shared between Demis Hassabis and John Jumper, reflecting the collaborative nature of their groundbreaking research endeavors.

This recognition by the Nobel Committee for Chemistry highlights not only the monumental discoveries made by these scientists but also their potential to significantly impact various fields, from biotechnology to environmental science.

The awarding of the Nobel Prize in Chemistry to David Baker, Demis Hassabis, and John Jumper is a testament to the extraordinary advances in protein research and the potential it holds for the future. Their groundbreaking work not only deepens our understanding of life’s fundamental processes but also sets the stage for innovative solutions to some of the world’s most pressing challenges. As we continue to explore the complexities of proteins and their structures, the contributions of these researchers will undoubtedly inspire and drive future developments in science.

Here is an official press release from nobelprize.org.

Also Read: Ambros & Ruvkun’s MicroRNA Breakthrough Wins 2024 Nobel Prize in Physiology & Medicine

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post 2024 Nobel Prize in Chemistry Awarded for Breakthroughs in Protein Research appeared first on Everyman Science.

The post 2024 Nobel Prize in Chemistry Awarded for Breakthroughs in Protein Research appeared first on Everyman Science.

Nano, Pro, and Ultra

Google CEO Sundar Pichai announced the beginning of the Gemini era, emphasizing the profound impact it will have on the company’s products. Gemini encompasses three distinct versions: Gemini Nano, a lightweight model for native and offline use on Android devices, Gemini Pro, a more robust version for powering Google AI services, and Gemini Ultra, a highly advanced model primarily designed for data centers and enterprise applications.

The model is currently being rolled out in various ways, including the integration of Gemini Pro into Bard and the introduction of new features for Pixel 8 Pro users thanks to Gemini Nano. Further, developers and enterprise customers will gain access to Gemini Pro through Google Generative AI Studio or Vertex AI in Google Cloud, starting on December 13th. While Gemini is currently only available in English, Google plans to integrate it into a wide array of its products globally in the near future.

Also Read: Chat GPT-4 vs. Its Predecessor: What You Need to Know

In comparison to OpenAI’s GPT-4, Google’s Gemini boasts advantages in numerous benchmarks, particularly in areas related to understanding and interacting with video and audio content. Additionally, Gemini’s versatility extends to processing images, video, and audio and is poised to further expand its capabilities to incorporate other sensory inputs.

Despite the promising benchmarks, the true test of Gemini’s potential will come from everyday users who will benefit from its diverse functionalities, such as idea generation, information lookup, and coding. Moreover, Google has emphasized Gemini’s efficiency, noting its superiority in terms of speed, cost-effectiveness, and reliability in comparison to its predecessors.

With the launch of Gemini, Google asserts its position as an AI innovator, aiming to reclaim and solidify its standing in the AI landscape. While the company prioritizes ambitious advancements in AI, it is also committed to ensuring the safety and responsibility of its AI systems, with a focus on thorough testing for potential vulnerabilities.

As Google embarks on this new phase of AI technology, Pichai and Hassabis signal that this is just the beginning of what they envision as a transformative era. They maintain a cautious yet optimistic approach as they push the boundaries of AI, recognizing the profound potential and the need for responsible development and deployment.

Ultimately, Google views Gemini as a monumental development that could potentially surpass the impact of the internet, positioning the company at the forefront of the AI landscape.

admin

www.everymansci.com

The post Google’s AI Revolution: Introducing Gemini to Rival GPT-4 appeared first on Everyman Science.

The post Google’s AI Revolution: Introducing Gemini to Rival GPT-4 appeared first on Everyman Science.

This isn’t the first time OpenAI has introduced a new model for ChatGPT, as they previously upgraded from GPT-3.5 to GPT-4 earlier this year. However, the forthcoming GPT-4 Turbo is expected to set itself apart in several ways when it becomes available later this year. It will likely be initially rolled out to subscribers of ChatGPT Plus, followed by a release to the general public.

While early access to the new ChatGPT model from OpenAI hasn’t been granted yet, we can still make informed speculations about the expected enhancements in GPT-4 Turbo using the available information.

Extending Knowledge: GPT-4 Turbo’s Expanded Cutoff

First and foremost, the new model will no longer be restricted to information only up until September 2021. OpenAI CEO Sam Altman expressed his frustration with this limitation and announced that GPT-4 Turbo will incorporate knowledge up until April 2023. This means that the chatbot will be able to provide responses with more up-to-date context, ensuring users receive the most current information. However, concerns about how OpenAI obtains this information without proper consent from authors and publishers continue to be a contentious issue.

Longer Prompts, Deeper Insights: GPT-4 Turbo’s Enhanced Capabilities

Another significant improvement is that GPT-4 Turbo will support longer and more detailed prompts. Altman explained that the model will accommodate up to 128,000 tokens of context, equivalent to approximately 300 book pages. This enhancement enables users to input extensive documents and receive comprehensive summaries or analyses from the chatbot.

Related: Google’s AI Revolution: Introducing Gemini to Rival GPT-4

Precise Instruction Following

OpenAI also promises better instruction following with the new model. GPT-4 Turbo is designed to be a more attentive listener and demonstrates improved performance in tasks that require precise adherence to instructions. Whether it’s generating content in specific formats like XML or assisting with code writing, the chatbot’s ability to follow instructions accurately will greatly benefit users.

Affordable AI: Pricing Options for Developers

In terms of pricing, developers will be pleased to learn that OpenAI is offering more affordable options. Altman announced that GPT-4 Turbo will be priced at one cent for every thousand prompt tokens and three cents for every thousand completion tokens. This revised pricing structure makes it more accessible and cost-effective for developers to utilize the capabilities of the chatbot.

User-Friendly Evolution: GPT-4 Turbo’s Streamlined Interface

Lastly, OpenAI is simplifying the user experience by integrating multiple tools into the chat interface. Previously, users had to navigate a cumbersome dropdown menu to select various chatbot tools. However, with GPT-4 Turbo, this confusing feature will be replaced. The chatbot will automatically select the appropriate tools based on user requests, eliminating the need for manual selection. For example, if you request an image, the chatbot will utilize Dall-E 3 to generate it, streamlining the process.

OpenAI’s latest model, GPT-4 Turbo, introduces several significant updates that enhance the functionality of ChatGPT. From expanding its knowledge cutoff and accommodating longer prompts to improving instruction following and offering more affordable pricing, these advancements aim to provide users with a more powerful and user-friendly experience.

You Might Also Like:

Chat GPT-4 vs. Its Predecessor: What You Need to Know

admin

www.everymansci.com

The post From GPT-3.5 to GPT-4 Turbo: OpenAI’s Journey of Innovation appeared first on Everyman Science.

The post From GPT-3.5 to GPT-4 Turbo: OpenAI’s Journey of Innovation appeared first on Everyman Science.

Experiments in Problem Solving

The experiments involved displaying different stimuli on a screen, including lines of varying width, placement, and orientation, as well as segmented and concentric rings. The pigeons had to peck a button on the left or right to determine which category the stimuli belonged to. Correct choices were rewarded with food pellets, while incorrect choices received no reward. The pigeons did not require a specific rule, learning primarily through trial and error. For instance, if presented with a visual labeled “category A,” they would categorize anything resembling it as “category A” based on their ability to identify similarities.

Associative Minds and Cognitive Insights

Throughout the experiments, the pigeons demonstrated an improved ability to make correct choices, increasing their accuracy from 55% to 95% in simpler tasks. When faced with more complex challenges, their accuracy rose from 55% to 68%. Onur Güntürkün, a professor of behavioral neuroscience at Ruhr University Bochum, who was not involved in the study, remarked that using animals like pigeons allowed researchers to gauge the potential of mostly associative minds. The findings underscore the strength of associative systems and their cognitive-like nature.

AI models aim to identify patterns and make decisions, a process that pigeons are shown to perform as well. Pigeons learn through consequences, exhibiting a remarkable ability to correct errors when no food pellet is provided. Similarity plays a crucial role for pigeons as they use their capacity to recognize resemblances between objects. Lead author of the study, Brandon Turner, emphasized the significance of these mechanisms, suggesting that neural networks or AI machines can employ them to solve categorization problems. Turner believes that the mechanisms present in AI are also present in pigeons.

The researchers intend to collaborate with experts who study pigeons and their brains, hoping to leverage these findings for a better understanding of human brain damage. Wasserman expressed optimism about gaining further insights into pigeon cognition, praising their extraordinary learning capacity despite their small brain size. It is important to note that no harm was inflicted upon pigeons during the course of the study.

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post Pigeons, AI, and Cognitive Resemblances appeared first on Everyman Science.

The post Pigeons, AI, and Cognitive Resemblances appeared first on Everyman Science.

AI systems have the ability to estimate the composition of contaminants in water and optimize the efficiency of water treatment plants. By leveraging neural networks, these systems can reduce energy costs by up to 40%. Additionally, AI can enhance the detection of bacteria in water, surpassing the limitations of traditional methods that rely on chemical strips. AI systems employ optics to identify both known and previously unknown contaminants, drastically reducing time and costs.

AI’s Impact on Wastewater and Agriculture

Wastewater treatment is another area where AI plays a crucial role. Real-time visibility into effluent levels at remote wastewater sites allows for effective management and early problem detection. Fault detection software and continuous monitoring increase operational efficiency and minimize potential issues. Furthermore, AI systems streamline water use in agriculture, which currently accounts for 70% of freshwater consumption but operates at less than 50% efficiency. AI-based systems can analyze soil moisture, crop growth stage, weather conditions, and other factors to optimize irrigation, thereby conserving water resources.

AI-Powered Smart Meters Revolutionizing Water Management

In terms of water conservation, AI-powered smart meters can identify leaks and pipe bursts in residential areas, reducing wastage and improving system efficiency. Governments and water corporations are utilizing software analytics to manage and analyze sensor data, enabling better management of water systems and reducing unaccounted-for water. Telcos are partnering with IoT platform vendors to develop solutions that address the growing demand for IoT in the water sector. The installation of smart meters worldwide is projected to experience a 28% growth rate, reaching nearly 400 million units by 2026.

As AI continues to advance, the possibilities for its application in the water sector are immense. With accurate data on water quality, consumers can choose personalized filtering solutions tailored to their specific needs. Moreover, as adoption increases and costs decrease, AI systems become more efficient, creating a positive feedback loop that helps preserve our planet’s most precious resource – water. The time to harness AI’s potential in the water sector is now.

Also Read: Unleashing the Power of Artificial Intelligence in Business

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post How Artificial Intelligence is Revolutionizing the Water Sector appeared first on Everyman Science.

The post How Artificial Intelligence is Revolutionizing the Water Sector appeared first on Everyman Science.

AI’s Smell Prediction Capability Principal Odour Map (POM)

By analyzing the relationship between a chemical’s structure and its smell, researchers at a start-up called Osmo utilized the AI to develop a principal odour map (POM) that can be used to predict the scent of new molecules. The POM was tested against human subjects who were trained to associate specific smells with descriptive words used by the AI, and the AI’s predictions closely matched the average responses given by the humans. This technology could be valuable in designing new synthetic scents and may also provide insights into how the human brain interprets smell.

The sense of smell is unique in that it directly connects sensory information from the nose to the brain’s memory and emotional centers, evoking specific and intense memories. Neurobiologist Alexander Wiltschko believes that smell has a distinct quality, which has led to the creation of his start-up company, Osmo, that aims to design new odorants. Osmo’s team used a neural network AI system to analyze the structures of approximately 5,000 odorants and assign descriptive words to each one, allowing the AI to predict the aroma of new molecules. Through this analysis, the system identified 250 correlations between specific patterns in a chemical’s structure and its smell, which were then combined into the POM for future predictions.

Human Testing and AI Accuracy

To test the accuracy of the POM, 15 volunteers were trained to associate smells with descriptive words used by the AI. The researchers collected hundreds of odorants that do not naturally exist but are familiar enough for people to describe, and asked both the humans and the AI to predict the scent of these molecules. The AI’s predictions were often closer to the average response given by the humans than any individual’s guess.

The Unique Nature of the Sense of Smell

While the POM holds potential for industries such as food and cleaning products, some experts believe that it doesn’t reveal much about the biology behind the human sense of smell. Stuart Firestein, a neuroscientist at Columbia University, says that the POM doesn’t explain how different molecules interact with the approximately 350 odor receptors in the human nose. Pablo Meyer, a systems biologist at the IBM Center for Computational Health, highlights the subjective nature of smell and challenges the idea of there being a “correct” perception of a smell. Despite these criticisms, Wiltschko and his team plan to further investigate how odorants combine and compete with each other to create unique smells, although this task poses challenges due to the vast number of possible combinations.

The ability to predict the smells of different combinations of odorants is the next frontier in this field. Even something as singular as the smell of coffee is comprised of hundreds of odorant chemicals. By delving into this complexity, researchers hope to gain a deeper understanding of our sense of smell.

References: Lee, B. K. et al. Science 381, 999–1006 (2023).

Mohsin Rasheed

Co-Founder & Chief Editor of Everyman Science. I view science not just as a collection of facts, but as the ultimate guide for human survival. From medical breakthroughs to the logistics of space exploration, I am dedicated to documenting how scientific reasoning uplifts the human spirit and provides the blueprints to save our planet. I believe that by unleashing the power of nature through disciplined inquiry, we can secure a sustainable future for humanity.

The post Artificial Intelligence Sniffs Out Scents from Chemical Structures appeared first on Everyman Science.

The post Artificial Intelligence Sniffs Out Scents from Chemical Structures appeared first on Everyman Science.