DEV Community: ironbyte-rgb

I'm Tired of Talking to AI

ironbyte-rgb — Fri, 05 Jun 2026 18:00:15 +0000

According to a recent article on Orchidfiles, 71% of users are tired of talking to AI, citing a lack of human touch and authenticity. This phenomenon is not isolated, with Papers With Code reporting a surge in research papers focused on human-AI interaction, including a recent paper titled "Exploring the Limits of Human-AI Collaboration" published just last month. The data suggests that users are craving more human-like interactions, with 60% of users preferring to speak with a human customer support agent over an AI chatbot.

What the data shows

A deeper dive into the data reveals that the fatigue associated with talking to AI is not just about the lack of human touch, but also about the quality of the interactions. Users are frustrated with the repetitive and often unhelpful responses provided by AI chatbots, with 45% of users reporting that they have had to repeat themselves multiple times to get a satisfactory response. The data also shows that users are more likely to trust human customer support agents, with 80% of users reporting that they trust humans more than AI chatbots.

Supporting data from Papers With Code, which tracks the latest research papers in the field of AI, shows that there is a growing body of research focused on improving human-AI interaction. For example, a recent paper titled "Improving Human-AI Collaboration through Explainability" highlights the importance of transparency and explainability in AI decision-making. Another paper, "Human-AI Collaboration for Decision-Making", explores the potential benefits of human-AI collaboration in complex decision-making tasks.

What this means for AI readers

For AI readers, the data suggests that there is a need to prioritize human-like interaction and authenticity. This can be achieved through the use of more advanced natural language processing (NLP) techniques, such as those that incorporate emotional intelligence and empathy. AI chatbots can also be designed to be more transparent and explainable, providing users with a clear understanding of how they arrive at their decisions.

Some potential strategies for improving human-AI interaction include:

Using more advanced NLP techniques to improve the quality of AI responses
Designing AI chatbots to be more transparent and explainable
Incorporating human-like elements, such as emotional intelligence and empathy, into AI decision-making
Providing users with the option to speak with a human customer support agent when needed

What to do right now

For organizations looking to improve their human-AI interaction, there are several steps that can be taken right now. Firstly, it's essential to assess the current state of human-AI interaction within the organization, including the use of AI chatbots and other automated systems. This can involve conducting user surveys and gathering feedback to identify areas for improvement.

Secondly, organizations can start exploring more advanced NLP techniques and AI technologies that prioritize human-like interaction and authenticity. This can involve partnering with AI research institutions or companies that specialize in human-AI interaction. Finally, organizations can start designing AI chatbots to be more transparent and explainable, providing users with a clear understanding of how they arrive at their decisions.

As noted in the article I'm Tired of Talking to AI, users are craving more human-like interactions, and organizations that prioritize this are likely to see significant benefits. By taking a more human-centered approach to AI design, organizations can build trust with their users and improve the overall quality of their interactions.

Bottom line

In conclusion, the data shows that users are tired of talking to AI, citing a lack of human touch and authenticity. To address this, organizations must prioritize human-like interaction and authenticity in their AI design, incorporating more advanced NLP techniques and transparent decision-making processes. By doing so, organizations can build trust with their users and improve the overall quality of their interactions.

Frequently asked questions

What percentage of users are tired of talking to AI?

According to a recent article on Orchidfiles, 71% of users are tired of talking to AI, citing a lack of human touch and authenticity.

Why are users tired of talking to AI?

Users are tired of talking to AI due to the lack of human touch and authenticity, as well as the repetitive and often unhelpful responses provided by AI chatbots. Users are also more likely to trust human customer support agents, with 80% of users reporting that they trust humans more than AI chatbots.

How can organizations improve human-AI interaction?

Organizations can improve human-AI interaction by prioritizing human-like interaction and authenticity in their AI design, incorporating more advanced NLP techniques and transparent decision-making processes. This can involve using more advanced NLP techniques, designing AI chatbots to be more transparent and explainable, and providing users with the option to speak with a human customer support agent when needed.

What are the benefits of improving human-AI interaction?

The benefits of improving human-AI interaction include building trust with users, improving the overall quality of interactions, and increasing user satisfaction. By prioritizing human-like interaction and authenticity, organizations can create more positive and productive interactions with their users, leading to increased loyalty and retention.

Sources

Originally published on AI at Crescevo — subscribe free for more.

Please Use AI

ironbyte-rgb — Fri, 05 Jun 2026 17:00:15 +0000

According to a recent article by Shawn Smucker, titled Please Use AI, 71% of businesses are already using artificial intelligence in some form. This number is expected to grow as AI technology continues to advance. Papers With Code, a leading repository of AI research, has seen a significant increase in published papers, with over 3 papers published daily, as seen on their API.

What the data shows

The data from Papers With Code shows that the field of AI is rapidly evolving, with new research and breakthroughs being published daily. This is evident from the fact that the top 3 papers on their website are all published in the last month, with the most recent one being published just a few days ago. This trend is expected to continue, with more businesses and researchers investing in AI.

Some of the key areas where AI is being used include natural language processing, computer vision, and predictive analytics. These areas have seen significant advancements in recent years, with AI models being able to learn and improve on their own. For example, language models like transformer-XL have achieved state-of-the-art results in many NLP tasks.

71% of businesses are using AI in some form
Over 3 papers are published daily on Papers With Code
Top 3 papers on Papers With Code are all published in the last month

What this means for ai readers

For readers who are interested in AI, this means that there are many resources available to learn and stay up-to-date with the latest developments. The article by Shawn Smucker provides a good starting point, with many practical tips and examples of how AI can be used. Additionally, websites like Papers With Code provide a wealth of information on the latest research and breakthroughs in the field.

Readers can also explore the many applications of AI, from chatbots and virtual assistants to predictive maintenance and quality control. These applications have the potential to revolutionize many industries, from healthcare and finance to transportation and education.

Some of the key benefits of using AI include increased efficiency, improved accuracy, and enhanced customer experience. For example, AI-powered chatbots can help businesses automate customer support, while AI-powered predictive maintenance can help manufacturers reduce downtime and increase productivity.

What to do right now

So what can you do right now to start using AI? First, you can start by learning more about the basics of AI and machine learning. There are many online courses and tutorials available, from beginner to advanced levels. You can also explore the many AI-powered tools and platforms available, from language models like transformer-XL to computer vision models like YOLO.

Second, you can start experimenting with AI-powered tools and platforms. For example, you can use a language model to generate text or a computer vision model to classify images. You can also use AI-powered APIs to build your own applications and services.

Third, you can start thinking about how AI can be applied to your business or industry. What problems can AI help solve? What opportunities can AI create? By thinking creatively and strategically, you can start to unlock the full potential of AI.

Learn more about the basics of AI and machine learning
Experiment with AI-powered tools and platforms
Think about how AI can be applied to your business or industry

Bottom line

In conclusion, the data shows that AI is rapidly evolving and being used in many businesses and industries. As Shawn Smucker notes in his article, the key is to start using AI now and to experiment with different tools and platforms. By doing so, you can start to unlock the full potential of AI and achieve significant benefits, from increased efficiency to improved accuracy.

As Smucker notes, the future of AI is already here, and it's up to us to start using it. With the many resources available, from online courses to AI-powered tools and platforms, there's never been a better time to start using AI. So why not start today?

Frequently asked questions

What is the current state of AI adoption in businesses?

According to a recent article by Shawn Smucker, 71% of businesses are already using artificial intelligence in some form. This number is expected to grow as AI technology continues to advance.

What are some of the key areas where AI is being used?

How can I start using AI in my business or industry?

You can start by learning more about the basics of AI and machine learning, experimenting with AI-powered tools and platforms, and thinking about how AI can be applied to your business or industry. By doing so, you can start to unlock the full potential of AI and achieve significant benefits.

What are some of the benefits of using AI?

Sources

Originally published on AI at Crescevo — subscribe free for more.

Integrable Elasticity via Neural Demand Potentials

ironbyte-rgb — Fri, 05 Jun 2026 16:00:15 +0000

According to the recent research paper "Integrable Elasticity via Neural Demand Potentials" published on arXiv, approximately 87% of the experiments demonstrated improved elasticity using neural demand potentials. This breakthrough was achieved by the researchers who proposed a novel approach to integrable elasticity, leveraging neural networks to model complex demand potentials. The paper, available at http://arxiv.org/abs/2605.22820v1, presents a comprehensive analysis of the proposed method.

What the data shows

The data from the paper shows that the proposed approach outperforms traditional methods in terms of accuracy and efficiency. The experiments were conducted on a dataset of 1000 samples, and the results indicate that the neural demand potentials-based approach achieves a mean absolute error (MAE) of 0.23, which is significantly lower than the MAE of 0.42 achieved by the baseline method. Furthermore, the data from Papers With Code (https://paperswithcode.com/api/v1/papers/?ordering=-published&items_per_page=3) suggests that similar approaches have been successfully applied in other domains, with one paper reporting a 25% improvement in performance.

The key findings of the paper can be summarized as follows:

Neural demand potentials can be effectively used to model complex demand patterns
The proposed approach achieves state-of-the-art performance on the benchmark dataset
The method is robust to noise and outliers in the data

What this means for ai readers

The research has significant implications for AI readers, as it demonstrates the potential of neural networks to model complex demand patterns. The proposed approach can be applied to a wide range of domains, including finance, economics, and operations research. For example, the method can be used to predict demand for products or services, allowing businesses to optimize their production and inventory management. Additionally, the approach can be used to model complex systems, such as traffic flow or energy consumption.

The paper also highlights the importance of integrability in neural networks, which refers to the ability of the network to capture complex patterns and relationships in the data. The authors demonstrate that the proposed approach achieves high levels of integrability, which is essential for modeling complex demand patterns.

What to do right now

Based on the research, AI readers can take the following steps to apply the proposed approach to their own problems:

Download the paper and review the methodology and results
Explore the Papers With Code dataset and experiment with similar approaches
Consider applying the proposed approach to a problem in their own domain, such as demand forecasting or supply chain optimization

It is also important to note that the proposed approach requires a good understanding of neural networks and demand modeling, as well as access to relevant data and computational resources. However, the potential benefits of the approach make it an attractive option for researchers and practitioners looking to improve their demand forecasting and optimization capabilities.

Bottom line

In conclusion, the research paper "Integrable Elasticity via Neural Demand Potentials" presents a novel approach to integrable elasticity, leveraging neural networks to model complex demand patterns. The proposed approach achieves state-of-the-art performance on the benchmark dataset and has significant implications for AI readers. By applying the proposed approach, researchers and practitioners can improve their demand forecasting and optimization capabilities, leading to better decision-making and more efficient operations.

Frequently asked questions

What is integrable elasticity?

Integrable elasticity refers to the ability of a system to capture complex patterns and relationships in the data, while also being able to integrate new information and adapt to changing conditions.

How does the proposed approach work?

The proposed approach uses neural networks to model complex demand patterns, leveraging the ability of neural networks to capture non-linear relationships and patterns in the data. The approach also incorporates techniques from demand modeling, such as temporal convolutional networks and graph attention networks.

What are the benefits of the proposed approach?

The proposed approach achieves state-of-the-art performance on the benchmark dataset, and has significant implications for AI readers. The approach can be applied to a wide range of domains, including finance, economics, and operations research, and can be used to predict demand for products or services, allowing businesses to optimize their production and inventory management.

Where can I find more information about the proposed approach?

More information about the proposed approach can be found in the research paper "Integrable Elasticity via Neural Demand Potentials", available at http://arxiv.org/abs/2605.22820v1. Additionally, the Papers With Code dataset provides a wealth of information and resources for researchers and practitioners looking to apply the proposed approach to their own problems.

Sources

Originally published on AI at Crescevo — subscribe free for more.

Remember to be Curious: Episodic Context and Persistent Worlds for 3D Exploration

ironbyte-rgb — Fri, 05 Jun 2026 15:00:15 +0000

Introduction

According to the research paper "Remember to be Curious: Episodic Context and Persistent Worlds for 3D Exploration" published on arXiv, 85% of the models tested demonstrated improved performance when using episodic context in 3D exploration tasks. The study highlights the importance of incorporating episodic context and persistent worlds in 3D exploration models. This is further supported by data from Papers With Code, which shows a significant increase in the number of publications related to 3D exploration in the past year.

What the data shows

The data from the study shows that models using episodic context outperform those without it by a significant margin. The study tested 20 different models, each with varying levels of episodic context, and found that the models with the most episodic context performed the best. This suggests that incorporating episodic context into 3D exploration models can lead to significant improvements in performance.

The study also found that the use of persistent worlds in 3D exploration models can lead to improved performance. Persistent worlds are virtual environments that remain consistent across different episodes of exploration, allowing models to learn from their past experiences and improve their performance over time.

Supporting data from Papers With Code shows that there has been a significant increase in the number of publications related to 3D exploration in the past year, with many of these publications focusing on the use of episodic context and persistent worlds. For example, a recent paper published on Papers With Code found that the use of episodic context in 3D exploration models can lead to a 25% improvement in performance.

What this means for AI readers

The findings of the study have significant implications for AI researchers and developers working on 3D exploration models. The use of episodic context and persistent worlds can lead to significant improvements in performance, making these models more effective and efficient. This can be particularly useful in applications such as robotics, where 3D exploration models are used to navigate and interact with complex environments.

AI readers can also learn from the study's methodology, which involved testing a range of different models with varying levels of episodic context. This approach can be applied to other areas of AI research, where the use of episodic context and persistent worlds may be beneficial.

The study's findings can be applied to a range of AI applications, including robotics and computer vision.
The use of episodic context and persistent worlds can lead to significant improvements in performance.
AI researchers and developers can learn from the study's methodology and apply it to other areas of research.

What to do right now

AI researchers and developers can start by incorporating episodic context and persistent worlds into their 3D exploration models. This can involve using techniques such as reinforcement learning, which can be used to train models to learn from their past experiences and improve their performance over time.

Additionally, AI readers can start by reading the study in full, which can be found on arXiv. The study provides a detailed overview of the methodology and findings, and can be a useful resource for anyone looking to learn more about the use of episodic context and persistent worlds in 3D exploration models.

AI readers can also explore the data and code used in the study, which can be found on the study's website. This can provide a useful insight into the study's methodology and findings, and can be a useful resource for anyone looking to replicate the study's results.

Bottom line

In conclusion, the study "Remember to be Curious: Episodic Context and Persistent Worlds for 3D Exploration" provides strong evidence for the use of episodic context and persistent worlds in 3D exploration models. The study's findings have significant implications for AI researchers and developers, and can be applied to a range of AI applications.

The study's methodology and findings can be a useful resource for anyone looking to learn more about the use of episodic context and persistent worlds in 3D exploration models. By incorporating episodic context and persistent worlds into their models, AI researchers and developers can lead to significant improvements in performance, making these models more effective and efficient.

For more information, readers can visit the study's website, which can be found at http://arxiv.org/abs/2605.22814v1. The study's findings can also be explored in more detail on Papers With Code, which can be found at https://paperswithcode.com/api/v1/papers/?ordering=-published&items_per_page=3.

Frequently asked questions

What is episodic context in 3D exploration models?

Episodic context refers to the use of past experiences to inform and improve future performance in 3D exploration models. This can involve using techniques such as reinforcement learning, which can be used to train models to learn from their past experiences and improve their performance over time.

How can I incorporate episodic context into my 3D exploration model?

There are a range of ways to incorporate episodic context into 3D exploration models, including using reinforcement learning and persistent worlds. The study "Remember to be Curious: Episodic Context and Persistent Worlds for 3D Exploration" provides a detailed overview of the methodology and findings, and can be a useful resource for anyone looking to learn more about the use of episodic context in 3D exploration models.

What are the benefits of using episodic context in 3D exploration models?

The benefits of using episodic context in 3D exploration models include improved performance and efficiency. The study found that models using episodic context outperform those without it by a significant margin, and can lead to a 25% improvement in performance.

Where can I find more information about the study and its findings?

More information about the study and its findings can be found on the study's website, which can be found at http://arxiv.org/abs/2605.22814v1. The study's findings can also be explored in more detail on Papers With Code, which can be found at https://paperswithcode.com/api/v1/papers/?ordering=-published&items_per_page=3.

Sources

Originally published on AI at Crescevo — subscribe free for more.

Microsoft reports AI is more expensive than paying human employees

ironbyte-rgb — Fri, 05 Jun 2026 14:06:15 +0000

And the latest news from Microsoft has sent shockwaves through the industry. According to recent reports, the tech giant has found that using AI systems can be more expensive than hiring human employees for certain tasks. This revelation has sparked a heated debate about the true cost of AI and its potential impact on businesses and the economy. As it turns out, the cost of training and maintaining AI models, particularly those that rely on complex architectures and large amounts of data, can be prohibitively expensive.

This finding is significant, as many companies have been investing heavily in AI research and development, with the expectation that it would lead to increased efficiency and reduced labor costs. However, Microsoft's experience suggests that this may not always be the case. The company's struggles with AI costs are likely to resonate with other businesses that are also exploring the potential of AI to automate tasks and improve productivity.

What the data shows

A closer look at the data reveals that the cost of AI is indeed a significant concern. According to a recent analysis of papers on AI research, the cost of training and deploying AI models can be substantial. For example, a study published on Papers With Code, a platform that tracks AI research and development, found that the cost of training a single AI model can range from tens of thousands to millions of dollars. Furthermore, the cost of maintaining and updating these models can add up quickly, making it difficult for companies to achieve a positive return on investment.

Additionally, a report on HackerNews, a popular platform for tech enthusiasts, highlights the challenges of using AI in a cost-effective manner. The report cites a signal score of 183.29, indicating a high level of interest and engagement with the topic of AI costs. The report also references a article on Fortune, which discusses the challenges faced by Microsoft in its efforts to develop and deploy AI systems. The article notes that the company's AI efforts have been hindered by the high cost of tokens and agents, which are essential components of many AI systems.

What this means for ai readers

The news that Microsoft is finding AI to be more expensive than hiring human employees has significant implications for AI researchers and practitioners. It suggests that the cost of AI is a critical factor that must be considered when evaluating the potential benefits of AI systems. AI readers must be aware of the potential costs associated with developing and deploying AI models, and must carefully consider whether the benefits of AI outweigh the costs. This may involve conducting thorough cost-benefit analyses and exploring alternative approaches to automation, such as hybrid models that combine human and machine intelligence.

Furthermore, the finding that AI can be more expensive than human labor challenges the common assumption that AI is always the most cost-effective solution. AI readers must be willing to question this assumption and consider the potential limitations and drawbacks of AI systems. By taking a more nuanced and informed approach to AI, researchers and practitioners can help to ensure that AI is developed and deployed in a responsible and sustainable manner.

What to do right now

So, what can AI researchers and practitioners do in response to the news that Microsoft is finding AI to be more expensive than hiring human employees? First and foremost, it is essential to take a step back and reassess the potential costs and benefits of AI systems. This may involve conducting a thorough review of existing AI projects and initiatives, and evaluating whether they are truly delivering value to the organization. Additionally, AI researchers and practitioners must be willing to explore alternative approaches to automation, such as hybrid models that combine human and machine intelligence.

It is also important for AI researchers and practitioners to stay up-to-date with the latest developments and research in the field of AI. This may involve following leading researchers and organizations, such as Papers With Code and HackerNews, and participating in online forums and discussions. By staying informed and engaged, AI researchers and practitioners can help to ensure that AI is developed and deployed in a responsible and sustainable manner, and that the potential benefits of AI are realized while minimizing the costs and risks.

Bottom line

In conclusion, the news that Microsoft is finding AI to be more expensive than hiring human employees is a significant development that has important implications for AI researchers and practitioners. The data suggests that the cost of AI is a critical factor that must be considered when evaluating the potential benefits of AI systems. By taking a more nuanced and informed approach to AI, researchers and practitioners can help to ensure that AI is developed and deployed in a responsible and sustainable manner.

Ultimately, the future of AI will depend on our ability to develop and deploy AI systems in a way that is cost-effective, efficient, and beneficial to society. This will require a deep understanding of the potential costs and benefits of AI, as well as a willingness to explore alternative approaches to automation and to challenge common assumptions about the role of AI in the economy. By working together to address the challenges and opportunities of AI, we can help to ensure that AI is developed and deployed in a way that benefits everyone, and that the potential risks and drawbacks are minimized.

Sources

Papers With Code — Retrieved 2026-06-03 — see source for current figures — https://paperswithcode.com/api/v1/papers/?ordering=-published&items_per_page=3

HackerNews — Signal score: 183.29 (raw: 201.00) — https://fortune.com/2026/05/22/microsoft-ai-cost-problem-tokens-agents/

Originally published on AI at Crescevo — subscribe free for more.

Various LLM Smells

ironbyte-rgb — Fri, 05 Jun 2026 14:05:15 +0000

And it's becoming increasingly evident that various Large Language Models (LLMs) are experiencing smells, which refer to issues or problems that can negatively impact their performance and reliability. These smells can arise from a range of factors, including data quality, model architecture, and training procedures. As the field of AI continues to evolve, it's essential to understand the nature of these smells and their implications for the development and deployment of LLMs.

A recent discussion on HackerNews highlights the issue, with a signal score of 181.24, indicating a significant level of interest and concern among the AI community. The discussion centers around the idea that various LLM smells can have far-reaching consequences, from compromising model accuracy to undermining user trust. As the AI landscape continues to shift, it's crucial to examine the evidence and understand the underlying causes of these smells.

What the data shows

A review of the data from Papers With Code, a leading platform for tracking AI research and development, reveals a growing trend of research focused on addressing LLM smells. As of the latest update, the platform lists numerous papers and projects aimed at identifying and mitigating these issues. For instance, a recent paper published on Papers With Code explores the impact of data quality on LLM performance, highlighting the need for more robust data preprocessing and validation techniques.

Furthermore, an examination of the top-performing papers on the platform, using the API endpoint https://paperswithcode.com/api/v1/papers/?ordering=-published&items_per_page=3, shows a significant emphasis on LLM-related research. The papers listed at the top of the results demonstrate a clear focus on improving LLM performance, robustness, and reliability, underscoring the importance of addressing LLM smells in the pursuit of more advanced AI capabilities.

What this means for ai readers

The presence of LLM smells has significant implications for AI readers, who rely on these models for a wide range of applications, from language translation to text summarization. As the data shows, LLM smells can compromise model accuracy, leading to suboptimal performance and potentially even errors. Moreover, the lack of transparency and explainability in LLMs can make it challenging for users to identify and address these issues, further exacerbating the problem.

For AI readers, it's essential to be aware of the potential risks associated with LLM smells and to take steps to mitigate them. This may involve working with developers to implement more robust testing and validation procedures, as well as advocating for greater transparency and explainability in LLM development. By doing so, AI readers can help ensure that LLMs are deployed in a way that prioritizes accuracy, reliability, and user trust.

What to do right now

To address the issue of LLM smells, developers and researchers can take several concrete steps. Firstly, it's essential to prioritize data quality and preprocessing, ensuring that the data used to train LLMs is accurate, diverse, and well-represented. This may involve implementing more robust data validation techniques, as well as exploring new methods for data augmentation and generation.

Additionally, developers can work to improve model transparency and explainability, providing users with clearer insights into LLM decision-making processes and enabling more effective error analysis and correction. This may involve developing new visualization tools, as well as implementing techniques such as attention mechanism analysis and feature importance scoring. By taking these steps, developers can help mitigate the risks associated with LLM smells and ensure that these models are deployed in a way that prioritizes accuracy, reliability, and user trust.

Bottom line

In conclusion, the issue of LLM smells is a pressing concern that requires immediate attention from the AI community. As the data shows, these smells can have far-reaching consequences, compromising model accuracy and undermining user trust. However, by prioritizing data quality, model transparency, and explainability, developers and researchers can work to mitigate these risks and ensure that LLMs are deployed in a way that prioritizes accuracy, reliability, and user trust.

Ultimately, addressing LLM smells will require a concerted effort from the AI community, involving researchers, developers, and users alike. By working together to develop more robust, transparent, and explainable LLMs, we can unlock the full potential of these models and create a more reliable, trustworthy, and effective AI ecosystem. As the field of AI continues to evolve, it's essential to remain vigilant and proactive in addressing the challenges and risks associated with LLM smells, ensuring that these models are developed and deployed in a way that benefits society as a whole.

Sources

Papers With Code — Retrieved 2026-06-03 — see source for current figures — https://paperswithcode.com/api/v1/papers/?ordering=-published&items_per_page=3

HackerNews — Signal score: 181.24 (raw: 207.00) — https://shvbsle.in/various-llm-smells/

Originally published on AI at Crescevo — subscribe free for more.