It has never been easier or more cost-effective to develop software, thanks to AI-powered development assistance and a wealth of open-source tools. AI assistance, done well, can significantly speed up the development process, reduce errors, and lower costs. As a result, development is now cheaper than ever before.
While AI and open-source tools make development cheaper, it’s important not to get lost in complexity. Over-reliance on too many tools or overly complicated solutions can negate the cost benefits and slow down progress. Keep things simple and focus on what truly adds value. I think that the culprit behind negative feelings towards AI development assistance, through experience, is due to over-complicated projects, which can overwhelm developers, rise AI coding submission errors, and ultimately reduce the perceived value of AI assistance. And also making development decisions without a clear understanding of the tools and their implications can lead to inefficiencies and increased costs.
Breaking down complex tasks into smaller, manageable steps can help maintain focus and efficiency, for the engineer, and the AI assistant. By taking baby steps, you can gradually build robust solutions without getting overwhelmed by complexity. Sloppy code and rushed development can lead to AI submission obfuscation, making it harder for the AI to understand and assist effectively, ultimately reducing the benefits of AI-powered development assistance. The engineer also needs to maintain clear communication with the AI assistant, providing well-structured prompts, and ultimately context, to ensure that the AI can provide accurate and relevant assistance throughout the development process.
Open-source tools can significantly reduce development costs and accelerate progress. However, it’s important to carefully evaluate and choose the right tools for your specific needs. Avoid blindly adopting every new tool, and ensure that the tools you use align with your project’s goals and long-term maintainability. It can be tempting to use every AI suggested tool or library, but doing so without proper consideration can lead to unnecessary complexity and potential integration issues. The better the engineer stears the AI assistant, the more effectively the AI can provide relevant and valuable assistance, ultimately enhancing development efficiency and cost-effectiveness.
Manual intervention is still sometimes necessary to ensure the quality and correctness of the development process–either for quick adjustments, critical tasks that require human judgment, or for resolving issues that the AI cannot handle effectively.
Monitoring and managing costs is crucial to maintaining the benefits of cheaper development. Regularly review your expenses, identify areas where you can optimize, and make informed decisions to ensure that your development efforts remain cost-effective. There are more and more cheaper LLMs and AI-powered tools emerging, making it essential to stay informed about cost-effective options and evaluate their potential impact on your development process.
Also, the intermittent usage of different LLMs for a task or project can help optimize costs and leverage the strengths of various models. By strategically selecting the most suitable LLM for each specific requirement, you can maximize efficiency while keeping expenses under control.
Regularly assess the effectiveness of your tools and development processes. Identify bottlenecks, inefficiencies, and areas for improvement. By continuously evaluating and refining your approach, you can maintain high productivity and cost efficiency. Just don’t forget to feed less complex tasks to the AI, and prioritize baby steps from initial implementation to full-scale deployment.
Keeping up with the latest trends and advancements in the software development industry can help you leverage new opportunities and stay competitive. Embrace relevant innovations that can enhance your development process and reduce costs.
Fostering a culture of collaboration and knowledge sharing within your team can lead to more efficient development and better problem-solving. Encourage open communication, code reviews, and the sharing of best practices to maximize the benefits of cheaper development.
Prioritizing quality over quantity ensures that your development efforts result in reliable, maintainable, and valuable software. Avoid cutting corners to save time or costs, as this can lead to technical debt and higher expenses in the long run. If building a software from scratch, focus on creating a solid foundation before adding more features or complexity. Always through baby steps to gradually enhance your software while maintaining quality. Always think in terms of incremental improvements and iterative development to ensure sustainable progress.
Just don’t feed a complex assembly of tasks to the AI all at once; break them down into manually-intervened, manageable steps to ensure better results and maintain control over the development process.
The landscape of software development has changed dramatically, making it more accessible and cost-effective than ever before. By leveraging AI assistance, open-source tools, and best practices, while maintaining simplicity, cost efficiency, and quality, you can achieve remarkable results at a fraction of the traditional cost. Embrace these strategies to stay ahead in the evolving world of software development.
]]>I’m a big user of OpenCode and it fits perfectly into my workflow for developing and managing macOS applications. But lately, I’ve been exploring the idea of running an MCP server on top of some macOS apps API to streamline my creativity process even further. I am already seeing some promising results with this approach. Some macOS apps already provide APIs that are well-suited for this kind of integration, making it easier to implement and experiment with MCP servers. One of them is MindNode, a mind mapping application that offers a robust API for creating and managing nodes programmatically. But lately, this app has been shipping with an integrated MCP server, which opens up new possibilities for automation and interaction within the app itself. Quickly, I wanted to have my iWork apps, like Pages and Numbers, also support this kind of integration to further enhance my productivity. It would be a game-changer for my workflow.
The iWork MCP integration project demonstrates how an MCP server can be layered on top of iWork apps, enabling enhanced automation and interaction capabilities. And it works surprisingly well, providing a seamless way to extend the functionality of these apps without modifying their core code. I was able to scaffold a Pages document with some tentative content for a novel using the MCP server, which significantly sped up my writing process and allowed me to focus more on creativity rather than manual formatting and structuring. The only hic was using latest Grok model as the prime LLM–it didn’t always generate perfect content, and the document formatting was horrible, but it was good enough to get the creative juices flowing.
Now, MCP servers are often criticized for adding complexity and potential security risks to macOS apps. Some engineers argue that layering an MCP server on top of an existing app could introduce vulnerabilities or performance issues. While these concerns are valid, I believe that with careful implementation and proper security measures, the benefits of enhanced automation and interaction can outweigh the potential downsides. It’s a trade-off that each developer needs to consider based on their specific use case and risk tolerance.
Looking ahead, I envision a future where MCP servers become a standard layer on top of macOS apps, providing a unified interface for automation and interaction. This could lead to a more modular and flexible ecosystem, where developers can easily extend the capabilities of existing apps without waiting for official API support. As more apps adopt this approach, we might see a shift in how macOS applications are designed and used, with MCP servers acting as a bridge between user creativity and app functionality.
In conclusion, running an MCP server on top of macOS apps offers exciting possibilities for enhancing productivity and creativity. While there are valid concerns regarding complexity and security, the potential benefits make it a worthwhile exploration for developers willing to navigate the trade-offs. As the ecosystem evolves, I am optimistic that MCP servers will play a significant role in shaping the future of macOS app development and user experience.
]]>While the CEOs of OpenAI, Anthropic, and others have called for a pause in AI development, the reality is that the global race for AI supremacy is accelerating and the competitive pressures are immense. Many nations and corporations are investing heavily in AI research and development, driven by the potential economic and strategic advantages that advanced AI systems can provide.
Some experts argue that a slowdown in AI development is necessary to ensure safety and ethical considerations, others contend that such efforts are unlikely to succeed in the face of intense competition. The challenge lies in balancing the need for innovation with the imperative to mitigate risks associated with powerful AI technologies.
The slowdown of AI development is further complicated by the fact that many nations are pursuing their own AI agendas, often with little regard for international norms or ethical considerations. This creates a fragmented landscape where some countries may prioritize rapid advancement over safety, potentially leading to a global arms race in AI capabilities.
Nations that are less concerned with ethical implications may continue to push the boundaries of AI development, while others may advocate for stricter regulations and oversight. This divergence in approaches can undermine collective efforts to slow down AI progress and ensure that it is developed responsibly.
With the rapid pace of AI advancements, the competitive pressures to achieve breakthroughs in AI capabilities are intensifying. Companies and nations are racing to develop more powerful AI systems, often prioritizing speed over safety. This competitive environment can lead to a “race to the bottom” scenario, where the focus shifts from responsible development to achieving technological supremacy at any cost.
But AI slowdown is not just a technological challenge; it is also a geopolitical one. The race for AI dominance is intertwined with national security concerns, economic competitiveness, and global influence. As a result, efforts to slow down AI development may be undermined by the strategic interests of various stakeholders, making it difficult to achieve a consensus on the need for caution and regulation.
Competitive pressures in the AI industry are a significant barrier to slowing down development. Companies and nations are incentivized to push the boundaries of AI capabilities to gain a competitive edge, often at the expense of safety and ethical considerations. This creates a challenging environment for those advocating for a slowdown, as they must contend with the economic and strategic motivations driving rapid AI advancement.
So expecting a global slowdown in AI development may be unrealistic, given the competitive pressures and strategic interests at play. Instead, efforts may need to focus on establishing international norms, regulations, and collaborative frameworks that prioritize safety and ethical considerations while still allowing for innovation and progress in AI technologies.
Most of humanity is not aware of the AI arms race, and the implications of this race are profound. The pursuit of AI supremacy is not just about technological advancement; it is also about national security, economic power, and global influence. As nations and corporations compete to develop the most advanced AI systems, the risks associated with these technologies increase, including potential misuse, unintended consequences, and ethical dilemmas.
Let’s not forget that the AI arms race is a complex issue that requires careful consideration and collaboration among stakeholders. While the competitive pressures driving AI development are significant, it is essential to recognize the potential risks and ethical implications associated with these technologies. By fostering dialogue, establishing international norms, and promoting responsible development practices, we can work towards a future where AI is developed in a safe and ethical manner, even in the face of intense competition.
Let’s not create a Skynet scenario. The future of AI development should be guided by principles that prioritize safety, ethics, and the well-being of humanity. But that requires a collective effort from governments, corporations, and the global community to ensure that AI technologies are developed responsibly and with consideration for their long-term impact on society.
]]>I’ve been doing tarot for the lst 20-ish years, and recently I came across a tarot deck that is about to enter the public domain. This got me thinking about how images, like those in the tarot deck, might be used to train AI models, and whether there’s a way to determine if a specific image has been included in a model’s training data.
The idea is simple: recreate the about to become public deck using AI, and then compare the generated images with the original ones to see if the model has been influenced by the specific images from the deck. If the generated images closely resemble the originals, it could indicate that the model has been fed by those specific images.
Now, my inquest is to reuse those to be released images for a maybe only personal deck, or maybe doing physical copies of the deck for resale. The ethical and legal implications of this are still something I’m trying to navigate, especially considering the images will soon be in the public domain. But if I’m giving trouble to my as regarding permissions and copyrights, it might be worth waiting until the images are officially in the public domain to avoid any potential issues. Still, the prototype phase of this experiment feels like a gray area, and I’m curious to see how it unfolds once the images are publicly available.
The thing is that images can be identified by unique patterns, features, or metadata that distinguish them from other images. AI models often learn these distinctive characteristics during training, which means that if a model has been fed a specific image, it might generate outputs that closely resemble the original image’s unique elements. By analyzing the similarities between generated images and the original ones, it may be possible to infer whether the model has been influenced by the specific images in question.
And this can be easily tested by generating multiple images using the AI model and comparing them with the original images from the tarot deck. By examining the degree of similarity, one can make an informed guess about whether the model has been influenced by the specific images in question.
Of course, this method is not foolproof. AI models can sometimes produce outputs that coincidentally resemble certain images without having been explicitly trained on them. Therefore, while this approach can provide valuable insights, it should be complemented with other investigative techniques and a healthy dose of skepticism. Ultimately, the goal is to better understand the relationship between training data and model outputs, especially as it pertains to images that are on the verge of entering the public domain.
The experiment involves several steps to test whether an AI model has been influenced by the specific images from the tarot deck:
By following these steps, one can systematically investigate whether an AI model has been fed by specific images, while keeping in mind the limitations and potential for coincidental similarities. It ain’t foolproof, but it provides a structured approach to understanding the influence of specific images on AI models. Without being too explicit on the placement of the elements within the images, this method allows for a careful and nuanced analysis of potential training data influence.
In conclusion, while it is challenging to definitively determine whether an AI model has been trained on specific images, the outlined experiment provides a structured approach to investigate potential influences. By carefully selecting images, creating detailed digital tracings, recording comprehensive metadata, generating AI images, and systematically comparing them, one can gain valuable insights into the relationship between training data and model outputs. However, it is essential to remain cautious and consider the possibility of coincidental similarities, as no method can guarantee absolute certainty, but only with a high probability.
]]>While the AI industry has been booming, with investments pouring in and startups sprouting up everywhere, there are growing concerns about the sustainability of this growth. The rapid advancements in AI technology have led to inflated valuations and a frenzy of investment that may not be grounded in reality.
Greediness and hype have driven many investors to pour money into AI startups without fully understanding the technology or the market potential. This has created a bubble that could burst if the promised returns do not materialize.
Many AI startups are struggling to generate revenue and are burning through cash at an alarming rate. If these companies fail to deliver on their promises, investors may lose confidence, leading to a sharp decline in valuations and a potential market crash.
And Chinese models are taking over the world, with their advanced capabilities and lower costs, they are outcompeting many Western AI companies, but yet using the same strategies of American AI companies, by stealing through distillage.
I’m not an economist, and I mostly read the big titles of the news, sometimes delving into the details, but I have a few thoughts on the current state of AI investments. The AI industry has been experiencing a surge in investment, with venture capitalists and tech giants pouring money into startups and research initiatives. However, this rapid influx of capital has led to inflated valuations and a frenzy of hype that may not be sustainable in the long run.
Investment abherrations are evident in the market, with some AI startups receiving astronomical valuations despite having little to no revenue or proven business models. This has created a bubble-like environment where investors are chasing the next big thing without fully understanding the risks involved.
And restricting the Chinese AI companies from accessing AI chips and other resources may not be a long-term solution, as they are likely to find alternative ways to develop their AI capabilities. This could lead to a more competitive landscape, with Chinese companies potentially outpacing their Western counterparts in the AI race. GLM-5.2 and Kimi K3 being the latest examples of Chinese AI models that are gaining traction in the market, showcasing their advanced capabilities and potential to disrupt the industry, without the technical and financial support like in the West.
This bubble burst could have significant implications for investors, particularly those who have invested their retirement funds in AI startups. If the bubble bursts, many investors may face substantial losses, potentially jeopardizing their financial security and retirement plans.
Mutual funds and pension funds that have invested heavily in AI companies will also be at risk, potentially leading to a ripple effect throughout the financial markets. This could result in a broader economic downturn, affecting not only individual investors but also the overall stability of the financial system.
If the AI bubble bursts, it is possible that only a few companies will emerge as winners, while many others may face significant challenges. Apple, with its strong brand, loyal customer base, and diversified product portfolio, may be well-positioned to weather the storm and continue to thrive in the AI space.
Apple’s focus on privacy, security, and user experience could give it a competitive advantage over other companies that may struggle to maintain consumer trust in the wake of a market downturn. Additionally, Apple’s financial stability and strong cash reserves may allow it to invest in AI research and development, further solidifying its position as a leader in the industry. And let’s face it, Apple is not an AI company, but it is a technology company that has been able to integrate AI into its products and services effectively. This could give Apple an edge in the market, as it may be able to leverage its existing ecosystem and customer base to drive adoption of AI technologies.
]]>I once dreamt of a world where everything was tokenized, where DAOs governed every aspect of our lives, and where decentralization was the norm. In this world, we would have no need for traditional governments or corporations, as DAOs would handle everything from healthcare to education to transportation.
It was a utopia of sorts, where decisions were made collectively, and everyone had a voice. The power of the people was truly realized, and the potential for innovation was limitless.
But actually, the reality was far from this dream. The world of DAOs was plagued with issues of governance, security, and scalability. Many DAOs were poorly designed, leading to conflicts and inefficiencies. The promise of decentralization often clashed with the need for effective decision-making, resulting in gridlock and frustration.
Then came the rise of Central Bank Digital Currencies (CBDCs), which promised to bring stability and efficiency to the financial system. However, they also introduced new challenges, such as privacy concerns and the potential for increased government control.
But CBDCs are heavy and slow to changes, and they often lack the flexibility and innovation that DAOs can offer. At least, the liberal DAOs are faster than the CBDCs DAOs, which are often bogged down by bureaucracy and regulation.
Time, time is what we need to make the world a better place. We need to find a balance between the ideals of decentralization and the practicalities of governance. We need to learn from the mistakes of the past and build a future that is both innovative and sustainable.
But CBDCs DAOs are slow and heavy, and they often lack the flexibility and innovation that liberal DAOs can offer…
Despite the challenges, I believe that DAOs have the potential to reshape our world for the better. Or even for the worse, depending on how we approach their development and implementation. The key is to strike a balance between decentralization and effective governance, ensuring that DAOs can operate efficiently while still empowering individuals.
But still, are DAOs better that traditional organizations? Are they more efficient, more transparent, and more equitable? I think they are different, but not necessarily better nor worse. They offer a new way of organizing and governing, but they also come with their own set of challenges and limitations.
Let’s take a bitcoin wallet as an example. It is a decentralized system that allows for peer-to-peer transactions without the need for intermediaries. However, it also requires users to take on the responsibility of securing their own funds, which can be a daunting task for those who are not tech-savvy.
The technology of bitcoin and blockchain is difficult to grasp for too many people, and it can be intimidating to navigate the complexities of decentralized systems. Even too much. Only tech-savvy individuals can fully understand and utilize these systems, which can create barriers to entry and limit their adoption.
Is it fair to say that DAOs and decentralized systems are only accessible to those who are already tech-savvy? What about the unbanked, the marginalized, and those who lack access to technology? How can we ensure that these systems are inclusive and equitable for all?
And what about the techy baked, those who are deeply entrenched in the world of technology and blockchain? Are they the only ones who can truly benefit from these systems, or is there a way to make them more accessible to a wider audience?
For the moment, it seems that the world of DAOs and decentralized systems is still largely dominated by those who are already familiar with technology and blockchain. However, there is hope that as these systems continue to evolve and mature, they will become more user-friendly and accessible to a wider range of people.
Take CBDCs for example, they are designed to be more user-friendly and accessible to the general public, with the goal of promoting financial inclusion. However, they also come with their own set of challenges and limitations, such as privacy concerns and the potential for increased government control.
Let’s say that CBDCs are like national IDs, with their own NFTs and unique identifiers. They can be used to access a wide range of services and benefits, from healthcare to education to social welfare programs. However, they also raise concerns about privacy and government surveillance, as well as the potential for exclusion and discrimination.
But still, we aren’t talking about the unbanked, we are talking about the techy baked. The ones who are already familiar with technology and blockchain, and who have the skills and knowledge to navigate these systems. For them, CBDCs may offer a more user-friendly and accessible option, but for others, they may still be out of reach.
In conclusion, the world of DAOs and decentralized systems is still evolving, and there are many challenges and limitations to overcome. However, there is also great potential for innovation and positive change. By finding a balance between decentralization and effective governance, we can create a future that is both inclusive and equitable for all.
]]>If you haven’t tried GLM-5.2 yet, go do it now. It’s a Chinese distilled model that is outperforming most American models in terms of price and performance. The implications for the American AI industry are profound, and we need to ask ourselves some hard questions.
Is using Chinese distilled models stealing the future of AI from the US? Are American models’ prices competitive with Chinese ones? What does this mean for the future of AI in America? And what does it mean for the future of the internet? Will firewalls and censorship become the norm in the US as they are in China to protect the domestic AI industry? These are questions that need to be answered in the next 12 months, and we need to be prepared for the answers.
Because the competition is fierce, and the stakes are high. The American AI industry is at a crossroads, and we need to make sure that we are on the right path. We need to be prepared for the future, and we need to be ready to adapt to the changing landscape of AI. The next 12 months will be critical for the future of AI in America, and we need to be ready for whatever comes our way.
Now, I’m Canadian, and Mr. Trump doesn’t like us very much, we don’t shine enough for him. And eventhough I don’t like him either, I do have to say that he is right about one thing: the Chinese are coming, and they are coming fast. The American AI industry needs to wake up and smell the coffee, because the Chinese are not going to wait for us to catch up. They are going to take the lead, and we need to be ready for that. We need to be prepared for the future of AI in America, and we need to be ready to adapt to the changing landscape of AI.
I’ve been testing GLM-5.2 for the last month, and I have to say that I’m impressed. The model is fast, accurate, and cheap. It’s a Chinese distilled model that is outperforming most American models in terms of price and performance. The implications for the American AI industry are profound, and we need to ask ourselves some hard questions. Can American models’ prices compete with Chinese ones? Are American models’ prices competitive with Chinese ones? No. Not even close.
The performance of GLM-5.2 is impressive, and the price is unbeatable.
Let’s imagine a world, like now, where national internet ports are mostly opened, pirating is rampant, and the Chinese are distilling American models and selling them back to us at a fraction of the price. Is this stealing the future of AI from the US? I would say yes. The Chinese are taking their technology and using it to their advantage, while the Americans and the rest of the buzzword eaters are using Claude Code and OpenAI GPT Codex models that are overpriced and similarly performing. The Chinese are stealing the stealers, and the Americans, or the whole internet enchilada, is letting them do it.
Should we cry for Anthropic and OpenAI? Where they stole the whole freaking internet, dark and light, and now they are being stolen from? Should we cry for the American AI industry? I haven’t been using Mistral models, mostly because they are not as marketed here in Canada, nor Cohere which is Canadian. I’ve been using OpenAI since GPT-3 and Claude models since Claude 2.
The Chinese have become good at copying stuff. We have seen it it the last 10 years on Amazon for example, where Chinese sellers have been copying American products and selling them at a fraction of the price. The same is happening in the AI industry.
Now we’ve mostly heard the prosecutions of Anthropic and OpenAI for stealing the book authors’ works, and some paywalled content, but in the end, the Chinese are stealing the stealers…
We must take this into account, because billions of dollars have been invested in the American AI industry, and if the Chinese are stealing the stealers, then the American AI bubble is going to burst soon. The American AI industry is at a crossroads, and we need to make sure that we are on the right path. At least for me as a Canadian and good friend the American AI industry.
Now, Mr. Trump would say that I should pay the whole enchilada for the American AI industry, but I don’t think that is the right path. I want the whole AI stuff, cheap. Really cheap…
GLM-5.2 is flabbergastingly good, and the price is unbeatable. The Chinese are not taking the lead in this way, but we are seeing a trend of model to distillation, and the Chinese are taking the lead in this way. The American AI industry needs to wake up and smell the new teen spirit, because the Chinese are not going to wait for us to catch up. They are going to take the lead, with undistillation, if I may say so, by inflating weights and biases, and we need to be ready for that. We need to be prepared for the future of AI in America, and we need to be ready to adapt to the changing landscape of AI.
At least for me as a Canadian and good friend of the American AI industry…
]]>Two weeks ago, I wrote a post about my experience with AI and how it has helped me in my work. Since then, I’ve had the opportunity to work on a third app in just one month, and I wanted to share the story of how it all came together.
While the app is still in its early stages, I am excited about the potential it has to help me study astrology and astronomy. The app is designed to provide users with a comprehensive understanding of the stars and planets, and how they influence our lives. It will also offer personalized insights based on the user’s birth chart, making it a valuable tool for anyone interested in astrology.
The (short) journey of creating this third app has been both challenging and rewarding. I started by identifying the key features that would make the app unique and valuable to users. This involved extensive research into astrology and astronomy, as well as gathering UI/UX experiences from other astrology apps. But still, I needed to figure out how to use the Swiss Ephemeris, which is a critical component for accurate astrological calculations. I first lurked around the Swiss Ephemeris website and read through the documentation, but what if somebody had a starting point done that I could reuse? I searched online and found Eduardo’s AstroMalik for macOS Swift app, that wa ready to be compiled on macOS.
The AstroMalik app was a great starting point for me, as it provided a working implementation of the Swiss Ephemeris in a Swift app. It also had many interesting features prebuilt, such as a user-friendly interface and the ability to generate birth charts.
I wrote to Eduardo to thank him for his work sharing. He answered a few hours later, and within Eduardo’s reply, he slipped me a little secret: he had used AI assistance to help him develop and implement his app. This was a revelation for me, as I had been struggling use AI assistance to adapt his work. Now I realized that I could simply fork his app and continue to build upon it with AI assistance also. I felt less irritated and more confident in my ability to create a successful app.
I don’t have the answer of why I prefer that the project be done by a machine from the beginning, instead of continuing one done by a human, but I guess it is because I feel that the machine can be more objective, normative, and efficient in its approach. Continuing a project that was started by a human can be challenging, as it may not align with the normative machine’s way of thinking and may require more effort to adapt and modify. However, I also recognize that there are benefits to working with human-created projects, such as the ability to learn from their experience and insights.
With the help of AI, I was able to quickly refactor the code and implement other features that I needed for my app. Translation happened to be the costlier part of the app, but I was able to use AI to generate multiple complex views, and as usual, some tasks and tweaks have been manually done. The AI assistance allowed me to focus on the creative aspects of the app, while also ensuring that the technical details were handled efficiently.
I quickly realized that the AI model I was using for translation was costing me a lot of credits/money, and I needed to find a more cost-effective solution. I had used a mix of Claude Sonnet 5 and GPT-Codex 5.3 for the translation. After spending 70% of my monthly budget (70% of Github’s Max plan in a few hours) on translation, I decided to explore other options. I discovered that GLM-5.2 on OpenRouter was a more cost-effective solution, and I was able to use it to handle the rest of the translation tasks for my app. This allowed me to save money while still ensuring that the app was fully functional and user-friendly.
But not only for translation, I also used GLM-5.2 for code generation and refactoring, and it performed exceptionally well. The model was able to understand the context of my code and provide accurate suggestions for improvements. It was also able to generate new code snippets that were both efficient and effective.
Notes of high loads of translation and code generation tasks: I remarked that the model, especially
Claude Sonnet 5, had better costs for smaller loads and tasks than for big ones. Model reasoning seemed to be more efficient with smaller, more manageable tasks. The higher the amount of text to process, the more the model seemed to struggle with reasoning and providing accurate suggestions. It has to reason more on longuer tasks, duh! I found that breaking down larger tasks into smaller, more manageable ones allowed the model to perform better and provide more accurate results. It can mean higher interaction with the AI assistant, but it can be scripted also.
Overall, I was deceived by Claude Sonnet 5 over-reasonning functionality and over-priced model, but I was very impressed with the capabilities of GLM-5.2 along with its tiny price usage. I would highly recommend GLM-5.2 to anyone looking for a cost/job-effective solution for agentic tasks. It was a game-changer for me, and I am grateful for the opportunity to work with such a powerful tool. This model is quite impressive by its specs, and I am excited to see how it will continue to evolve and improve in the future.
Yes, I know what you’re thinking. AI is a hot topic these days, and it’s easy to get caught up in the hype. But hear me out: I believe that embracing AI is not just about accepting the technology itself, but also about embracing the potential it brings to our lives and society.
AI slop is like a machine-child’s art and gibberish. AI slop is the messy, imperfect output that AI can produce. It’s the result of the AI learning from vast amounts of data, and we about assembling it and controlling it. But just like a child’s art, AI slop can be beautiful, creative, and even insightful. It can challenge our assumptions, spark new ideas, and inspire us to think differently.
But still, a lot of people don’t accept AI slop…
AI has the potential to revolutionize various aspects of our lives, from healthcare and education to transportation and entertainment. It can help us solve complex problems, automate mundane tasks, and make more informed decisions. By embracing AI, we can leverage its capabilities to enhance our productivity, creativity, and overall quality of life.
Let’s look as some fictional examples of how AI can be beneficial:
Star Trek’s Data: In the Star Trek universe, the character Data is an android with advanced AI capabilities. He serves as a valuable member of the crew, assisting in problem-solving and decision-making. By embracing AI like Data, we can enhance our own abilities and achieve greater success in various endeavors.
Jarvis from Iron Man: In the Iron Man films, Tony Stark’s AI assistant Jarvis helps him manage his technology and make critical decisions. By embracing AI like Jarvis, we can streamline our workflows and improve our efficiency in both personal and professional settings.
The Doctor from Star Trek: Voyager: The Doctor is an AI holographic program that serves as the ship’s medical officer. He demonstrates the potential of AI in healthcare, providing medical expertise and assistance to the crew. By embracing AI like The Doctor, we can improve healthcare outcomes and provide better support to medical professionals.
AI, despite its impressive capabilities, is still in its infancy. It learns from the data we provide, and its understanding of the world is limited by the scope and quality of that data. Just like a child, AI needs guidance, education, and ethical boundaries to grow into a responsible and beneficial entity. By embracing AI, we acknowledge its potential while also recognizing our responsibility to shape its development in a way that aligns with our values and societal goals.
And we also need to master data ingestion and curation. The quality of the data we feed into AI systems directly impacts their performance and reliability. By embracing AI, we commit to providing it with accurate, diverse, and representative data, ensuring that it learns from a wide range of perspectives and experiences. But are we doing it the right way? Are we curating data responsibly, or are we inadvertently introducing biases and limitations into AI systems? Embracing AI means taking these questions seriously and striving for continuous improvement in our data practices.
AI slop refers to the imperfections and limitations inherent in current AI systems. While AI can be incredibly powerful, it is not infallible. Embracing AI slop means acknowledging these limitations and working with them, rather than expecting perfection. By understanding and accepting the imperfections of AI, we can better integrate it into our lives and make more informed decisions about its use.
And yet still, are we feeding it the right data? Are we curating it responsibly? Should we better parameterize data? Are the labels correct for the best AI possible? Is there still a lot to do for a slop-free AI? What would you say if the Voyager’s Doctor injected you with embalming fluid instead of a vaccine? Would you trust him if it was recently built? Are we ready to embrace AI telemedecine? Are we ready to embrace AI in our homes, workplaces, and communities? These are important questions that we must consider as we navigate the evolving landscape of AI technology.
Yet, a lot is coming…
]]>… so we might as well learn to live with it. The rise of artificial intelligence has sparked both excitement and fear, with many predicting a future where AI dominates every aspect of our lives. But instead of succumbing to panic, we should embrace the power of reading as a means to navigate this new world. And not just any navigation with AI, but a navigation that is informed, critical, and deeply human. It all starts with books and pre-school education.
Where AI starts to be fed, book reading is the first thing to be taught. The more we read, the more we understand the world around us, and the better equipped we are to make informed decisions about how to interact with AI. Reading allows us to develop critical thinking skills, empathy, and a sense of morality that can guide us in our interactions with AI systems.
If AI is going to be a part of our lives, we need to ensure that we are not left behind. Reading is a powerful tool that can help us stay informed and engaged with the world around us. By reading books, articles, and other materials, we can gain a deeper understanding of the implications of AI and how it may impact our lives. But even more importantly, reading can help us develop the skills and knowledge necessary to evaluate AI systems critically and make informed decisions about their use.
If AI makes us less intelligent, that is if we take for granted what is talked about in computer development environments and their usage of AI, we will be left behind. We need to ensure that we are not just passive consumers of AI, but active participants in shaping its development and use. Reading is a crucial part of this process, as it allows us to engage with the ideas and perspectives of others, and to develop our own informed opinions about the role of AI in society.
So developers using AI must know how to read, and evaluate code. Artists must know how to write (as I think it goes with reading) and prompt AI to create art. Architects must know how to read and understand the implications of AI in their designs. Cooks must know how to read and write to AI on how to get new recipes and create new dishes. And so on. Reading is the foundation upon which we can build a more informed and engaged society, one that is capable of navigating the challenges and opportunities presented by AI.
Reading can be a love-affair, or it can be hated by pure ignorance. But the more we read, the more we understand the world around us, and the better equipped we are to make informed decisions about how to interact with technology. The new generations (the “Z” and “Alpha” generations) are growing up in a world where AI is ubiquitous, and it is essential that they develop the skills and knowledge necessary to navigate this new landscape. By encouraging children to read from an early age, we insure that they are equipped with the tools they need to engage with AI in a thoughtful and informed manner.
I personally believe that reading is the solution to an AI apocalypse. By fostering a love of reading and encouraging critical thinking, we can ensure that future generations are prepared to navigate the challenges and opportunities presented by AI. It is our responsibility to equip our children with the skills and knowledge they need to thrive in a world where AI is an integral part of our lives. So let us make reading a priority, and let us ensure that all kids have access to the books and resources they need, and make them love it, to develop the skills and knowledge necessary to navigate this new world. By doing so, we can help ensure that the rise of AI is a positive force for humanity, rather than a source of fear and uncertainty.
]]>