DEV Community: Tambet Väli

CoPilot is for entertainment

Tambet Väli — Fri, 10 Apr 2026 09:43:43 +0000

It states that CoPilot is for entertainment purposes only, the MS post I would like to comment.

In company where I am, able to watch it knowing the aspects, AI made an estimation of income and people could not believe it, cutting it off in something like a gold rush - money was too important to notice that for a few thousand euros of input, AI expectancy is enough to prove only that this input is brought back; it was magnitudes larger per year, so even if AI error was in 3 it/in magnitudes, where it is iteration and in is input, one-time fee, in this unit the error must be like 3 magnitudes in optimistic direction to make our nerves tick. As a programmer, I am not interested in psychological needs of others, such as nervous system activity when they see money, which make their mental muscle tick and yield so much that they just cancel it; it's more interesting: substantially, there was good enough average estimation to conclude that with some error, an income would be made and whether it's a bright future as well, is an interesting adventure as soon as the real cost is covered. This poker-like discussion was way too much, as well as the voices people do as they shake their hands and fail buttons in poker, just bringing the lower level of clarity - while it's illegal to play poker well, remember the cards and I was trained for this when I was child because I studied magic and it's mostly about such tricks; it's still seen as trick in real poker field and not allowed: it's also funny if you cannot play this real game well, because the public wants to be "equal" and decide in where it's "shared" and "positioned".

You do not have to believe in exact nuanced story, but rather the archetypes - it's a typical story in our society, appearing in eyes of rich, intelligent, historic genius or jesus christ, each time the income game is seen as poker rather than calm estimation. What turns a genius into an introspective autist in history: a small land, and there were magnitudes less of people, produced stable amount of genius especially in it's favourite fields, but now with so many people genius must be able to form clubs and societies, and be a valuable member of their own community: in culture, which appears, they are not autist. Still, the jewish people tend to create collective genius effects, and for other collectives: it's as dangerous as genius for people; collective genius almost inevitably will produce material money and weapons.

What is the result?

I saw the people can produce reframing bias: the "hallucination" of an AI, when it was taught on one framework, for similar language or shallow meaning, but asked or conditioned in another framework: based on selection of details, it will make estimations as if this was another framework or paradigm. Such misfit can also produce ethical errors: if domain is different, such as militaries behaviour with peaceful cityzens in war, but the initial domain was treating pigs in farms and business rules, then maybe bluffing to people about how they were treated would be the absolute answer based on business ethics and how we could hide pig killing from children or neighbours, even to fight with the loud sound which would reach nearby blocks: almost every day in a productive factory. YouTube was just providing a series of videos about that - I could not even understand, is this AI fake or where this is happening, and what is happening exactly, but they were told to provide for "luxury dogs" and a typical image of half-killed dog as production entry was displayed in main image, but not inside the video I could not see it. Finally I did not understand is this advertisement or against this process. But it's not interesting enough to study deeply, because in fact I do not organize factories currently and I do not know what constitutes an "idea" in this constituent. I am not fair believer in synthetical meat for example and not vegan or fan of soya meat. I do not know why I live in ecosystem, but mathematically it's hard to exclude, like a quantum law: I still live in ecosystem. But now bring this framework and discussion context to train an AI: what it would say about holocaust now? This is kind of, an example of any human hallucination: such domain misfits or drugs or spiritual experiences and aura seeing, none of them distort the actual image or cognition as a primary mean; complete dysfunction of brain areas based on damage or growth problem, or when it's extremely tired or sick, is producing rather the real distortion: in desert, under every condition at the same time, the actual thing you see wrong has huge number of such mistakes. Normally, we reframe things and if input domain and output domain, or random condition, is different, the normal brain activity will produce hallucinations without change. Philosophy is enough in these conditions, but trained spiritual ethics to control spirit and inner world in right direction and not conclude to external world, is rather useful for real hallucination: while your brain, such as dreams which are real hallucinations, can block you from consciousness that such thing is an artefact, symbolical, metaphorical or fantasy field whatever is the current synthesis, and in dreamtime the muscle control is limited which is real protection against such imagination realm to also take over the stdout channel if it's mocking the stdin with something definitely generated somehow, from something, now this is a safe hallucination. If you are sick, such as person dying in war and in addition to seeing death also physically damaged, tired and unable to heal up or rest properly in these conditions, fears and challenges of everybody, in movies they see hallucinations about what is going on, such as childhood memories, funny things which would, could happen with their body but do not, and hallucinations about solutions such as proper food but it disappears as you are awaken up by yourself, such as their woman or childhood dream or vacation from war with their dead family. Usually the muscle output is limited. Now in reframing hallucinations, our body is not ready because it seems straight go or suppose, and happens often in poker-like conditions of chance and money, or dreams such as women or not appearing similar to a criminal, some would fight to look like local saint or appear with women, better than they are. All this would produce a social hallucination, an infection syndrome, in those subjects and in actual reality behaviour which might reinduce the hallucination to it's producer; if nature or society counts this as money, it runs out and the thing appears like a witchcraft, a magic replaced by science: science of collapse, not a human ideal fairy tale which would be seen in working conditions.

In this case: human would produce worse ethics, but triggered by same synthetic output as an AI, which could be labelled as same "hallucination" in the same language; other AI errors, even in-productibility or inability create in some conditions, would be repeated by low intelligence in humans; they are not really able to resolve the conflicts of the rich, but sometimes the poor unintelligent person would think they won't have the chance, and rich are evil just because of money: this is extreme simplification in many problems we can see in business realities and hallucinations, where in material hallucination the money would move into the bankcrupcy or fall when seen otherwise, or there is unexpected surprise for someone who was initially rather stupid, but not in conditional change such as intelligent child making success of their father's fade-off money and company which seemed as a dead branch and was sold cheaply enough.

In something like counter-Turing test: AI ethics must be compared with humans; I have seen the same story repeated in some human "professionals", who could not answer the interdisciplinary challenge. Based on this test, on average line, different levels of AI in domains or general sphere would compete humans in same domain, and also less ethical intelligence vs. straight corruption and low motivation could be the winning factor of a general test, where sometimes human has more personal gain than AI would learn from patterns, even if AI repeats some kind of false scientific belief or public propaganda with less results.

Iterative Freedom

Tambet Väli — Fri, 17 Oct 2025 13:49:59 +0000

https://cloud.uibakery.io/share/Etck7En4t5/home - I made this webpage about open source in agile development. This continues topics of my posts here, and also links back here in the articles section, as I do not have separate popular article system myself.

I link it under my main webpage, at licence paragraph: https://spireason.neocities.org/.

An introduction to AI and Deep Learning

Tambet Väli — Tue, 15 Apr 2025 16:03:49 +0000

https://github.com/tambetvali/LaegnaAIBasics - this is the main page for my Deep Learning introduction.

Main page: this is the introduction document, where I tried to represent the ideas more simply, than they were presented to me.
- invariants.md: to also explain invariants, a central concept in developing AI.
- visualizingai.md: an intro, how to visualize multidimension to screen, to debug or learn an AI.
- compassion.md: to integrate AI systems more, as well with humans.

Iterative Freedom

Tambet Väli — Fri, 14 Mar 2025 11:53:58 +0000

Iterative Freedom: The Path to a Sustainable, Free Service for Users

The idea of Iterative Freedom revolves around creating a working, free service that can scale over time through small, manageable iterations. This concept, exemplified by chatbots or other AI systems, balances the freedom to create and serve with the financial sustainability required for growth.

It’s not just about offering a free chatbot or AI service; it’s about giving developers and users the freedom to evolve the product together—without the developer having to worry about money in the early stages of development.

Programmer's Agile Journey: Starting with a Free Service

A programmer's journey through iterations involves starting with a basic, free service and gradually adding more advanced features over time. The first version, or first iterations, is fully functional and free, but as the service scales, additional iterations with new features require payment—either through user contributions, ads, or a recurring trial model.

Key Elements of Agile Iterations:

First Iterations: Free and Functional: In the beginning, the system is built with minimal functionality to serve users. The first iteration (and its subsequent subversions) is free, and it should remain functional enough to fulfill the basic needs of users. Documentation is kept minimal but sufficient to guide the bot’s responses based on the programmer’s initial theoretical framework.
Preprogrammed Iterations: The chatbot follows an agile development process with preprogrammed iterations:
- The first version is free, as it’s the minimal viable product.
- Each successive iteration introduces new features, and these become paid versions once they are enabled.

This means that while the first iteration (and its updates) is free, as the developer progresses, more advanced versions of the chatbot with additional features can be enabled for a fee.

User Experience: Free to Start, Pay as You Grow

For users, the first iteration and its subversions are available free of charge, which allows them to start using the service immediately. As the developer progresses with new features, users may choose to pay for these advanced features once they’re ready.

Agile Development with Iterations: The concept of agile development plays a significant role here. Users can use the free version and access subversions of the bot. However, when the developer creates new features, these might require payment for access. The goal is to ensure that users have access to the latest, fully-functional version of the bot without unnecessary limitations in the early stages.
Developer Freedom with Iterations: The key idea here is that the developer’s freedom lies in the process of development. They don’t need to worry about monetizing the service immediately. The initial free versions serve as a base, but as the service evolves, the developer and users can decide together when it makes sense to charge for advanced versions.

This is where agile methodology plays an important role. As new features are added, they are provided iteratively, aligning with the developer’s progress. Users benefit from improvements, but they only pay for what they use, and they know in advance when they need to upgrade.

The Role of Ads and User-Driven Monetization

As the service progresses and more advanced features are developed, it’s important to have mechanisms for monetizing the service—but without making it disruptive for users or developers. Ads or user-driven monetization can allow the service to scale while remaining accessible.

Ads and Automated Revenue: Instead of charging users directly from the start, the service can incorporate ads to generate revenue. This allows for the system to remain free for users while providing enough income to keep the service running. As more users join, ads can be used to ensure that resources are available for both the basic and advanced features.
User-Driven Payments: While the initial versions are free, users can opt to pay for advanced features as they become available. This system ensures that developer freedom remains intact while still providing a way to support the service financially. This self-sustaining model creates a win-win for both the developers and the users: developers get paid for advanced features, and users can enjoy the service without upfront costs.

Developer Freedom: Not About Avoiding Money, But About Not Worrying About It

The core of developer freedom lies not in avoiding monetary flow, but in having the freedom to not worry about it. Developer freedom is about focusing on creating and innovating, without being burdened by the practicalities of managing payments, pricing, or monetization at the outset of development.

The Role of Managers in Enabling Developer Freedom

Freedom from Business Concerns: In a well-structured development environment, managers handle the business side of things—such as monetizing the service, negotiating payments, securing investments, and ensuring sustainable income. This freedom from business concerns allows developers to focus on building the project rather than worrying about finances.
Invisible Monetary Flow: For developers, the monetary flow should remain invisible to them. This can be achieved through ads, user payments, or other automated revenue-generating mechanisms. Developers are free to code, create, and iterate without needing to worry about where the next payment is coming from. This gives them the space to focus on what matters most: delivering a high-quality, innovative solution to users.
Managers Enable Freedom: It’s important to recognize that while developers enjoy freedom in terms of coding and creating, they still need managers to handle the practicalities of running the service—whether that’s monitoring user growth, handling payments, or negotiating with advertisers. Developers don’t need to become the managers; they can focus on what they do best, while managers ensure the financial sustainability of the service.

The Expensive Reality of AI: Balancing Free Services and Costs

Providing AI services, especially through chatbots, is expensive. Running the infrastructure, maintaining AI models, and ensuring reliability requires significant resources. The open-source community may not have the means to offer high-end AI services completely for free, and companies cannot continue to operate these services at a loss.

Thus, while we respect the ethos of free services, we must acknowledge that advanced AI services require financial backing. This backing can come from ads, user payments, or investors, but it should not interfere with the freedom of developers or users.

Revenue from Ads: As ads become part of the service, they can provide a way to cover operational costs. However, the ad revenue should be distributed fairly, not just focusing on the chatbot but on the entire service. For example, if the chatbot is only 10% of the service, it should receive only 10% of the revenue, while the rest of the income can support the infrastructure and other resources necessary to run the service smoothly.

Conclusion: Iterative Freedom in Action

The Iterative Freedom model allows developers to create and scale a service without being burdened by money concerns in the initial stages. As the system grows, monetization through ads, user payments, or advanced features can help sustain the service while maintaining the free access that users need.

Developer Freedom is about the ability to create without worrying about the financial side, made possible by managers and automated systems.
Users can enjoy the service for free while paying for advanced features as they are introduced.
Agile Iterations allow both developers and users to grow the service together in manageable, scalable steps, ensuring the service remains functional and sustainable.

By following this iterative, agile model, developers can keep the system free and accessible, while ensuring the financial stability needed to scale the service over time.

No Fake Freedom

Tambet Väli — Fri, 14 Mar 2025 10:46:11 +0000

Freedom vs. Witchcraft: The Illusion of Free Software

I've been on a quest to find truly free software and open-source solutions, and I can't help but feel that there’s a sort of "witchcraft" at play. It's as if some charlatans are peddling the idea of free software, only for me to discover their offerings are riddled with hidden costs and limitations. For those who might not have easy access to a bank card, this situation is frustrating, but for me, it feels more like illegal manipulation. I genuinely want to avoid these pitfalls for personal use.

The Illusion of "Community Freedom"

Take platforms like Odoo and various personal management systems. They advertise "free" versions, but what you often find is a lack of critical features, which makes these offerings feel more like demos.
I remember using advanced commercial software that had genuine limitations—like the early days of Windows with its demo discs that simulated functions without providing any real capability. That is not what I consider a "free version."

The Restrictions on Custom Chatbots

When it comes to creating custom chatbots, I could accept a small model being offered for free. As someone who lacks the resources to build models myself, these tools can be valuable. However, I also understand the need for systems to manage user load, which can result in slowness or temporary unavailability.
Platforms like ChatGPT have managed to balance offering free access while maintaining quality, but when "free" comes with limits—like a capped number of messages or users—it feels more like a "trial" than a legitimate free offering.
I wouldn't mind features like image generation or web searches being trial-based, as long as the core functionalities remain accessible for testing purposes.

The Burden of Unexpected Sign-Ins

Another frustration arises when I encounter platforms that suddenly require me to log in or provide personal information like phone numbers or credit card details. While obtaining a credit card might be easy for some, it's not the case for everyone, especially in regions where credit systems differ significantly.
Here’s the thing: Americans have self-made cars, free credit cards, and can get a phone without paying a penny. They even have WhatsApp for free, right? It's almost as if everything is free in the States! Meanwhile, Europeans are still buying their cars from a dealership for thousands of euros, and often need a phone number or bank details just to use "free" software. It’s like the definition of "free" changes depending on where you are. Over here, it’s not so free if I don’t already have assets, a phone number, or a credit card. And if I need to register, guess what? I’m asked for that phone number. It's like the true cost of "free" is having to prove you have enough resources to access it.

System Limitations vs. Full-Service Philosophy

I’ve encountered limited systems that are designed to offer "free" services but fail to deliver quality when you need them the most. It’s frustrating when a system shows an “Access Disabled” message because they have oversold their resources, but you didn’t know that until it’s too late.
On the other hand, a full-service system, even if it’s slower or has some limitations in performance, can still serve each user without the sudden, unexpected downtimes. I’d rather have something that's always up, even if not fast, than something that has a powerful system but fails when too many users try to access it. This is an important distinction—some systems prioritize resource distribution over providing perfect service, and I think that’s where the balance of "free" vs. "open-source" should be recognized.
I see value in starting with a small, agile system that evolves over time. In its first iterations, the focus should be on delivering core functionality to users, with some acceptable compromises in speed and scale. As I grow, I can optimize the system, improving quality while still maintaining the flexibility to scale and adapt.

The Open-Source vs. Free Debate

When I think of open-source, I’m looking for something that allows me to modify and iterate on my own terms, with full transparency about how the system works. These systems tend to prioritize freedom, which is valuable for long-term projects.
However, when I’m looking for free systems, I’m generally open to paying for advanced features that make sense for the platform's sustainability. There’s no problem with this, as long as the core features I need are available for free. I think of this as a tiered system, where the base functionality is free and the more powerful features are paid—this gives me access to developer-level features or even free article writing tools without hitting a wall. As long as the fundamentals are free and accessible, I can use the platform to serve others in advance while deciding if I want to unlock higher quality features later.
I have no issue with a system offering premium, advanced features for a price. As long as the basics serve my needs, I can optimize my system in its first iterations with small, functional tools. Let the more complex, high-quality iterations come later.

Conclusion

The journey to find truly free and open-source software is fraught with challenges. As users, we deserve transparency, genuine access, and a system that aligns with our needs without hidden agendas. The concept of "free" isn't always as simple as it sounds, especially when the expectations around resources differ across regions. It's time we call out the practices that make the "free" label feel more like a mirage than a reality.

Class Society in Open Source

Tambet Väli — Wed, 18 Dec 2024 15:42:23 +0000

I am new to this community, and I just want to write one initial post about what I have concerns with lately.

When we look at the Open Source world from programmers perspective, it's pretty complete - we have games, science and programming software, office tools and operating system(s). We also have some basic AI implementations, yet with expensive hardware making it less free.

Looking at these categories, these are things we actually do ourselves as programmers. We all need operating systems and we have heavy load of understanding of their limitations; many capable programmers also play games, even some on professional level. Science is a must-be for an engineer to do anything practical at all, everybody comes with some specific formulaes to calculate things about circles, acceleration or even the table of chemical elements. And, finally, using office tools is as trivial as having the pen and papers. We can relate ourselves to mathematicians, who turn coffee and cigarettes to formulaes, and to bridge builders, who revolutionalized the access; people, who invented new architectures and tools, are our kind. So, natively, each programmer knows at least some formulations of physics and chemistry, and in any forum we have someone, who is highly competent to discuss those topics.

So, what we can see, we have built a perfect set of tools we can easily use.

Now, looking at the world of cooking, carpeting, farming etc. - the tools from open source community are really very basic, struggling to cover the critical parts of functionality, and mostly the term "open source" refers to some unknown concepts such as "limited community editions", and it's impossible to convince Google to give some real open source - probably, because it does not exist in it's initial sources. When I happened to set up a CNC machine for wood cutting, and when I was looking for human resource management software, I really got into it - things, which are really so basic that they are worth of a few days of programmer, needing only a list or a controller, they are implemented in low-quality manner, the library versions are outdated, and the communities provide very funny guidelines about programming. Somebody worked on the code for 6 months, 10 years ago, and this version is still the most decent without any updates, and people are using it as a quality standard.

So we have kind of class society we kind of wanted to avoid - people, who know, how to code, are the privileged class of open source, who gets any service. The reason is very basic - as we have sciences, offices and utilizing-operating-system computers very much in programmer's lives, moreover the scientists themselves easily learn to code at least something and they are careful to make at least the basic algorithms work - when we want to know something about a waiter or carpenter, we hardly know any of their problems and solutions. Afaik in any programming company, there are programmers, who build house, grow plants, all of us do some carpeting and things such as fixing the cables; these are easy to learn for an engineer of any kind and indeed they do those things themselves - we have minorities around those less-developing areas, with many of us left them simply with boredom of having grandfather-invented technologies. Each programming company has some people, who work in farms for their fathers, who cook very carefully and who build something for themselves and construct their moble; indeed we are men - but these numbers are very small to create big, united collectives. Finally, those people really concentrate, for 6 months, on their almost-hobby, but they do not form long-term, stable teams.

I think this is because we don't imagine those things - theoretically, probably we really have implemented each of the important equations of those people, and also some UI's, but those look like coming from someone's school-time, containing simply an algorithm from something in the book. We don't have a range of usage scenarios, high level of competition etc., in those areas.

Thus, I would suggest to do something like "Open Project" in addition to "Open Source", where people from any field are invited to create a project, an AI prototype, screenshots of UI's and explanations of needs and use cases. Such team might involve less programmers and create more decent team of multiple areas. People, for example, from farming - they can be logical and clear-thinking, and they might understand the flow of screens and forms of calculations, and some of the possibilities, very naturally like men understand the wood. Basically, we need those people to carry out the project plans; and for different teams it might be requirements, specifications or even ai-assisted prototypes. We should have a friendly environment, where programmers, instead of looking for other programmers, look for such people.

There should also be business models for such things. In general, it's also hard case, why we have rather programmers in open source communities, than designers, testers etc. - while we have designs and testing of open source projects, it's hard to write specifically a home page of open source designer.