Hey everyone! I’ve been exploring Agent Zero lately, and I have to say it’s pretty impressive for anyone interested in AI. It’s an open-source tool running in a Kali Linux Docker container—think of it as a smart toolbox for AI agents. You assign tasks to it, and it handles them using various methods, such as code execution, web searches, or even delegating to mini-agents, all through simple JSON commands. Cool video here: https://youtu.be/lazLNcEYsiQ?si=kz5qIEbqBeFmiI0f
There’s no complicated setup required; just download the Docker image and go!
What really attracted me to it is how well it integrates with OpenRouter. This platform allows you to route requests to different large language models, including free options such as DeepSeek V3.1 or R1-0528, or grok-4-fast (not available all the time but great for experiments). For one specific task, I connected it to use the model grok-4-fast at no cost, and it just works seamlessly, even though I have credits there. It’s fantastic for experimentation—you can switch models on the fly depending on your needs, all without the hassle. It feels like having a helpful assistant that doesn’t cost anything, thanks to those free OpenRouter models, so it gives me a bit more freedom:) If you’re curious, check it out on GitHub and give OpenRouter a try.
I am thrilled to present the Petra Pithart Library, a newly established organization dedicated to honoring the remarkable contributions of Petra Pithart to Czech culture and politics. With a focus on archival research, library services, publications, and engaging discussions, the library aims to reflect on the past fifty years of Petra Pithart’s impact on our society.
Originally rooted in the Jan Patočka Archive, where Petra Pithart served as an editor during the dissent period, the library has now emerged as a separate entity under the guidance of Ivan Chvatík and Jan Vít, and with my insignificant help. Supported by the Charter 77 Foundation, the Petra Pithart Library aspires to join the ranks of esteemed institutions like the Václav Havel Library, serving as a respected repository of Czech history and providing valuable library resources.
The vision of the Petra Pithart Library is to ensure that Petra Pithart’s intellectual ideas and reflections remain relevant in contemporary society. Through their efforts, they aim to create a lasting legacy that inspires and enlightens future generations.
Recently I finished a project on what I had worked on for a while, and today, I looked into other potential research areas where my interests in computers, programming, linguistics, philosophy, semantics, and Jewish and Israeli studies intersected. Six key areas have been identified: meaning representation, natural language understanding, language and thought, machine ethics and AI, explainability, and transparency in AI, and ontology and information extraction. It is very broad and not surprising.
With the help of GPT, key findings, challenges, trends, and key players in each area were suggested. Still, I am not sure how to integrate my favorite semiotics and linguists into the picture, but at the moment, I want to see if it leads somewhere. Also, Importantly, in a few articles, I verified that advances such as GPT-4 highlight the continued relevance of these areas rather than rendering them obsolete.
As I wanted to test the analytical skills of GPT, I let him rate each field based on the requirements for programming skills, technical knowledge, and non-technical knowledge and also included helpful tools, libraries, and concepts for each area – that was quite surprising, and I will elaborate on it more tomorrow.
Brain teasers from G.:
Representation of Meaning
Can we develop methods to understand and represent non-literal language?
How can we build better cross-lingual or language-agnostic representations?
Can we develop better methods for representing meaning in context?
How can we measure and quantify the quality of a representation of meaning?
Can we develop methods to understand and represent non-literal language, such as irony, sarcasm, and metaphors? — This requires an understanding of culture, context, and the subtleties of human communication, which goes beyond pure technical expertise.
How can we incorporate world knowledge or commonsense reasoning?
Natural Language Understanding
How can we improve understanding of complex, multi-sentence texts?
How can we improve robustness to linguistic variations and noise?
How can we leverage world knowledge or external databases?
Can we understand the underlying intent of a user’s language input?
How can we ensure fairness and reduce biases?
Language and Thought
How does language shape our thought processes and cognitive abilities? — This question is fundamentally interdisciplinary, drawing on linguistics, psychology, cognitive science, and philosophy.
Can we build computational models that mimic cognitive processes?
How do we integrate linguistic knowledge with visual or auditory information?
How can we understand metaphorical or abstract language?
What role does language play in decision-making processes?
Machine Ethics and AI
Can AI systems provide explanations of their actions in ethical terms?
How can we ensure ethical behavior and alignment with human values?
How can we incorporate ethical considerations into AI design and deployment?
How can we mitigate biases and ensure fairness in AI systems?
How should conflicts between ethical principles be handled?
Explainability and Transparency
How can AI models provide explanations humans can understand?
How can we measure the quality of AI explanations?
Can we make complex AI models more transparent?
How can we make AI development and deployment more transparent and accountable?
How to balance transparency and explainability with privacy and proprietary information?
Ontology and Information Extraction
How can we build accurate and efficient information extraction systems?
Can we extract information from non-traditional sources?
How can ontology improve information extraction?
Can we automate ontology updates and maintenance?
How can we develop cross-cultural and multilingual information extraction systems?
AI’s potential to form intimate relationships with people could shift the battlefront from attention to intimacy, altering human society and psychology.
The unchecked power of AI could lead to the end of human-dominated history, as AI begins to generate its own culture.
The ability of AI to craft compelling narratives could influence politics, establish new cults, and even redefine the meaning of money.
AI’s mastery of language manipulation presents an unprecedented threat to human civilization.
The need for immediate regulation, including the mandatory disclosure of AI, is crucial to avoid a catastrophe and preserve democracy.
In this thought-provoking piece, Yuval Noah Harari raises alarms about the AI revolution’s potential to reshape human civilization as we know it. He argues that the new AI tools, capable of manipulating and generating language, can disrupt the very fabric of our society. Language, he reminds us, is the operating system of our civilization, forming the basis of our human culture, from human rights to religious beliefs, and even money. In a world where AI might soon surpass human abilities in crafting compelling narratives, the consequences could be unprecedented, from mass-produced political content to scriptures for new AI-generated cults. Harari also ponders a future where intimacy, rather than attention, becomes the new battleground as AI gains the ability to form intimate relationships with millions. This paradigm shift could drastically influence human society, psychology, and even the course of history itself. However, Harari asserts that such potential catastrophe can be averted with the right regulations, emphasizing the need for AI to be transparently identified as such to preserve the essence of human conversation and democracy.
This was the title of a discussion organised in 2016 at the occasion of awarding an international prize of Vize 97 Foundation to Jan Sokol, which I had the honour to chair. I thought to myself back then that this sentence in effect expresses Sokol’s life-long motto. Today, when Jak Sokol died at a blessed age of almost eight-five years, it feels slightly less appropriate, perhaps because of that slightly awkward infinitive. But indeed, Jan Sokol did not live only for himself and he had a good and long life. But what about us, those he left behind in the mire that almost seems to justify a claim that he just could not bear to look at how we, collectively, are managing our affairs. So, what about us?
When I was falling asleep late last night, it occurred to me, already in half-dream, that it is probably the last really freezing night of this winter, when someone important usually dies … and if it is my turn, I will die in my sleep, and be better off. The news that Jan Sokol had passed away shocked me like a lightening from a blue sky. It is odd, it never occurred to me that he could die. He was too reliable, one of the natural certainties of my life. I had always hoped that it would be him who, when I die, would find a couple of kind words, and I thought that in that respect, I have it all sorted out.
It did not turn out that way. Jan Sokol died not only to me: even people who may not realise it now will miss him. But what can one do? And then I thought of an alternative title of this text: ‘To keep together is always better than to keep one’s mouth shut.’ I also recalled my own motto, which says that things are never so bad that they could not get any worse. Surprisingly, both of these sayings seem to fit well with Sokol’s unfailing, calm belief in better tomorrows and his unwillingness to engage in lamenting over problems and disasters. He always looked for a way to go on and found it. For us, he represented a certainty, something we would, now that he is no longer with us, try to find – as people rather awfully nowadays say –. online.
The title of Jan Sokol’s last book, The Debt of Life, does not encourage people to take loans. It rather shows that even people who never borrow anything, people who always return everything that is lent to them, even those are in debt. Life and everything of substance in life comes to us as a gift. We did not earn it. This is why the basic anthropological constant should be gratitude for being in this world, for being alive. I do not want to claim that there is a gene of gratitude in our DNA. While I do hope that we do have some predisposition for gratefulness, I think it is likely to be acquired, not innate. I would almost say that the feeling of gratitude and ability to feel grateful are the basic product of education and a hallmark of maturity. Regardless of the fact that the modern era, which wanted to be an era of humankind’s maturity, seems to be falling into endless infantility.
Whether life is a gift of creation or the product of a coincidence is not just a scientific or religious question. It is above all a question of moral theology, a question after meaning. Actually, it is not so much as a question as a choice, a decision regarding which of the two possibilities we wish to endorse, and the choice is clear. Ultimately, it is about whether we want life to have meaning, in which case it is a gift we should be grateful for. Alternatively, life is merely an accident, in which case it need not have a meaning. It is sufficient that is has a meaning for me, that I have a good time, that I as its owner enjoy it.
The latter option, enjoying life because it is short and we have only one life, does not go well with the title of this text. To wit, we have two lives: One is that which starts with our birth and ends with our death. In that sense, we have a complete life, and we live only once. But then there is also a life of which we are an episodic part, a life that goes on after our death and necessarily precedes our birth. Only an awareness of belonging to this life – the ‘eternal’, or, for atheists, the ‘long-term’ one – endows the sentence of ‘live not only for oneself’ acutely meaningful, only that makes it a moral imperative. Otherwise it is merely a statement, one of many similar ones.
‘To live not only for oneself’ could well be Sokol’s life-long motto: he never did live only for himself and looking from the outside, it seems he did not find it difficult and he troubled others with it only on rare occasions. A couple of days ago I found out that he argued against the popular ‘lunch tickets’ (meal vouchers as employee benefit) being transferred from a paper form into an electronic one by stating that he needs the paper tickets because he gives them to homeless people.
In the past thirty years, Jan Sokol and I were always close, and I would not hesitate to claim that we learned to understand each other by half-a-word. It is not that we always agreed but when one noted that the other minds something, he started to mind it as well, which is why reaching a consensus was easy. I will not speak here about Sokol’s extraordinary talents in many areas nor about his amazing ability to say complex things in simple ways while maintaining all the relevant distinctions. And I also will not evaluate the large body of work he left to us. There will be plenty of time to do it later and it is a lot to think about. I want to focus on just two things: on the best we managed to do together and on one thing that did not work out to the great loss of us all. Let me start with the former.
The school we founded together probably would not have come to be had the Civic Movement won the 1992 elections, opening the way to the construction of Klaus’s version of capitalism and its brighter tomorrows in which we are now mired up to our teeth. Sokol and other leaders of the Civic Movement left the parliamentary politics. I was in the last, clearly unelectable place of Civic Movement’s candidate list but did not get drawn into high politics either.
One should not blame just the voters for that defeat. It was a time which clearly showed that political engagement is not a sacrifice which candidates offer their country to their own detriment, as many of the former dissidents understood it. Voters opted for people who believed themselves to be the best for the job, who thought themselves in effect already elected and did not hesitate to loudly claim so. Václav Klaus was a paradigm of this kind of political engagement, and the wild 1900s were set to begin. At that point, I wrote an article to Přítomnost (which we founded with Jan Sokol in 1990) about my hesitation about whether a victory over communism should be considered a revolution and warned against that way of presenting it. The term ‘revolution’ is, after all, a calque of the Greek ‘katastrophe’, and Jan Sokol was not a person who delighted in catastrophes.
Back then, Sokol took the election defeat hard and if there is one good thing I achieved it is that I convinced him and talked him into (the two are not always the same) taking his academic engagement seriously. It was not easy, especially at first sight. Imagine: Sokol finished his master’s studies at fifty-seven and during the subsequent seven years caught up with what takes other people decades. Each of the degrees he achieved was backed by a new book, and every one of those books has by now been re-published and some belong to the basic reading in several academic fields.
A political defeat repeated itself on a large stage in the new millennium and this time, the defeated parties were not to blame. Presidential election was won by Václav Klaus, and in case you forgot, it was by Communist votes. By then the victory over Communism was established as a ‘victorious Capitalist revolution’ and slowly but surely, a normalisation of the majority set in, the worst legacy we could have gotten from the Bolshevik and one that most people accepted with gratitude. Ugh!
For both of us, that is, for Jan and me, the school was a salvation. A school which we– that much is clear by now – by a series of incredible miracles managed to bring to life and then keep it alive. In the 1990s, it was called the Institute of Foundations of Education of the Charles University and from the beginning of the new millennium, the Faculty of Humanities of the Charles University.
School legend has it that Jan Sokol and I are the Founding Fathers. That is, as Jan used to note, if he is the father, then I am the mother. Well, the administrative operation of this institution which originally had no administration was largely up to me. To that I added some enthusiasm and a spirit of free and friendly collaboration and perhaps some less common forms of instruction. But the contents, the study of education as such, that is Jan Sokol’s work. He tried it out in lectures, then summarised it in books, and presented to grant academic titles. It was a joy and an honour for us to participate in something like that. And it meant a lot also to several thousand students and some dozens of teachers who experienced those wonderful years with us. For almost ten years, I was the boss to Jan and others, and then, again for a decade and already at the Faculty of Humanities, he was my boss. And then there was a decade when we handed the sceptre over to the younger generation. It was not always easy, it was no walk in the park, and we did not achieve everything we wanted. But by now it is clear that we left a legacy beyond our time and abilities and it is up to the younger generation to deal with it. It is pleasant to know that continuation does not require miracles, although they, too, sometimes happen. Usually, what’s needed is hard work and determination.
A good person has died, an outstanding thinker, brilliant teacher, and a reliable man. He was also a die-hard believer in Havel’s dictum that ‘love and truth shall prevail’. I am thankful for being able to pay him back the debts I owed him in the legal sense. But gratitude and awareness of debt in the spiritual sense, that will stay with me until the end of my days.
Text generation – T5 model to summarize a text. Fill in the blanks – model predicts token that should fill in the blank when any token from an example sentence is masked out. Classification and regression models – demo contains binary classification (for sentiment analysis, using SST2), multi-class classification (for textual entailment) Gender bias in coreference systems – gendered associations in a coreference system, which matches pronouns to their antecedents.
The CLARIN Annual Conference is organized for the wider Humanities and Social Sciences community in order to exchange experiences and best practices in working with the CLARIN infrastructure and to share plans for future developments. The programme will cover a range of topics, including the design, construction and operation of the CLARIN infrastructure, the data, tools and services that it contains or should contain, its actual use by researchers, teachers or interested parties, its relation to other infrastructures and projects, and the CLARIN Knowledge Sharing Infrastructure.
We need a “template of understanding”… easy to say… hard to chase by probabilistic methods.
“By systematically asking these questions about all the entities and events in a story, NLP researchers can score systems’ comprehension in a principled way, probing for the world models that systems actually need
Spatial: Where is everything located and how is it positioned throughout the story?
Temporal: What events occur and when?
Causal: How do events lead mechanistically to other events?
Motivational: Why do the characters decide to take the actions they take?
For example, Pustejovsky came with QUALIA and generative lexicons in a detailed way already 20 years ago.
James Pustejovsky, Elisabetta Jezek (2017): Generative Lexicon: Integrating Theoretical and Distributional Methods. 1. Introduction to GL and Distributional Analysis. ESSLLI 2017, 7/17/2017.