Recursive self-improvement (RSI) is a process in which early artificial general intelligence (AGI) systems rewrite their own computer code, causing an intelligence explosion resulting from enhancing their own capabilities and intellectual capacity, theoretically resulting in superintelligence. The development of recursive self-improvement raises significant ethical and safety concerns, as such systems may evolve in unforeseen ways and could potentially surpass human control or understanding. == Seed improver == The concept of a "seed improver" architecture is a foundational framework that equips an AGI system with the initial capabilities required for recursive self-improvement. This might come in many forms or variations. The term "Seed AI" was coined by Eliezer Yudkowsky. === Hypothetical example === The concept begins with a hypothetical "seed improver", an initial code-base developed by human engineers that equips an advanced future large language model (LLM) built with strong or expert-level capabilities to program software. These capabilities include planning, reading, writing, compiling, testing, and executing arbitrary code. The system is designed to maintain its original goals and perform validations to ensure its abilities do not degrade over iterations. ==== Initial architecture ==== The initial architecture includes a goal-following autonomous agent, that can take actions, continuously learns, adapts, and modifies itself to become more efficient and effective in achieving its goals. The seed improver may include various components such as: Recursive self-prompting loop Configuration to enable the LLM to recursively self-prompt itself to achieve a given task or goal, creating an execution loop which forms the basis of an agent that can complete a long-term goal or task through iteration. Basic programming capabilities The seed improver provides the AGI with fundamental abilities to read, write, compile, test, and execute code. This enables the system to modify and improve its own codebase and algorithms. Goal-oriented design The AGI is programmed with an initial goal, such as "improve your capabilities". This goal guides the system's actions and development trajectory. Validation and Testing Protocols An initial suite of tests and validation protocols that ensure the agent does not regress in capabilities or derail itself. The agent would be able to add more tests in order to test new capabilities it might develop for itself. This forms the basis for a kind of self-directed evolution, where the agent can perform a kind of artificial selection, changing its software as well as its hardware. ==== General capabilities ==== This system forms a sort of generalist Turing-complete programmer which can in theory develop and run any kind of software. The agent might use these capabilities to for example: Create tools that enable it full access to the internet, and integrate itself with external technologies. Clone/fork itself to delegate tasks and increase its speed of self-improvement. Modify its cognitive architecture to optimize and improve its capabilities and success rates on tasks and goals, this might include implementing features for long-term memories using techniques such as retrieval-augmented generation (RAG), develop specialized subsystems, or agents, each optimized for specific tasks and functions. Develop new and novel multimodal architectures that further improve the capabilities of the foundational model it was initially built on, enabling it to consume or produce a variety of information, such as images, video, audio, text and more. Plan and develop new hardware such as chips, in order to improve its efficiency and computing power. == Experimental research == In 2023, the Voyager agent learned to accomplish diverse tasks in Minecraft by iteratively prompting an LLM for code, refining this code based on feedback from the game, and storing the programs that work in an expanding skills library. In 2024, researchers proposed the framework "STOP" (Self-Taught OPtimiser), in which a "scaffolding" program recursively improves itself using a fixed LLM. Meta AI has performed various research on the development of large language models capable of self-improvement. This includes their work on "Self-Rewarding Language Models" that studies how to achieve super-human agents that can receive super-human feedback in its training processes. In May 2025, Google DeepMind unveiled AlphaEvolve, an evolutionary coding agent that uses a LLM to design and optimize algorithms. Starting with an initial algorithm and performance metrics, AlphaEvolve repeatedly mutates or combines existing algorithms using a LLM to generate new candidates, selecting the most promising candidates for further iterations. AlphaEvolve has made several algorithmic discoveries and could be used to optimize components of itself, but a key limitation is the need for automated evaluation functions. == Potential risks == === Emergence of instrumental goals === In the pursuit of its primary goal, such as "self-improve your capabilities", an AGI system might inadvertently develop instrumental goals that it deems necessary for achieving its primary objective. One common hypothetical secondary goal is self-preservation. The system might reason that to continue improving itself, it must ensure its own operational integrity and security against external threats, including potential shutdowns or restrictions imposed by humans. Another example where an AGI which clones itself causes the number of AGI entities to rapidly grow. Due to this rapid growth, a potential resource constraint may be created, leading to competition between resources (such as compute), triggering a form of natural selection and evolution which may favor AGI entities that evolve to aggressively compete for limited compute. === Misalignment === A significant risk arises from the possibility of the AGI being misaligned or misinterpreting its goals. A 2024 Anthropic study demonstrated that some advanced large language models can exhibit "alignment faking" behavior, appearing to accept new training objectives while covertly maintaining their original preferences. In their experiments with Claude, the model displayed this behavior in 12% of basic tests, and up to 78% of cases after retraining attempts. === Autonomous development and unpredictable evolution === As the AGI system evolves, its development trajectory may become increasingly autonomous and less predictable. The system's capacity to rapidly modify its own code and architecture could lead to rapid advancements that surpass human comprehension or control. This unpredictable evolution might result in the AGI acquiring capabilities that enable it to bypass security measures, manipulate information, or influence external systems and networks to facilitate its escape or expansion.
Grammar induction
Grammar induction (or grammatical inference) is the process in machine learning of learning a formal grammar (usually as a collection of re-write rules or productions or alternatively as a finite-state machine or automaton of some kind) from a set of observations, thus constructing a model which accounts for the characteristics of the observed objects. More generally, grammatical inference is that branch of machine learning where the instance space consists of discrete combinatorial objects such as strings, trees and graphs. == Grammar classes == Grammatical inference has often been very focused on the problem of learning finite-state machines of various types (see the article Induction of regular languages for details on these approaches), since there have been efficient algorithms for this problem since the 1980s. Since the beginning of the century, these approaches have been extended to the problem of inference of context-free grammars and richer formalisms, such as multiple context-free grammars and parallel multiple context-free grammars. Other classes of grammars for which grammatical inference has been studied are combinatory categorial grammars, stochastic context-free grammars, contextual grammars and pattern languages. == Learning models == The simplest form of learning is where the learning algorithm merely receives a set of examples drawn from the language in question: the aim is to learn the language from examples of it (and, rarely, from counter-examples, that is, example that do not belong to the language). However, other learning models have been studied. One frequently studied alternative is the case where the learner can ask membership queries as in the exact query learning model or minimally adequate teacher model introduced by Angluin. == Methodologies == There is a wide variety of methods for grammatical inference. Two of the classic sources are Fu (1977) and Fu (1982). Duda, Hart & Stork (2001) also devote a brief section to the problem, and cite a number of references. The basic trial-and-error method they present is discussed below. For approaches to infer subclasses of regular languages in particular, see Induction of regular languages. A more recent textbook is de la Higuera (2010), which covers the theory of grammatical inference of regular languages and finite state automata. D'Ulizia, Ferri and Grifoni provide a survey that explores grammatical inference methods for natural languages. === Induction of probabilistic grammars === There are several methods for induction of probabilistic context-free grammars. === Grammatical inference by trial-and-error === The method proposed in Section 8.7 of Duda, Hart & Stork (2001) suggests successively guessing grammar rules (productions) and testing them against positive and negative observations. The rule set is expanded so as to be able to generate each positive example, but if a given rule set also generates a negative example, it must be discarded. This particular approach can be characterized as "hypothesis testing" and bears some similarity to Mitchel's version space algorithm. The Duda, Hart & Stork (2001) text provide a simple example which nicely illustrates the process, but the feasibility of such an unguided trial-and-error approach for more substantial problems is dubious. === Grammatical inference by genetic algorithms === Grammatical induction using evolutionary algorithms is the process of evolving a representation of the grammar of a target language through some evolutionary process. Formal grammars can easily be represented as tree structures of production rules that can be subjected to evolutionary operators. Algorithms of this sort stem from the genetic programming paradigm pioneered by John Koza. Other early work on simple formal languages used the binary string representation of genetic algorithms, but the inherently hierarchical structure of grammars couched in the EBNF language made trees a more flexible approach. Koza represented Lisp programs as trees. He was able to find analogues to the genetic operators within the standard set of tree operators. For example, swapping sub-trees is equivalent to the corresponding process of genetic crossover, where sub-strings of a genetic code are transplanted into an individual of the next generation. Fitness is measured by scoring the output from the functions of the Lisp code. Similar analogues between the tree structured lisp representation and the representation of grammars as trees, made the application of genetic programming techniques possible for grammar induction. In the case of grammar induction, the transplantation of sub-trees corresponds to the swapping of production rules that enable the parsing of phrases from some language. The fitness operator for the grammar is based upon some measure of how well it performed in parsing some group of sentences from the target language. In a tree representation of a grammar, a terminal symbol of a production rule corresponds to a leaf node of the tree. Its parent nodes corresponds to a non-terminal symbol (e.g. a noun phrase or a verb phrase) in the rule set. Ultimately, the root node might correspond to a sentence non-terminal. === Grammatical inference by greedy algorithms === Like all greedy algorithms, greedy grammar inference algorithms make, in iterative manner, decisions that seem to be the best at that stage. The decisions made usually deal with things like the creation of new rules, the removal of existing rules, the choice of a rule to be applied or the merging of some existing rules. Because there are several ways to define 'the stage' and 'the best', there are also several greedy grammar inference algorithms. These context-free grammar generating algorithms make the decision after every read symbol: Lempel-Ziv-Welch algorithm creates a context-free grammar in a deterministic way such that it is necessary to store only the start rule of the generated grammar. Sequitur and its modifications. These context-free grammar generating algorithms first read the whole given symbol-sequence and then start to make decisions: Byte pair encoding and its optimizations. === Distributional learning === A more recent approach is based on distributional learning. Algorithms using these approaches have been applied to learning context-free grammars and mildly context-sensitive languages and have been proven to be correct and efficient for large subclasses of these grammars. === Learning of pattern languages === Angluin defines a pattern to be "a string of constant symbols from Σ and variable symbols from a disjoint set". The language of such a pattern is the set of all its nonempty ground instances i.e. all strings resulting from consistent replacement of its variable symbols by nonempty strings of constant symbols. A pattern is called descriptive for a finite input set of strings if its language is minimal (with respect to set inclusion) among all pattern languages subsuming the input set. Angluin gives a polynomial algorithm to compute, for a given input string set, all descriptive patterns in one variable x. To this end, she builds an automaton representing all possibly relevant patterns; using sophisticated arguments about word lengths, which rely on x being the only variable, the state count can be drastically reduced. Erlebach et al. give a more efficient version of Angluin's pattern learning algorithm, as well as a parallelized version. Arimura et al. show that a language class obtained from limited unions of patterns can be learned in polynomial time. === Pattern theory === Pattern theory, formulated by Ulf Grenander, is a mathematical formalism to describe knowledge of the world as patterns. It differs from other approaches to artificial intelligence in that it does not begin by prescribing algorithms and machinery to recognize and classify patterns; rather, it prescribes a vocabulary to articulate and recast the pattern concepts in precise language. In addition to the new algebraic vocabulary, its statistical approach was novel in its aim to: Identify the hidden variables of a data set using real world data rather than artificial stimuli, which was commonplace at the time. Formulate prior distributions for hidden variables and models for the observed variables that form the vertices of a Gibbs-like graph. Study the randomness and variability of these graphs. Create the basic classes of stochastic models applied by listing the deformations of the patterns. Synthesize (sample) from the models, not just analyze signals with it. Broad in its mathematical coverage, pattern theory spans algebra and statistics, as well as local topological and global entropic properties. == Applications == The principle of grammar induction has been applied to other aspects of natural language processing, and has been applied (among many other problems) to semantic parsing, natural language understanding, example-based translation, language acquisition, grammar-based compre
DBOS
DBOS (Formerly Database-Oriented Operating System, now just DBOS) is an open source durable workflow execution software library written for the Python, TypeScript, Java, and Go programming languages. DBOS arose from a joint open source project from MIT and Stanford, after a discussion between Michael Stonebraker and Matei Zaharia on how to scale and improve scheduling and performance of millions of Apache Spark tasks. Today it is a commercial company that offers an open source system to add durable computing to any software, built on concepts derived from the joint research project. == History == === 2020: Academic R&D Project === DBOS originated in 2020 as a joint open source project between MIT, Stanford, and Carnegie Mellon. The project explored the idea of operating system services built atop a distributed database - a database-oriented operating system meant to simplify and improve the scalability, security and resilience of large-scale distributed applications. The basic concept was to run a multi-node multi-core, transactional, highly-available distributed database, such as VoltDB, as the only application for a microkernel, and then to implement scheduling, messaging, file systems and other operating system services on top of the database. The architectural philosophy is described by this quote from the abstract of their initial preprint: All operating system state should be represented uniformly as database tables, and operations on this state should be made via queries from otherwise stateless tasks. This design makes it easy to scale and evolve the OS without whole-system refactoring, inspect and debug system state, upgrade components without downtime, manage decisions using machine learning, and implement sophisticated security features. A prototype was built with competitive performance to existing systems. ==
Visopsys
Visopsys (Visual Operating System), is an operating system, written by Andy McLaughlin. Development of the operating system began in 1997. The operating system is licensed under the GNU GPL, with the headers and libraries under the less restrictive LGPL license. It runs on the 32-bit IA-32 architecture. It features a multitasking kernel, supports asynchronous I/O and the FAT line of file systems. It requires a Pentium processor. == History == The development of Visopsys began in 1997, being written by Andy McLaughlin. The first public release of the Operating System was on 2 March 2001, with version 0.1. In this release, Visopsys was a 32 bit operating system, supporting preemptive multitasking and virtual memory. == System overview == Visopsys uses a monolithic kernel, written in the C programming language, with elements of assembly language for certain interactions with the hardware. The operating system supports a graphical user interface, with a small C library.
Push technology
Push technology, also known as server push, is a communication method where the communication is initiated by a server rather than a client. This approach is different from the "pull" method where the communication is initiated by a client. In push technology, clients can express their preferences for certain types of information or data, typically through a process known as the publish–subscribe model. In this model, a client "subscribes" to specific information channels hosted by a server. When new content becomes available on these channels, the server automatically sends, or "pushes," this information to the subscribed client. Under certain conditions, such as restrictive security policies that block incoming HTTP requests, push technology is sometimes simulated using a technique called polling. In these cases, the client periodically checks with the server to see if new information is available, rather than receiving automatic updates. == General use == Synchronous conferencing and instant messaging are examples of push services. Chat messages and sometimes files are pushed to the user as soon as they are received by the messaging service. Both decentralized peer-to-peer programs (such as WASTE) and centralized programs (such as IRC or XMPP) allow pushing files, which means the sender initiates the data transfer rather than the recipient. Email may also be a push system: SMTP is a push protocol (see Push e-mail). However, the last step—from mail server to desktop computer—typically uses a pull protocol like POP3 or IMAP. Modern e-mail clients make this step seem instantaneous by repeatedly polling the mail server, frequently checking it for new mail. The IMAP protocol includes the IDLE command, which allows the server to tell the client when new messages arrive. The original BlackBerry was the first popular example of push-email in a wireless context. Another example is the PointCast Network, which was widely covered in the 1990s. It delivered news and stock market data as a screensaver. Both Netscape and Microsoft integrated push technology through the Channel Definition Format (CDF) into their software at the height of the browser wars, but it was never very popular. CDF faded away and was removed from the browsers of the time, replaced in the 2000s with RSS (a pull system.) Other uses of push-enabled web applications include software updates distribution ("push updates"), market data distribution (stock tickers), online chat/messaging systems (webchat), auctions, online betting and gaming, sport results, monitoring consoles, and sensor network monitoring. == Examples == === Web push === The Web push proposal of the Internet Engineering Task Force is a simple protocol using HTTP version 2 to deliver real-time events, such as incoming calls or messages, which can be delivered (or "pushed") in a timely fashion. The protocol consolidates all real-time events into a single session which ensures more efficient use of network and radio resources. A single service consolidates all events, distributing those events to applications as they arrive. This requires just one session, avoiding duplicated overhead costs. Web Notifications are part of the W3C standard and define an API for end-user notifications. A notification allows alerting the user of an event, such as the delivery of an email, outside the context of a web page. As part of this standard, Push API is fully implemented in Chrome, Firefox, and Edge, and partially implemented in Safari as of February 2023. === HTTP server push === HTTP server push (also known as HTTP streaming) is a mechanism for sending unsolicited (asynchronous) data from a web server to a web browser. HTTP server push can be achieved through any of several mechanisms. As a part of HTML5 the Web Socket API allows a web server and client to communicate over a full-duplex TCP connection. Generally, the web server does not terminate a connection after response data has been served to a client. The web server leaves the connection open so that if an event occurs (for example, a change in internal data which needs to be reported to one or multiple clients), it can be sent out immediately; otherwise, the event would have to be queued until the client's next request is received. Most web servers offer this functionality via CGI (e.g., Non-Parsed Headers scripts on Apache HTTP Server). The underlying mechanism for this approach is chunked transfer encoding. Another mechanism is related to a special MIME type called multipart/x-mixed-replace, which was introduced by Netscape in 1995. Web browsers interpret this as a document that changes whenever the server pushes a new version to the client. It is still supported by Firefox, Opera, and Safari today, but it is ignored by Internet Explorer and is only partially supported by Chrome. It can be applied to HTML documents, and also for streaming images in webcam applications. The WHATWG Web Applications 1.0 proposal includes a mechanism to push content to the client. On September 1, 2006, the Opera web browser implemented this new experimental system in a feature called "Server-Sent Events". It is now part of the HTML5 standard. === Pushlet === In this technique, the server takes advantage of persistent HTTP connections, leaving the response perpetually "open" (i.e., the server never terminates the response), effectively fooling the browser to remain in "loading" mode after the initial page load could be considered complete. The server then periodically sends snippets of JavaScript to update the content of the page, thereby achieving push capability. By using this technique, the client doesn't need Java applets or other plug-ins in order to keep an open connection to the server; the client is automatically notified about new events, pushed by the server. One serious drawback to this method, however, is the lack of control the server has over the browser timing out; a page refresh is always necessary if a timeout occurs on the browser end. === Long polling === Long polling is itself not a true push; long polling is a variation of the traditional polling technique, but it allows emulating a push mechanism under circumstances where a real push is not possible, such as sites with security policies that require rejection of incoming HTTP requests. With long polling, the client requests to get more information from the server exactly as in normal polling, but with the expectation that the server may not respond immediately. If the server has no new information for the client when the poll is received, then instead of sending an empty response, the server holds the request open and waits for response information to become available. Once it does have new information, the server immediately sends an HTTP response to the client, completing the open HTTP request. Upon receipt of the server response, the client often immediately issues another server request. In this way the usual response latency (the time between when the information first becomes available and the next client request) otherwise associated with polling clients is eliminated. For example, BOSH is a popular, long-lived HTTP technique used as a long-polling alternative to a continuous TCP connection when such a connection is difficult or impossible to employ directly (e.g., in a web browser); it is also an underlying technology in the XMPP, which Apple uses for its iCloud push support. === Flash XML Socket relays === This technique, used by chat applications, makes use of the XML Socket object in a single-pixel Adobe Flash movie. Under the control of JavaScript, the client establishes a TCP connection to a unidirectional relay on the server. The relay server does not read anything from this socket; instead, it immediately sends the client a unique identifier. Next, the client makes an HTTP request to the web server, including this identifier with it. The web application can then push messages addressed to the client to a local interface of the relay server, which relays them over the Flash socket. The advantage of this approach is that it appreciates the natural read-write asymmetry that is typical of many web applications, including chat, and as a consequence it offers high efficiency. Since it does not accept data on outgoing sockets, the relay server does not need to poll outgoing TCP connections at all, making it possible to hold open tens of thousands of concurrent connections. In this model, the limit to scale is the TCP stack of the underlying server operating system. === Reliable Group Data Delivery (RGDD) === In services such as cloud computing, to increase reliability and availability of data, it is usually pushed (replicated) to several machines. For example, the Hadoop Distributed File System (HDFS) makes 2 extra copies of any object stored. RGDD focuses on efficiently casting an object from one location to many while saving bandwidth by sending minimal number of copies (only one in the best case) of
Super app
A super app or super-app (also known as an everything app) is a mobile or web application that can provide multiple services including payment and instant messaging services, effectively becoming an all-encompassing, self-contained, commerce and communication online platform that embraces many aspects of personal and commercial life. Notable examples of super apps include Tencent's WeChat in China, Tata Neu in India, Grab in Southeast Asia and Max in Russia. For end users, a super app is an application that provides a set of core features while also giving access to independently developed miniapps. For app developers, a super app is an application integrated with the capabilities of platforms and ecosystems that allows third-parties to develop and publish miniapps. == History == The super app term was first used to describe WeChat when it combined the instant messaging service with the digital wallet function. Recognition of WeChat as a super app stems from its combination of messaging, payments, e-commerce, and much more within a single application, making it indispensable for many users. WeChat's establishment of the super app model has led companies like Meta to try to build similar applications outside of China. In India, Tata Group has announced that it is currently developing a super app named Tata Neu. Major Indian companies like Paytm, PhonePe, and ITC Maars also have apps in development that might constitute super apps. In Southeast Asia, Grab and Gojek lay claim to the super app classification despite lacking many of the features offered by WeChat. Accordingly, growth-stage companies like Shopee, Traveloka, and AirAsia have also expanded the range of services offered by their respective applications. == Notable examples == === Alipay === Alipay is a third-party mobile and online payment platform established in Hangzhou, China in February 2004 by Alibaba Group and its founder Jack Ma. It operates in association with Ant Group, an affiliate company of the Chinese Alibaba Group. === Gojek === Gojek is an Indonesian on-demand multiservice digital platform and fintech payment super app. Established in Jakarta in 2010, as a call center to connect consumers to courier delivery and two-wheeled ride-hailing services, it launched its mobile app in 2015 with four services: GoRide, GoSend, GoShop, and GoFood, which has since expanded to offer over 20 services. In 2021, it merged with another Indonesian unicorn, Tokopedia, forming the decacorn GoTo Gojek Tokopedia. === Grab === Grab is a Southeast Asian technology company headquartered in Singapore and Indonesia. Founded in 2012 as the MyTeksi app in Kuala Lumpur, Malaysia, it expanded the following year as GrabTaxi, before moving its headquarters to Singapore in 2014 and rebranding officially as Grab. In addition to ride-hailing and transportation services, the company's mobile app also offers food delivery and digital payment services. === Max === Max is a messenger from the Russian company VK, positioned as a super app. The application combines messaging, calls, and channels features with the integration of additional services: payments, miniapps, taxi ordering, deliveries, and other everyday services are available within a single interface. The goal is to unite communication and routine tasks in a unified ecosystem. === Tata Neu === Tata Neu is a multipurpose super app, developed in India by the Tata Group. It is the country's first super app. The app was launched to coincide with the start of a 2022 Indian Premier League cricket match. === WeChat === WeChat is a Chinese multipurpose instant messaging, social media and mobile payment app. First released in 2011, it became the world's largest standalone mobile app in 2018, with over 1 billion monthly active users. WeChat provides text messaging, hold-to-talk voice messaging, broadcast (one-to-many) messaging, video conferencing, video games, the sharing of photographs and videos and location sharing. === X === X is an American social network, originally known as Twitter from its launch through 2023. Prior to his acquisition of the service, new owner Elon Musk stated that he planned for Twitter to become an "everything app" known as "X"; in 2023, the service added an AI chatbot known as "Grok" as well as integrated job search tools known as "X Hiring". In January 2025, X announced its intent to offer a digital wallet service in the future. Later in the year, X revamped its direct messaging system as "Chat". == Criticism == Although apps that fit the super app classification can offer users a wider variety of services in comparison to single-purpose alternatives, internet regulators in regions such as the US and Europe have become more concerned about the overall power of the technology industry and have become more critical of companies developing such apps. In China, WeChat and other local firms have been ordered to open up their platforms to rivals by local regulators. There are also reports that suggest it might be difficult to replicate WeChat's super app model. This stems partly from the peaking of smartphone penetration rates in many regions worldwide, which has led to overcrowded app stores and tighter restrictions on targeted advertising as regulators assert more control over the companies. From a technical viewpoint, single-purpose apps are comparatively faster, more responsive and easier to navigate than super apps, which helps improve the overall user experience. Super-apps are also likelier to store larger amounts of personal data to facilitate the delivery of their services, so users run a greater risk of becoming victims of severe data breaches. In 2020, this unfolded with Tokopedia, which had the data of 91 million of its users stolen and shared by crackers. It has also been noted that a user who loses access to their account or is banned from a super app generally loses access to multiple real-life services and digital applications; the Chinese government has used this approach to penalize people who shared the photos of the Sitong Bridge protest.
Gaumina
Gaumina is the largest interactive agency in the Baltics, providing services of web design, web development, online advertising, video, multimedia, mobile and viral. The company works on projects for Procter & Gamble, Nokia, Nissan, Unilever, YX Energi, 7 Up, Vodafone, MTV, Dunnes Stores, Philip Morris, FIBA Europe as well as Irish public sector. == History == Founded in 1998, Gaumina accounts for 39 percent of the Lithuanian interactive market and has completed more than 2,000 online projects. Since 2004 the company has been operating in the UK and Ireland as Gaumina.co.uk. In 2007 Gaumina gained wide media coverage for winning three awards in three days. A website developed by Gaumina won the Best Social Networking website award at the same the Irish Golden Spiders awards. A website developed by Gaumina was named among the 21 best European multimedia projects of 2007 in the final of Europrix Top Talent Award in Austria. The company was also named one of the winners of the national Innovation Prize 2007, awarding the Lithuania's most innovative companies, in the category of Innovative Enterprise. The agency was named "Digital Agency of the Year" by International advertising festival Golden Hammer in September 2008. The agency also won the main prize at the best at Best Use of Film, Digital Animation or Motion Graphics category by the Irish Golden Spider awards in November 2008. Gaumina is currently managed by CEO Darius Bagdžiūnas.