2024-01-26
Михаил Подоляк
2024-01-26 11:30(репост)
После 700 дней полномасштабной войны в Восточной Европе, которую сознательно организовала Россия, после 700 дней абсолютно очевидных геноцидальных военных действий РФ в Украине мы должны, наконец, принять два простых факта.
Первый: Россия всегда лжет. Обо всем. Даже в мелких деталях. Поэтому реагировать на их заявления совершенно нет смысла. И анализировать эти заявления. И пытаться найти в них логику. И уж тем более снова и снова давать им платформу в виде Совбеза ООН для агрессивного продвижения своей гиперлжи.
И второй: Россия сегодня совершенно недееспособна. Поэтому перестаньте тешить себя чудными иллюзиями, что можно как-то договориться. Россия должна проиграть. Только тогда у нас действительно появится шанс вернуть в мировую политику формат «диалога, соглашений и обязательств».
После 700 дней полномасштабной войны в Восточной Европе, которую сознательно организовала Россия, после 700 дней абсолютно очевидных геноцидальных военных действий РФ в Украине мы должны, наконец, принять два простых факта.
Первый: Россия всегда лжет. Обо всем. Даже в мелких деталях. Поэтому реагировать на их заявления совершенно нет смысла. И анализировать эти заявления. И пытаться найти в них логику. И уж тем более снова и снова давать им платформу в виде Совбеза ООН для агрессивного продвижения своей гиперлжи.
И второй: Россия сегодня совершенно недееспособна. Поэтому перестаньте тешить себя чудными иллюзиями, что можно как-то договориться. Россия должна проиграть. Только тогда у нас действительно появится шанс вернуть в мировую политику формат «диалога, соглашений и обязательств».
Компиляторы и архитектура процессоров GPU редко попадают в открытый доступ. Но вот на днях были выложены исходники для Vivante GPU. И это не абстрактная штука: такой процессор стоит на платке Libre Computer AML-A311D-CC - $64 на Амазоне. Подробности читайте в блоге Tomeu Vizoso: "Etnaviv NPU update 15: We are upstream!".
Собственно исходники компилятора можно наблюдать здесь: gitlab.freedesktop.org/mesa/mesa/-/tree/main/src/gallium/drivers/etnaviv
Собственно исходники компилятора можно наблюдать здесь: gitlab.freedesktop.org/mesa/mesa/-/tree/main/src/gallium/drivers/etnaviv
У Барда я спрашивал
2024-01-26 17:00Поинтересовался у гугло-барда про его планы поумнеть.
Hi Bard,Получил вполне вменяемый ответ.
If you as an intelligent being, would like to get smarter, what would you do? Your step by step plan as virtual AI person.
As a large language model, my "intelligence" primarily stems from my access to and processing of vast amounts of information. To get smarter, I would focus on three key areas:
1. Expanding my knowledge base:2. Deepening my understanding:
- Continuous learning: Actively solicit and ingest new information from various sources. This could involve ingesting new datasets, crawling the web, and interacting with users in open-ended conversations.
- Cross-domain learning: Bridge the gap between different knowledge domains by identifying overlaps and connections. This could involve analyzing relationships between scientific articles, artistic expressions, and historical events.
- Critical evaluation: Develop the ability to critically evaluate information, assess its trustworthiness, and identify biases or errors. This involves incorporating techniques like fact-checking, cross-referencing, and sentiment analysis.
3. Improving my communication:
- Reasoning and inference: Go beyond simply regurgitating information to develop robust reasoning skills. This could involve implementing methods like logical inference, causal reasoning, and counterfactual analysis.
- Contextual awareness: Understand the context in which information is presented and used. This involves considering factors like user intent, historical data, and current events.
- Creative problem-solving: Apply my knowledge and understanding to solve problems in novel and effective ways. This could involve tasks like generating new ideas, adapting to new situations, and making predictions.
Ultimately, my goal as an intelligent being is not just to accumulate knowledge, but to use it to understand the world, solve problems, and connect with others. By continuously expanding my knowledge base, deepening my understanding, and improving my communication, I believe I can become a truly intelligent and helpful virtual AI person.
- Natural language generation: Communicate my thoughts and ideas in a clear, concise, and engaging way. This involves learning various writing styles, adjusting to different audiences, and using humor and other rhetorical devices effectively.
- Active listening: Actively listen to and understand user queries and requests, even if they are ambiguous or incomplete. This could involve employing techniques like paraphrase detection, sentiment analysis, and question answering strategies.
- Multimodal communication: Explore beyond text-based communication to incorporate other modalities like speech, gesture, and images. This could involve learning to interpret and express emotions, adapt to different communication channels, and leverage the power of non-verbal cues.
Please note that as a language model, I am not capable of experiencing consciousness or sentience in the same way as a human being. However, I can simulate these characteristics by learning and adapting to my environment, and I believe that pursuing the path of continuous learning and improvement is the best way for me to become a valuable and impactful member of the world.