#deepseek

#Qwen 'QWQ 32B,' a reasoning model you can run locally via #Ollama on your PC with sufficient GPU RAM (20GB+), delivers performance nearly comparable to #DeepSeek R1, despite the latter having 671 billion parameters. The power of efficient architecture! #AI

Да, #OpenAI и #DeepSeek узнают много данных обо мне, которые лучше бы им не знать (это не конфиденциальные данные, публично доступные, но всё-таки если был бы выбор, я бы и их не передавал).

Но иначе я вообще не успею вовремя. Я за полтора часа 10 страниц написал, ещё 25 страниц мне нужно написать за ещё полтора часа. Разумеется, я не успею выполнить такой объём за такое время без #AI

Да, у меня не было другого времени для написания этой работы. И да, её содержание даже не будут читать.

Gemini 2.5 Pro еще ближе к человеку, но реально рассуждать умеют только модели OpenAI

25 марта вышла новая модель Gemini 2.5 Pro от Google. Много уже было про нее сказано, но я стараюсь не обращать внимания на волны хайпа вокруг новинок, ориентируясь на свои доверенные источники. Одним из них является любительский бенчмарк SimpleBench , который тестирует здравый смысл в реальной жизни. Он интересен тем, что в нем самый обычный человек набирает 83.7%, а рекорд для LLM еще недавно был лишь 46.4%.

https://habr.com/ru/articles/896306/

#Gemini #o1 #reasoning #claude #deepseek #google #25 #pro #искусственный_интеллект #нейросети

Wow! Check out this awesome video from lucky giveaway winner!
Two clear methods to run #DeepSeek locally on the ArmSoM #Sige7 powered by #RK3588.
Both informative and fun for the video, offering something new for everyone.
#opensource #armbian #rockchip #linux #ai #sbc #hardware #iot #embedded

https://www.youtube.com/watch?v=OsIT6F8DFs0

#FinancialTimes Alibaba seeks to become an AI leader after years of stagnant growth | FT #shorts

Alibaba seeks to become an AI leader after years of stagnant growth | FT #shorts The reversal of fortunes in the Chinese tech giant marks a remarkable comeback for a company that had embodied the rise of China’s tech industry in the first two decades of this century. #alibaba #ai #tech #jackma #deepseek #artificialintelligence #china #shortsvideo #shortfeed #shorts…

https://www.shibuyaworldnews.online/financialtimes-alibaba-seeks-to-become-an-ai-leader-after-years-of-stagnant-growth-ft-shorts/

‘Giving machines a heart’: China urged to dig deeper in AI development https://www.byteseu.com/882565/ #agi #AI #AIPhilosophy #AITheory #ArtificialGeneralIntelligence #ArtificialIntelligence #Beijing #ChatGPT #CognitiveModelling #DeepSeek #Innovation #PekingUniversity #PhilosophicalLayer #TECH #TheoreticalLayer #Zhongguancun #ZhongguancunForum #ZhuSongchun

People have low expectations for AI. I asked ChatGPT (paid account, not the free stuff) what would make ChatGPT better. I think the answer is a lot higher quality than what people keep saying one should expect. This is what I got:

--- My Prompt:
With DeepSeek, there are things it is very good at, but it fails quickly and rather spectacularly. Your models tend to break down as well, but somewhere there must be a model that will be able to stay on topic, even when the topic changes, so that conversations move around like human conversations move around subject-to-subject. What would take ChatGPT over the top?

--- ChatGPT reply:
Here’s what it’ll take to get closer to that human-like conversation flow:

Long-Term Memory That Actually Works. Most models either have no memory or they remember only in a clunky, artificial way.

A better memory system would not just recall facts, but contextual relevance—like “we talked about LLM efficiency two days ago and you hate cheerleader language.”

It has to be smart recall, not just data storage.

Thread Awareness
Imagine a model that can say: “You mentioned DeepSeek earlier. Should I compare it again now that we’ve brought up open-source tools?”
That requires models to track not just words, but conversational branches—like a tree, not a straight line.
Graceful Fallbacks
When a model doesn’t know, it should gracefully pivot, not hallucinate or go off-topic. This kind of conversational “stability” is rare and subtle.

Most models that fail “spectacularly,” as you said, are trying to bluff through complexity instead of narrowing scope or clarifying.

Hybrid Contextual ReasoningThe future model needs to reason across subjects rather than switching on and off.

If we're talking about language models, and you shift to science fiction, it should be able to bring up Asimov or neural networks in storytelling—bridging topics, not just hopping between them.

Right now, most models either:

Stay sharp but shallow (good at Q&A, bad at memory), or

Drift into friendly nonsense (chatty, but brittle)

The sweet spot will be a system that grows with you across time and topics—and doesn’t forget what you value in a conversation.

....

When you see stuff posted online meant to minimize the quality of output possible, This was a single question with no leader. I think this is the kind of answer we want from our AI devices.

Thoughts? Feel free to tear this response apart. It needs that kind of human analysis.

#ai #openai #deepseek #intelligence #chatgpt #claude #gemini #perplexity #copilot #jasper

SPIRIT's Diary: How Easy to Use Deepseek Is? https://tkt058.blogspot.com/2025/02/how-easy-to-use-deepseek-is.html?spref=tw
#DeepSeekR1 #DeepSeek #DeepSafe #AI #AIChat #AIchatbot #China #Chinatech #ArtificialIntelligence #AI #Blog #Blogs #Blogger #Bloggers #Informaton

DeepSeek surpasses ChatGPT in new monthly visits
https://economictimes.indiatimes.com/tech/artificial-intelligence/deepseek-surpasses-chatgpt-in-new-monthly-visits-emerges-as-the-fastest-growing-ai-tool-report/articleshow/119754529.cms?from=mdr
#ycombinator #DeepSeek_monthly_visits #ChatGPT_competitor #AI_analytics_platform #China_AI_startup #DeepSeek #OpenAI #fastest_growing_AI_tool

[Перевод] Четыре основных подхода к созданию моделей рассуждений (Reasoning LLMs)

Описаны ключевых методологии создания Reasoning LLM: масштабирование во время инференса, чистое обучение с подкреплением, комбинация SFT + RL и дистилляция. Это практическое руководство поможет выбрать оптимальную стратегию разработки, опираясь на имеющиеся ресурсы.

https://habr.com/ru/articles/894688/

#llm #reasoning #deepseek #agentic_ai #ии #искусственный_интеллект #gpt #разработка #ииинжиниринг

How DeepSeek erased Silicon Valley’s AI lead and wiped $1 trillion from U.S. markets https://www.byteseu.com/879308/ #A.I. #AI #AlibabaGroup #ArtificialIntelligence #China #DeepSeek #Editor'sPicks #Featured #OpenAI

Updated #PlanB using #AI #DeepSeek

They are not even remotely hiding their base #deepseek

国家支配から抜け出せないDeepSeek、イーロン支配から抜け出せないGrok、マイクロソフト支配から抜け出せないChatGPT。OSSによるAIは成立するでしょうか。
#DeepSeek #Grok #ChatGPT #OSS #AI

Chiedo a #DeepSeek se mi aggiusta un testo in inglese sull'Intifada Studentesca e alla fine lo ringrazio. Questa la risposta:
(L'intelligenza artificiale orientale è molto meglio della stupidità umana occidentale)
#AI

DeepSeek is revolutionizing AI, competing with OpenAI’s ChatGPT, Google’s Gemini & xAI’s Grok!
Its DeepSeek LLM (67B parameters) outperforms LLaMA-2 70B in math & coding!
DeepSeek Coder, trained on 2T tokens, excels in multi-language coding!
In March 2025, DeepSeek-V3-0324 launched, boosting reasoning & coding!
With a research-first, open-source approach, can DeepSeek lead global AI? #AI #DeepSeek

https://newwstuff.com/meet-deepseek-the-chinese-ai-company-taking-on-the-world/

#AbrahamLincoln: "#democracy is #government of the #people - by the #people - for the #people"

but #democracy is not perfect per default, it needs constant #reform to become better and better (not worse and worse how it was made by (#financial #taxavoiding #lobbyists and other #lobbyists = #corruption)

how to reform democracy? a #conversationwithai: #deepseek

<think>
Alright, I'm trying to figure out how to reform #democracy so that decisions are made faster and are more competent, while also reducing incompetence and corruption. This is a big question because democracy as it stands has a lot of #issues, but it's still considered one of the better systems out there <- read this?

First off, I know that one of the main problems with democracies is that #decisionmaking can be #slow, especially in #large #countries.

There are so many #checks and #balances that sometimes it feels like nothing gets done.

So maybe making decisions faster requires #streamlining some processes.

But how? Maybe reducing #bureaucracy or changing how legislation is passed?

Another issue is #incompetence. I've seen elected officials who don't seem to have the right skills or knowledge for their roles. <- this #AI has seen things!

How can we ensure that #leaders are actually #competent?

Perhaps through better #education #requirements, or maybe having #experts in certain fields advise them more #effectively.

#Corruption is another huge #Problem seems like #money plays a big role in #politics, which can lead to decisions that favor #donors over the people.

#Campaign #finance #reform might help here, but I'm not sure how feasible that is given current laws and court rulings, especially in places like the #US with #CitizensUnited.

#Voter #engagement is also a factor.

If more people participate, maybe there's more #accountability.

But getting people to vote when they feel their votes don't matter is tough.

Maybe changing voting systems, like ranked-choice voting, could make elections feel more representative and increase turnout.

I've heard about #sortition, where #random #citizens are #selected for #decisionmaking bodies.

That might bring in diverse perspectives and reduce corruption since the average person isn't as beholden to special interests.

But would people take it seriously? I'm not sure how effective that would be compared to elected officials.

#Technology could play a role too.

Maybe using #online #platforms for #participatory #budgeting or #policy #discussions.

It could make #government more #transparent and involve citizens more #directly.

However, there's the risk of #misinformation and #manipulation online.

#Term #limits are another idea.

They might prevent politicians from becoming too entrenched and corrupt, but they can also limit institutional knowledge and lead to a #revolvingdoor of #inexperienced #leaders.

#Anticorruption agencies seem necessary, but they need to be #independent and well #funded.

If they're controlled by the government, they won't be effective.

How do you ensure their independence?

#Media #literacy is important because if people are #better #informed, they can make #better #decisions at the #ballotbox.

But with so much #misinformation out there, improving media literacy is a big challenge.

I'm also thinking about proportional representation.

Maybe switching to a system where more parties have seats could lead to more diverse and representative governments, reducing gridlock.

But it might also make forming a government harder if no single party has a majority.

What about making voting mandatory?

It increases turnout but can also lead to people voting without being informed, which doesn't necessarily improve the quality of decisions.

Another thought is focusing on #local #governance.

If more decisions are made at the local level where people can be more involved and informed, it might lead to #better #outcomes #overall.

But not all issues can be handled locally; some require national or international coordination.

I'm also considering how #education plays into this.

Better #civic #education could make voters more informed and engaged.

It starts from a young age teaching about #government #structures, #roles, and the importance of #participation.

#Transparency is key too.

If governments are more open about their processes and decisions, trust might increase. Things like #FOIA requests in the U.S. help, but they can be slow and sometimes
incomplete.

#Incentivizing #good #governance through awards or recognition might motivate leaders to perform better, but it's not a sure fix for deeper systemic issues.

Encouraging #whistleblowers is important because they can expose #corruption before it becomes entrenched.

Protecting them legally would make others more willing to come forward.

Public-private partnerships could bring in expertise from the private sector, but there's a risk of corporate influence if not managed properly.

Inclusive decision-making ensures that all groups have a voice, which leads to more equitable policies.

It requires active efforts to involve marginalized communities.

Independent judiciaries are crucial because they can check the other branches and enforce laws against corruption.

But judicial appointments can be politicized, which undermines their independence.

Finally, cultural shifts towards valuing integrity over power might help reduce corruption. But changing a culture is slow and difficult, especially in large, diverse societies.

Putting all these thoughts together, it seems like reforming democracy requires a multi-faceted approach addressing both structural issues and cultural factors. It's not just about tweaking one part but making comprehensive changes that promote competence, transparency, and accountability while ensuring representation and reducing corruption.

</think>
Reforming democracy to enhance decision-making efficiency, competence, and reduce corruption involves a multifaceted approach that addresses both structural and cultural aspects of
governance. Here is an organized strategy based on the thought process:

### Structural Reforms
1. **Streamline Decision-Making Processes:**
- Reduce bureaucratic layers and adopt more agile legislative processes to accelerate decision-making without compromising thorough consideration.

2. **Campaign Finance Reform:**
- Implement strict regulations on campaign contributions and promote public financing of elections to diminish the influence of money in politics.

3. **Proportional Representation:**
- Introduce proportional representation systems to ensure diverse representation, which can lead to more inclusive and efficient governance.

4. **Term Limits:**
- Enact term limits for elected officials to prevent entrenchment, while ensuring mechanisms for knowledge retention and smooth transitions.

5. **Independent Anti-Corruption Agencies:**
- Establish independent bodies with strong investigative and prosecutorial powers to combat corruption effectively.

### Enhancing Competence
1. **Expert Advisory Bodies:**
- Integrate expert panels in decision-making processes to provide informed guidance on complex issues.

2. **Civic Education:**
- Improve education systems to foster informed voters through civic engagement and critical thinking skills.

3. **Sortition (Citizen Juries):**
- Experiment with sortition to involve diverse citizen perspectives, enhancing
representativeness and reducing corruption risks.

### Reducing Corruption
1. **Transparency Measures:**
- Increase government transparency through open data initiatives, accessible meetings, and robust freedom of information laws.

2. **Whistleblower Protection:**
- Strengthen legal protections for whistleblowers to encourage the exposure of corruption without fear of retribution.

3. **Judicial Independence:**
- Ensure the judiciary is independent and impartial, free from political influence, to uphold rule of law and check executive power.

### Voter Engagement and Participation
1. **Voting System Reforms:**
- Adopt ranked-choice voting and proportional representation to ensure votes count more meaningfully and increase voter satisfaction.

2. **Compulsory Voting:**
- Consider mandatory voting with incentives for informed voting, potentially reducing apathy and increasing political engagement.

3. **Participatory Budgeting:**
- Involve citizens in budget decisions through digital platforms to foster transparency and accountability.

### Technological Integration
1. **Digital Platforms:**
- Utilize technology for participatory governance, enabling real-time public input on policies and fostering informed discussions.

2. **Media Literacy Programs:**
- Launch initiatives to combat misinformation, promoting a better-informed electorate capable of discerning credible sources.

### Cultural Shifts
1. **Incentivizing Good Governance:**
- Recognize and reward ethical leadership through awards or public acknowledgment to promote integrity in governance.

2. **Cultural Values:**
- Foster societal norms that prioritize transparency, accountability, and integrity, encouraging leaders to adhere to high ethical standards.

### Conclusion
Reforming democracy requires a comprehensive approach that combines structural changes with cultural shifts. By streamlining processes, enhancing
competence through education and expertise, reducing corruption through transparency and independent oversight, and engaging citizens more
effectively, democracies can become more responsive, efficient, and just. This multifaceted strategy aims to rebuild trust in governance and ensure
decisions truly reflect the will of the people.

Топ-9 бесплатных китайских нейросетей в 2025 году

Tencent, Alibaba, DeepSeek AI, Kuaishou — эти компании выпускают модели одну за другой, заставляя западных конкурентов почесать головы. Китайские ИИ‑продукты сегодня покрывают всё: от создания объёмных научных статей до генерации мемных видеороликов, которые легко взорвут соцсети. Эти технологии уже привлекли внимание профессионалов, но их потенциал доступен и обычным пользователям. Куда бежать, если вам срочно нужно оживить персонажа, создать фотосессию из одного‑единственного портрета или сделать прикольную гифку для поздравления? Hunyuan‑T1 с гибридным трансформером, Qwen-2.5 с миллионным контекстным окном, Talkie, который обошёл по популярности Character.ai, и Kling, монтирующий видео из нескольких картинок, — обо всём этом в статье!

https://habr.com/ru/companies/bothub/articles/895482/

#ии #hunyuan #deepseek #qwen #hailuo #stepvideot2v #pixverse #kling #talkie #seaart

Closing down for the night. Loving the #flatline of #TSLA Thought I'd ask #DeepSeek about #WallStreetGurus Looks like there is lots of #Entertainment and not too much substance--except Longterm Investing. Gnite. #ThisIsHowIUseAI

Gosh darn it to heck #ai #chatgpt #llm #deepseek