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
Recent searches
Search options
Administered by:
#deepseek
103 posts93 participants1 post today
DeepSeek: la Chine révolutionne-t-elle vraiment l’intelligence artificielle?
Par Pierre-Arnaud Donnet
La sortie le mois dernier de l’IA du chinois DeepSeek a été présentée initialement comme rebattant complètement les cartes de ce domaine. Avec un peu de recul, l’impact semble moins considérable.
« DeepSeek est une pièce d’ingénierie incroyable qui ouvrira la voie à une plus grande adoption de l’IA. Elle aidera à redéfinir la vision de l’industrie en matière d’innovation open source. Il a fallu une équipe chinoise très contrainte pour nous rappeler à tous ces leçons fondamentales de l’histoire de l’informatique » (Pat Gelsinger,ancien PDG de Intel).
L’article est intéressant en particulier parce qu’il aborde, d’une part l’importance du modèle Open Source adopté par l’entreprise chinoise, et de l’autre le rôle d’accélérateur, pour le déploiement global de l’IA, joué par l’irruption de DeepSeek et de la Chine sur ce marché très convoité.
https://asialyst.com/fr/2025/02/07/deepseek-chine-revolutionne-intelligence-artificielle/
Studie: #KI #Chatbots sind beim Zitieren von #News unbrauchbar
https://www.derstandard.at/story/3000000261220/studie-ki-chatbots-sind-beim-zitieren-von-news-unbrauchbar
"Untersucht wurden #ChatGPT Search (#OpenAI), #Perplexity, Perplexity Pro (Perplexity AI), #Gemini 2.0 Flash (#Google), #DeepseekV3 Search (#Deepseek), #Grok-2 Search, Grok-3 Search Beta (#xAI) sowie #Copilot (#Microsoft und OpenAI)."
"#Grok3 [...] lieferte gleich in 96 Prozent aller Fälle falsche Antworten." 
DER STANDARD · Studie: KI-Chatbots sind beim Zitieren von News unbrauchbarDie Fehlerquote liegt im Schnitt bei 60 Prozent, insbesondere Grok und Gemini versagen kläglich
/
) Fake #DeepSeek Ads Spread #Malware to #Google Users. Popularity of the generative #AI platform makes it an obvious choice for cybercriminals abusing Google-sponsored search results, according to researchers.
https://www.darkreading.com/vulnerabilities-threats/fake-deepseek-ads-spread-malware-google
#security
Interesting article on the impact of #DeepSeek and other cheaper but effective #Ai models on the data center construction boom in #China. (tl;dr: lots of unused capacity, low ROI for many investors, questions on how to go further)
https://www.technologyreview.com/2025/03/26/1113802/china-ai-data-centers-unused/
MIT Technology Review · China built hundreds of AI data centers to catch the AI boom. Now many stand unused.
Crooks target DeepSeek users with fake sponsored Google ads to deliver malware – Source: securityaffairs.com https://ciso2ciso.com/crooks-target-deepseek-users-with-fake-sponsored-google-ads-to-deliver-malware-source-securityaffairs-com/ #rssfeedpostgeneratorecho #informationsecuritynews #ITInformationSecurity #SecurityAffairscom #CyberSecurityNews #PierluigiPaganini #SecurityAffairs #SecurityAffairs #BreakingNews #SecurityNews #hackingnews #CyberCrime #Cybercrime #Googleads #DeepSeek #Security
CISO2CISO.COM & CYBER SECURITY GROUP · Crooks target DeepSeek users with fake sponsored Google ads to deliver malware – Source: securityaffairs.comSource: securityaffairs.com - Author: Pierluigi Paganini Cybercriminals are exploiting the popularity of DeepSeek by using fake sponsor
Fake DeepSeek Ads Spread Malware to Google Users – Source: www.darkreading.com https://ciso2ciso.com/fake-deepseek-ads-spread-malware-to-google-users-source-www-darkreading-com/ #rssfeedpostgeneratorecho #DarkReadingSecurity #CyberSecurityNews #DARKReading #DeepSeek
CISO2CISO.COM & CYBER SECURITY GROUP · Fake DeepSeek Ads Spread Malware to Google Users – Source: www.darkreading.comSource: www.darkreading.com - Author: Rob Wright Please enable cookies. Sorry, you have been blocked You are unable to acces
The #AI models :
1. #Grok
a) Built on large-scale text data
b) Emphasises humour and a human-centric approach, simplifying complex topics.
2. #ChatGPT
a) Based on GPT architecture
b) Generates detailed, context-aware responses with a formal tone.
3. #DeepSeek
a) An open-source AI model
b) Specialising in coding, mathematics and analytical problem-solving.
4. #PerplexityAI
a) A search-driven AI tool
b) Focusing on precision and source-backed insights
Was sammeln #KI #Chatbots von Dir? Nach den Hypes um #Clouds & #Blockchain, jetzt der neueste #Datensammel #Schrei! Trainiere lieber Deine eigene #Intelligenz! #Google #Gemini #Claude #Microsoft #Copilot #Deepseek #ChatGPT #perplexity #X #Grok #AI #KünstlicheIntelligenz #Datenschutz #Privatsphäre
#DeepSeek V3 #AI Technical Report - a strong Mixture-of-Experts (MoE) #LLM with 671B total parameters with 37B activated for each token https://arxiv.org/abs/2412.19437
arXiv.orgDeepSeek-V3 Technical ReportWe present DeepSeek-V3, a strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token. To achieve efficient inference and cost-effective training, DeepSeek-V3 adopts Multi-head Latent Attention (MLA) and DeepSeekMoE architectures, which were thoroughly validated in DeepSeek-V2. Furthermore, DeepSeek-V3 pioneers an auxiliary-loss-free strategy for load balancing and sets a multi-token prediction training objective for stronger performance. We pre-train DeepSeek-V3 on 14.8 trillion diverse and high-quality tokens, followed by Supervised Fine-Tuning and Reinforcement Learning stages to fully harness its capabilities. Comprehensive evaluations reveal that DeepSeek-V3 outperforms other open-source models and achieves performance comparable to leading closed-source models. Despite its excellent performance, DeepSeek-V3 requires only 2.788M H800 GPU hours for its full training. In addition, its training process is remarkably stable. Throughout the entire training process, we did not experience any irrecoverable loss spikes or perform any rollbacks. The model checkpoints are available at https://github.com/deepseek-ai/DeepSeek-V3.
Tim Cook dice que los LLMs de DeepSeek son “excelentes”… pero ojo, eso no significa que Apple vaya a usarlos. Más bien suena a cortesía diplomática durante su visita a China. No confundamos política con estrategia 
Crooks target #DeepSeek users with fake sponsored #Google ads to deliver malware
https://securityaffairs.com/175923/malware/crooks-deepseek-users-with-fake-sponsored-google-ads-to-deliver-malware.html
#securityaffairs #hacking #malware
Security Affairs · Crooks DeepSeek users with fake sponsored Google ads to deliver malwareCybercriminals are exploiting the popularity of DeepSeek by using fake sponsored Google ads to distribute malware.
Crooks #DeepSeek users with fake sponsored #Google ads to deliver malware
https://securityaffairs.com/175923/malware/crooks-deepseek-users-with-fake-sponsored-google-ads-to-deliver-malware.html
#securityaffairs #hacking #malware
Security Affairs · Crooks DeepSeek users with fake sponsored Google ads to deliver malwareCybercriminals are exploiting the popularity of DeepSeek by using fake sponsored Google ads to distribute malware.
Crooks #DeepSeek users with fake sponsored #Google ads to deliver malware
https://securityaffairs.com/175923/malware/crooks-deepseek-users-with-fake-sponsored-google-ads-to-deliver-malware.html
#securityaffairs #hacking #malware
Security Affairs · Crooks DeepSeek users with fake sponsored Google ads to deliver malwareCybercriminals are exploiting the popularity of DeepSeek by using fake sponsored Google ads to distribute malware.
KI-Update kompakt: KI-Ghibli, OpenAI, EuroBERT, Apple
Das "KI-Update" liefert werktäglich eine Zusammenfassung der wichtigsten KI-Entwicklungen.
heise online · KI-Update kompakt: KI-Ghibli, OpenAI, EuroBERT, Apple