DeepSeek R1 AI Model Update Boosts Reasoning, Catching up With OpenAI o3 and Gemini 2.5 Pro

#AI #DeepSeek #GenAI #LLM #DeepSeekR1 #AIUpdate #OpenSourceAI #ReasoningModels #AIBenchmarks #MachineLearning #ChinaAI #China

https://winbuzzer.com/2025/05/29/deepseek-r1-ai-model-update-boosts-reasoning-catching-up-with-openai-o3-and-gemini-2-5-pro-xcxwbn/

DeepSeek quietly drops R1 model upgrade

The Chinese AI firm has released an improved version of its powerful R1 reasoning model on Hugging Face without formal announcement. The new model boosts logical reasoning, efficiency, and supports real-time decision-making ranking just behind OpenAI's top models.

#AI #DeepSeekR1 #ArtificialIntelligence #AGI #HuggingFace #TechInnovation #OpenSourceAI #TECHi

Read Full Article Here :- https://www.techi.com/deepseek-r1-update-ai-reasoning-model-enhancements/

Can your 8GB laptop handle DeepSeek R1?
We ran 250 sessions, built XGBoost models (R² = 0.91 ), and found the hidden levers behind RAM, latency & reasoning accuracy.
This isn't guesswork—it's LLM deployment as data science

Read the full breakdown:
https://medium.com/@rogt.x1997/can-you-run-deepseek-r1-on-8gb-ram-a-data-science-driven-breakdown-21340677a063
#LLM #EdgeAI #DeepSeekR1 #AIForecasting #MachineLearning #LocalInference
https://medium.com/@rogt.x1997/can-you-run-deepseek-r1-on-8gb-ram-a-data-science-driven-breakdown-21340677a063

I've tried #DeepseekR1, #Llama32, #Qwen3, and of course the end-case which I knew would work, #ChatGPT. (2/6)

#Apple #MacStudio #M3Ultra Runs #DeepSeekR1 With 671 Billion Parameters Using 448GB Of Unified Memory, Delivering High Bandwidth Performance At Under 200W Power Consumption, With No Need For A Multi-GPU Setup
https://wccftech.com/m3-ultra-chip-handles-deepseek-r1-model-with-671-billion-parameters/

ARC-AGI-2 mette in crisi i modelli IA più avanzati

https://gomoot.com/arc-agi-2-mette-in-crisi-i-modelli-ia-piu-avanzati/

NEW: Researchers use AI jailbreak on top LLMs, including ChatGPT, DeepSeek, and Copilot, to create functional Google Chrome infostealers.

Read: https://hackread.com/ai-jailbreak-on-top-llms-to-create-chrome-infostealer/

NEW: Researchers manipulate AI Chatbot #DeepSeekR1 to create ransomware code and a functioning keylogger.

Read: https://hackread.com/ai-chatbot-deepseek-r1-manipulated-to-create-malware/

Meet #DeepSeekR1 - an #opensource #LLM fine-tuned with reinforcement learning to improve reasoning capability!

DeepSeek-R1 achieves results on par with OpenAI's o1 model on several benchmarks, including MATH-500 and SWE-bench.

Explore more: https://www.infoq.com/news/2025/02/deepseek-r1-release/?utm_source=mastodon&utm_medium=link&utm_campaign=calendar

"On Thursday, mobile security company NowSecure reported that the app sends sensitive data over unencrypted channels, making the data readable to anyone who can monitor the traffic. More sophisticated attackers could also tamper with the data while it's in transit.
(...)
What’s more, the data is sent to servers that are controlled by ByteDance, the Chinese company that owns TikTok. While some of that data is properly encrypted using transport layer security, once it's decrypted on the ByteDance-controlled servers, it can be cross-referenced with user data collected elsewhere to identify specific users and potentially track queries and other usage.
(...)
A NowSecure audit of the app has found other behaviors that researchers found potentially concerning. For instance, the app uses a symmetric encryption scheme known as 3DES or triple DES. The scheme was deprecated by NIST following research in 2016 that showed it could be broken in practical attacks to decrypt web and VPN traffic. Another concern is that the symmetric keys, which are identical for every iOS user, are hardcoded into the app and stored on the device.

The app is “not equipped or willing to provide basic security protections of your data and identity,” NowSecure co-founder Andrew Hoog told Ars. “There are fundamental security practices that are not being observed, either intentionally or unintentionally. In the end, it puts your and your company’s data and identity at risk.”"

https://arstechnica.com/security/2025/02/deepseek-ios-app-sends-data-unencrypted-to-bytedance-controlled-servers/

Researchers find major security issues in DeepSeek AI! #DeepSeekR1 fails 58% of jailbreak tests, exposing compliance risks and vulnerabilities.

Read: https://hackread.com/deepseek-r1-llm-fail-jailbreak-attack-security-analysis/

Cisco study finds DeepSeek R1 fails all safety tests, with a 100% jailbreak success rate!

Read: https://hackread.com/cisco-finds-deepseek-r1-vulnerable-harmful-prompts/

"The Hangzhou-based company’s decision to release a low-cost, open-sourced AI model, alongside detailed disclosure of its training methods, means that everyone, from researchers in São Paulo to start-ups in Stockholm and doctors in Nairobi, can access state-of-the-art AI at little to no cost. 

Within the Chinese start-up sector a chain reaction is taking place. New AI applications are being created. Competition is going to become more fierce. Risk appetite from early-stage venture investment is increasing. DeepSeek’s decision to pursue an open-source AI model is inspiring and putting pressure on others to do the same. The first to react was Alibaba’s Qwen team, which released Qwen2.5 as open source last month on the eve of Chinese new year. 

This is a remarkable change. After the US start-up OpenAI released its generative AI model ChatGPT in late 2022, the global digital economy was edging towards control by a handful of tech giants. These players chase scale over efficiency — building ever-larger models that demand staggering compute, energy and capital while guarding their training methods as trade secrets. 

Centralised, closed models create a dangerous feedback loop. The more data they amass, the more powerful they become, further marginalising anyone outside their gates. For consumers this means large fees, surrendered data and watching AI’s future unfold without meaningful participation."

https://www.ft.com/content/3549cc33-e04d-41da-8c58-525d5bb2ba4c?accessToken=zwAGLSXfgO_okc81Scwz4E1B2tOMWFJdW7K6TA.MEUCIQC5tk93DqeDHWa4AouV5J9_FFlU2ddy0ltquDNoZ2tjdgIgFXHDwinDfvNpMJaFOsPPZWDivL5yE74shCNFELLhetQ&sharetype=gift&token=86e55301-a539-4320-8cd5-fdd68aa93cc6

This memo from May 2023 was spot on:
Google: “We Have No Moat, And Neither Does OpenAI”, Directly Competing With Open Source Is a Losing Proposition.
https://semianalysis.com/2023/05/04/google-we-have-no-moat-and-neither/
#ai #opensource #openai #deepseekR1

OpenAI , even after releasing o3-mini , has a problem in comparison with Deepseek R1 , it's a lot more expensive, also not opensource,(R1 is only partly opensource) you can't download it or run it local. The transparent thinking that R1 has, is also missing in o3.
#ai #openai #deepseekR1 #o3mini

to 14b or not to 14b
that is the question

DEEPSEEK..
I know what you're thinkin'
I don't need your reasons
Don't tell me 'cause it hurts

"DeepSeek has commoditized the Large Language Model, publishing both the source code and the guide to building your own. Whether or not someone chooses to pay DeepSeek is largely irrelevant — someone else will take what it’s created and build their own, or people will start running their own DeepSeek instances renting GPUs from one of the various cloud computing firms.

While NVIDIA will find other ways to make money — Jensen Huang always does — it's going to be a hard sell for any hyperscaler to justify spending billions more on GPUs to markets that now know that near-identical models can be built for a fraction of the cost with older hardware. Why do you need Blackwell? The narrative of "this is the only way to build powerful models" no longer holds water, and the only other selling point it has is "what if the Chinese do something?"

Well, the Chinese did something, and they've now proven that they can not only compete with American AI companies, but do so in such an effective way that they can effectively crash the market.

It still isn't clear if these models are going to be profitable — as discussed, it's unclear who funds DeepSeek and whether its current pricing is sustainable — but they are likely going to be a damn sight more profitable than anything OpenAI is flogging. After all, OpenAI loses money on every transaction — even its $200-a-month "ChatGPT Pro" subscription. And if OpenAI cuts its prices to compete with DeepSeek, its losses will only deepen."

https://www.wheresyoured.at/deep-impact/

Hugging Face plant Open-Source-Nachbau von DeepSeek R1
Das KI-Modell DeepSeek R1 hat seit seiner Veröffentlichung für Aufmerksamkeit gesorgt. Nun will das Team von Hugging Face eine offene Alternative schaffen. Das neue Projekt trägt den Namen Ope
https://www.apfeltalk.de/magazin/news/hugging-face-plant-open-source-nachbau-von-deepseek-r1/
#News #Services #DeepSeekR1 #HuggingFace #KI #KnstlicheIntelligenz #MachineLearning #OpenR1 #OpenSource #PermissiveLizenz #Trainingsdaten #Transparenz