ICYMI: We've partnered with @HuggingFace to bring 10x faster #AI inference speeds to devs! @AI at Meta's Llama 3 & @Alibaba_Qwen models SambaNova Cloud's Llama Guard & Qwen QwQ Easy integration with minimal code changes @deepseek_ai coming soon! Try it now
GENERATIVE AI
-
DeepSeek Shifts: China Advances, Open Models Commoditize AI
By
–
The buzz over DeepSeek this week crystallized, for many people, a few important trends that have been happening in plain sight: (i) China is catching up to the U.S. in generative AI, with implications for the AI supply chain. (ii) Open weight models are commoditizing the
-
Bloomberg segment on AI and DeepSeek developments
By
–
I got on Bloomberg News yesterday to talk about AI and the DeepSeek hullabaloo. You can find my segment starting on minute 58.
-
Mistral Small 3 Available on Partner Platforms
By
–
In addition to being available on la Plateforme (
http://
console.mistral.ai), Mistral Small 3 is available on our partner platforms -
Chaining Multiple AI Models for Optimized Workflow Performance
By
–
@RodrigoLiang shared with @jeremyakahn of @FortuneMagazine his insights on #AI: The next phase involves chaining multiple models to create seamless workflows
Our chips are 100x faster, with 1/10th the power
In workflows with multiple models, delays add up Read more -
Base Models Commoditized: Building AI Flywheels Strategy
By
–
It's increasingly clear that the base models will be commoditized (Deepseek) I would argue that that defining companies will invest not in models, but rather in flywheels around the model We're talking about how to build these flywheels next Tuesday: https://
lu.ma/w7y0bqwr -
Build Beautiful AI Apps with 21st.dev and Replit Agent
By
–
Need a beautiful design for your next Replit app?
— Replit ⠕ (@Replit) 30 janvier 2025
Head to 21st.dev and:
– Find a design
– Copy the prompt into Replit Agent
– Iterate and add Replit’s built-in tools and databases
– Deploy your app to the public
– Profit
Share what you build and tag us! pic.twitter.com/PoEREiNtveNeed a beautiful design for your next Replit app? Head to 21st.dev and: – Find a design
– Copy the prompt into Replit Agent
– Iterate and add Replit’s built-in tools and databases
– Deploy your app to the public
– Profit Share what you build and tag us! -
Cerebras launches world’s fastest DeepSeek R1 Llama-70B inference
By
–
Cerebras DeepSeek R1 Llama-70B is available now on http://
cerebras.ai, and we're offering API preview access to select customers. Read our blog to learn more. https://
cerebras.ai/blog/cerebras-
launches-worlds-fastest-deepseek-r1-llama-70b-inference
… -
Cerebras Inference Makes Reasoning Models Instant
By
–
Reasoning models are powerful but can take minutes to generate the final answer. Cerebras Inference makes AI instant again. In the coding example below, Cerebras Inference R1 70B returns the answer in 1.5 seconds vs. 22 seconds using o1 mini. pic.twitter.com/43TQ9YVl8l
— Cerebras (@cerebras) 30 janvier 2025Reasoning models are powerful but can take minutes to generate the final answer. Cerebras Inference makes AI instant again. In the coding example below, Cerebras Inference R1 70B returns the answer in 1.5 seconds vs. 22 seconds using o1 mini.
-
DeepSeek R1 70B Outperforms GPT-4o and o1-mini
By
–
DeepSeek’s R1 70B combines the powerful reasoning ability of the full R1 model with the size and speed of Llama 70B. R1 70B outperforms GPT-4o and o1-mini across a range of general and reasoning benchmarks, making it the most capable Llama 70B variant by far.