This update doesn’t support the o1-style thinking UI btw For now you can enjoy r1’s raw CoT
LLMS
-
Apollo 1.0.23 Brings Compact DeepSeek r1 Models to Mobile
By
–
Apollo 1.0.23 is In Review and has the smaller DeepSeek r1 models ready to try. I suspect these will be the best models you can run on a phone.
-
ChatGPT adds memory for personalized search
By
–
Memory in search is a new toggle which will allow ChatGPT to use context from memories when performing search to provide personalised answers
-
Hugging Face Inference Warm: Access Trending AI Models
By
–
yes it's inference=warm: https://
huggingface.co/models?inferen
ce=warm&sort=trending
… (doc https://
huggingface.co/docs/api-infer
ence/supported-models#supported-models
…) -
3 AI tools for creating captivating presentations with a ChatGPT trick
By
–
Done #PowerPoint With these 3 #AI tools you will create captivating presentations without spending all day → https://youtu.be/9TVJ2gauvwI Bonus: a #ChatGPT trick to create the perfect slide outline
-
Small AI Models Disappoint: Benchmark vs Real-World Performance
By
–
I think benchmark real and in-practice real are different. Maybe it’s true, but I need to try. All small models disappoint me, if it the theory of distilling out noise makes sense.
-
DeepSeek-R1 released, powers Reasoning mode
By
–
BREAKING : DeepSeek-R1 is out and powering Reasoning mode on DeepSeek apps already
-
ChatGPT Gains One IQ Point Every 4.5 Days
By
–
Il n'y a qu'un Français sur 16.000 qui est plus intelligent que la dernière version de ChatGPT ChatGPT gagné un point de QI tous les quatre jours et demi !
