You have to admit that OpenAI has delivered a lot. Not a new model, but still lots of improvements and new features. And as Adam says: the year is not yet ignited. I for one am not disappointed, on the contrary. SearchGPT alone is a great release.
LLMS
-

Smol LM v2 Released: New Open-Source Language Model
By
–
The freshly released Smol LM v2 https://
huggingface.co/collections/Hu
ggingFaceTB/smollm2-6723884218bcda64b34d7db9
… -
Matthew Berman News Update Information for AI Development
By
–
@MatthewBerman in case you missed the news 🙂 much information for the next AI update 🙂
-
OpenAI AMA Highlights: Sora, New Models, AGI, and o1
By
–
I read the whole AMA with @OpenAI and especially @sama . In this thread I post everything they said about Sora, new models, AGI, o1 and so on. Usually I don’t do threads, but this time: it’s a
-

Sam Altman Reddit AMA: GPT-5 delayed, AGI possible, new DALL-E coming
By
–
Sam Altman gave an AMA on Reddit. Here are the most important takeaways (I am dead serious): – no GPT-5 this year! But other great releases end of the year
– AGI is achievable with this architecture!
– new dalle is coming and it will worth the wait – larger context-window -

Sam Altman Predicts Benchmark Saturation Next Year
By
–
Sam Altman: Next year, the benchmarks will be "saturated".
On Reddit, Sam Altman holds an "ask me anything" session.
Insane. -

SmolLM v2 Releases State-of-Art Small Language Models
By
–
It’s a huge day for small models!! Super proud to release SmolLM v2 pushing the state-of-the-art performances of LLMs under 2B parameters with three sizes: 135M, 360M and 1.7B parameters. These models have been several months in the making with many lessons learned that we’ll
-

SmolLM2: Faster, Better, Cheaper Language Model
By
–
SmolLM2 – faster, better and cheaper! Intelligence is definitely too cheap to meter.
-

LLMs Enhanced by Emotional Stimuli Understanding
By
–
Large Language Models Understand and Can be Enhanced by Emotional Stimuli https://
bit.ly/3N8wt5f
#AI #MachineLearning #DeepLearning #LLMs #DataScience