AINews news.smol.ai

ALL of AI Engineering in One Place

news.smol.ai·103w

Anthropic's "LLM Genome Project": learning & clamping 34m features on Claude Sonnet

news.smol.ai·103w

Skyfall

news.smol.ai·103w

Chameleon: Meta's (unreleased) GPT4o-like Omnimodal Model

news.smol.ai·104w

Cursor reaches >1000 tok/s finetuning Llama3-70b for fast file editing

news.smol.ai·104w

Not much happened today

news.smol.ai·104w

Google I/O in 60 seconds

news.smol.ai·104w

GPT-4o: the new SOTA-EVERYTHING Frontier model (GPT4T version)

news.smol.ai·104w

GPT-4o: the new SOTA-EVERYTHING Frontier model (GPT4O version)

news.smol.ai·104w

Quis promptum ipso promptiet?

news.smol.ai·105w

LMSys advances Llama 3 eval analysis

news.smol.ai·105w

OpenAI's PR Campaign?

news.smol.ai·105w

Kolmogorov-Arnold Networks: MLP killers or just spicy MLPs?

news.smol.ai·105w

DeepSeek-V2 beats Mixtral 8x22B with >160 experts at HALF the cost

news.smol.ai·105w

$100k to predict LMSYS human preferences in a Kaggle contest

news.smol.ai·106w

Evals: The Next Generation

news.smol.ai·106w

Not much happened today

news.smol.ai·106w

LLMs-as-Juries

news.smol.ai·106w

A quiet weekend

news.smol.ai·106w

Apple's OpenELM beats OLMo with 50% of its dataset, using DeLighT

news.smol.ai·107w

Log in to enable infinite scrolling