10 results on this page
Meta is launching Meta Startup School, a three-month programme for early-stage consumer brands. Its first cohort will include 200 startups receiving training and support from venture capital firms and industry experts.
Google reportedly bid $10 million for Spirit Airlines’ internal enterprise data, including approximately 100 million emails and 500 million Microsoft Teams messages, in a bankruptcy asset auction. The deal signals the emergence of a market for acquiring sanitized corporate communications and work records to train AI systems and enterprise agents.
OpenAI has slowed parts of its AI development to strengthen security and safeguards. The company paused reinforcement-learning training for two weeks on its latest deployment-focused models and delayed its largest planned frontier RL run.
OpenAI has introduced a monitoring system for its riskiest AI work and says it adds about 20% to the compute cost of covered workloads. The company also paused some frontier training for two weeks as part of the changes.
AI data startup Micro1 has reached a $500 million gross run rate as demand for AI training data accelerates. The growth highlights expanding opportunities for data providers supporting model development.
AntLing has released six official Ling-3.0 base checkpoints spanning two model sizes and three training stages. The MIT-licensed, ungated repositories are intended for continued pretraining, fine-tuning, and research rather than direct chat use.
An independent builder pretrained a 1.02-billion-parameter Kimi-K3 replica on 5 billion decontaminated tokens for roughly $250. The model achieved 33.4% on HellaSwag, exceeding the reported 28% score of GPT-2 124M despite receiving no instruction tuning.
The Spectral Neuron is a proposed ML primitive based on the largest eigenvalue of an input-conditioned matrix. The preprint develops its mathematical properties, initialization and training methods, and evaluates scalability on synthetic and real-world datasets.
ByteDance’s Seed foundation-model division has reportedly completed another internal restructuring. Its foundation-model organization now includes four first-level departments focused on pretraining data, reinforcement learning, product post-training for work, and product post-training for chat.
Apple Machine Learning studies how to mix scarce target-domain data with abundant generic data during language-model pretraining. Across more than 2,000 training runs, the research examines the trade-off between insufficient target exposure, repeated examples, diminishing returns, and overfitting.