This post provides a comprehensive guide on training advanced text embedding models, highlighting key techniques and the importance of filtering negative examples for better performance.
About:
Jack Morris is the author of a Substack publication with hundreds of subscribers.
Website:
Specializations:
Outgoing Links:
Subscribe to RSS:
The blog post discusses Project CETI, an initiative aimed at decoding whale communication using AI models. It draws parallels between human semantic understanding and AI's ability to compress and model data. The author explains ho...
The blog post discusses the potential of reinforcement learning (RL) as a training technique for advanced AI models, particularly in the context of reasoning models. The author reflects on their journey from skepticism to recogniz...
The post discusses the challenges and considerations in building superintelligent AI, focusing on the use of neural networks, particularly Transformers, and the importance of text data for training. It explores the limitations of ...
The post discusses the progress of AI over the past fifteen years, particularly focusing on the last five. It highlights the concept of a 'Moore's Law for AI' and the continuous improvement in AI capabilities, driven by research f...
The post critiques the common speculation surrounding Artificial General Intelligence (AGI) timelines, arguing that the focus should shift from the mythical concept of AGI to the measurable productivity of AI systems relative to h...
The article introduces the concept of superintelligence and its historical context, tracing the evolution of artificial intelligence (AI) from early pioneers like Charles Babbage and Ada Lovelace to contemporary advancements in la...
The post discusses the stagnation in AI research despite an increase in the number of papers published. It argues that while the field has seen exponential growth in submissions, the actual scientific progress remains constant, su...
The author reflects on their experience working remotely at Google during the pandemic, highlighting concerns about the potential for employers to collect and utilize behavioral data to create AI models that replicate employee act...