Skip to content
{ ken ashe }
  • Building
  • Writing
  • About
  • Newsroom
  • Digest
  • Building
  • Writing
  • About
  • Newsroom
  • Digest
← Digest / Tags

// tag

model-training

38 posts tagged model-training.

  • Free local AI models are not charity Sep 27, 2026
  • Clifford-VAE puts pixels into symbolic memory space Sep 24, 2026
  • UECR-GRPO treats the teacher as evidence, not an oracle Sep 24, 2026
  • Critical-State RL trains the one agent call that actually matters Sep 22, 2026
  • onPanda turns alignment feedback into token-level steering Sep 22, 2026
  • Where Harness Self-Improvement Actually Stands in Late 2026 Sep 22, 2026
  • Membership Inference Gets an Entropy Correction: Reading the ETD Paper Sep 21, 2026
  • Mini-AGI makes local training the interesting part Sep 21, 2026
  • Softmax attention needs an off switch Sep 21, 2026
  • Microsoft’s AI scraping memo points to a data supply chain problem Sep 18, 2026
  • Paint-Anything makes hex colors a first-class diffusion control Sep 18, 2026
  • Double descent as an implicit regularization story Sep 17, 2026
  • What Actually Makes a Tokeniser Good: Search Beats Objective Sep 17, 2026
  • Distillation needs calibration when the teacher is biased Sep 16, 2026
  • Bellman Policy Optimization cuts one moving part from RLVR Sep 15, 2026
  • Federated learning gets more practical when privacy and timing are treated together Sep 15, 2026
  • OptiFlow treats offline RL policy learning as sample matching Sep 15, 2026
  • Garry Tan’s distillation argument is really about AI capability access Sep 14, 2026
  • The useful part of calling Nvidia AI’s central bank Sep 13, 2026
  • CRISPR screens need learned experiment pickers, not bigger chatbots Sep 11, 2026
  • Recursive Self-Improvement Has a Roadmap Now, and Most of It Isn't Built Sep 11, 2026
  • Layer-selective unlearning aims at the parts of a model that remember Sep 10, 2026
  • On-policy distillation may be pruning tails, not teaching Sep 1, 2026
  • Optimizers Are Becoming Systems Choices, Not AdamW Replacements Aug 31, 2026
  • Reward choice changes how LLM forecasters are wrong Aug 31, 2026
  • Alignment baked into pretraining, not bolted on later Aug 14, 2026
  • Selective Trust: Why RAG Systems Should Learn When to Ignore Their Own Context Aug 7, 2026
  • MindForge trains coding agents on blank-repo software work Jul 30, 2026
  • Hyperball optimizers still need learning-rate discipline Jul 27, 2026
  • MIRROR trains vision models by making each modality teach the others Jul 24, 2026
  • OPD2 tries to distill reasoning by subtracting the base model Jul 17, 2026
  • RL post-training as procedure compression, not just skill amplification Jul 9, 2026
  • Timestamp drift is the quiet ASR failure that breaks real workflows Jul 7, 2026
  • DemoPSD Treats Teacher Disagreement as a Training Signal Jul 3, 2026
  • RLVR needs a taste model, not just a grader Jul 2, 2026
  • Synthetic QA Has a Selection Problem Before It Has a Training Problem Jul 1, 2026
  • Autodata Turns Synthetic Data Generation Into an Agent You Train Jun 25, 2026
  • Self-distillation can make models better on the first try and worse on the fifth Jun 25, 2026

Ken Ashe ·AI application builder ·CPA ·PMP

Building with AI in public. No hype, no doom. Receipts only.

hello@kenashe.ai

Explore

  • Building
  • Writing
  • Digest
  • Topics

About & Press

  • About
  • Newsroom
  • Media Kit
  • Lucky Domains

Social

  • LinkedIn
  • X
  • GitHub
  • RSS

Legal

  • Privacy
  • Terms
  • Disclosure

© 2026 Ken Ashe ·Built with AI in public