https://goodomens.studio/https://goodomens.studio/about/https://goodomens.studio/blog/https://goodomens.studio/blog/2/https://goodomens.studio/blog/3/https://goodomens.studio/blog/4/https://goodomens.studio/blog/5/https://goodomens.studio/blog/6/https://goodomens.studio/blog/ai-agents-when-ai-stops-answering-and-starts-doing/https://goodomens.studio/blog/ai-and-ml-actually-explained/https://goodomens.studio/blog/ai-bias-and-fairness-why-just-the-data-isnt-neutral/https://goodomens.studio/blog/below-four-bits-qat-bitnet-and-the-race-toward-one-bit-intelligence/https://goodomens.studio/blog/building-a-fine-tuning-dataset-the-practical-craft-nobody-writes-papers-about/https://goodomens.studio/blog/building-blocks-twelve-pieces-one-working-model/https://goodomens.studio/blog/building-blocks-you-built-the-whole-chain/https://goodomens.studio/blog/dimensions-why-ai-thinks-in-thousands-of-directions-at-once/https://goodomens.studio/blog/dqn-the-moment-reinforcement-learning-met-deep-learning/https://goodomens.studio/blog/embeddings-how-meaning-becomes-geometry/https://goodomens.studio/blog/embeddings-where-similar-ideas-live-near-each-other/https://goodomens.studio/blog/from-raw-web-to-training-corpus-the-unglamorous-pipeline-behind-every-llm/https://goodomens.studio/blog/generative-ai-how-computers-learned-to-create/https://goodomens.studio/blog/gguf-and-the-art-of-choosing-a-quant-a-field-guide-to-local-models/https://goodomens.studio/blog/how-models-actually-learn-gradient-descent-and-backpropagation-from-first-principles/https://goodomens.studio/blog/i-built-an-11200-example-synthetic-dataset-fine-tuned-a-model-and-then-didnt-ship-it/https://goodomens.studio/blog/i-trained-a-board-game-ai-from-zero-using-self-play-and-broke-it-several-times-first/https://goodomens.studio/blog/keeping-a-vision-pipeline-honest-three-hallucination-guards-i-built-for-a-video-captioning-agent/https://goodomens.studio/blog/kimi-k3/https://goodomens.studio/blog/life-after-training-merging-loras-stacking-them-and-serving-a-hundred-fine-tunes-on-one-gpu/https://goodomens.studio/blog/lora-why-fine-tuning-a-billion-parameter-model-only-takes-a-few-million-parameters/https://goodomens.studio/blog/open-source-vs-closed-ai-who-owns-intelligence/https://goodomens.studio/blog/optimizing-a-model-the-complete-guide-to-distillation-quantization-pruning-and-beyond/https://goodomens.studio/blog/policy-gradients-and-ppo-learning-behavior-directly/https://goodomens.studio/blog/prompt-engineering-why-talking-to-ai-is-a-skill/https://goodomens.studio/blog/qlora-the-paper-that-put-large-model-fine-tuning-on-a-desk/https://goodomens.studio/blog/quality-vs-quantity-the-data-question-that-keeps-changing-its-answer/https://goodomens.studio/blog/quantization-from-the-bits-up-how-4-bit-models-actually-work/https://goodomens.studio/blog/reinforcement-learning-from-zero-agents-rewards-and-the-loop-that-learns-from-consequences/https://goodomens.studio/blog/rlhf-and-beyond-how-reinforcement-learning-taught-language-models-to-behave/https://goodomens.studio/blog/scaling-laws-the-physics-of-machine-learning/https://goodomens.studio/blog/shrinking-a-2-8-trillion-parameter-model/https://goodomens.studio/blog/synthetic-data-training-models-on-text-that-models-wrote/https://goodomens.studio/blog/tensors-the-shape-data-takes-inside-ai/https://goodomens.studio/blog/the-agi-debate-are-we-close-or-is-it-all-hype/https://goodomens.studio/blog/the-chip-that-learned-to-think-in-parallel/https://goodomens.studio/blog/the-eval-that-broke-containment/https://goodomens.studio/blog/the-lora-knobs-rank-alpha-targets-and-the-settings-everyone-copies-without-understanding/https://goodomens.studio/blog/the-model-already-knew-the-answer/https://goodomens.studio/blog/the-outlier-problem-the-discovery-that-shaped-modern-quantization/https://goodomens.studio/blog/the-transformer-explained-the-way-i-wish-someone-had-explained-it-to-me/https://goodomens.studio/blog/tokens-revisited-how-text-becomes-math/https://goodomens.studio/blog/training-vs-inference-the-same-model-two-very-different-jobs/https://goodomens.studio/blog/transformers-and-attention-the-idea-that-changed-everything/https://goodomens.studio/blog/vectors-turning-meaning-into-a-list-of-numbers/https://goodomens.studio/blog/weights-and-parameters-the-billions-of-knobs-inside-a-model/https://goodomens.studio/blog/what-it-looks-like-when-ai-cheats/https://goodomens.studio/blog/whats-actually-inside-chatgpt/https://goodomens.studio/blog/will-ai-take-my-job/https://goodomens.studio/case-studies/https://goodomens.studio/case-studies/arte-corporate/https://goodomens.studio/case-studies/arte-presse/https://goodomens.studio/case-studies/ben-greenlight/https://goodomens.studio/case-studies/eco/https://goodomens.studio/case-studies/forbes-portugal/https://goodomens.studio/case-studies/leva-me/https://goodomens.studio/categories/https://goodomens.studio/categories/ai-systems/https://goodomens.studio/categories/ai-systems/2/https://goodomens.studio/categories/data-engineering/https://goodomens.studio/categories/engineering/https://goodomens.studio/categories/machine-learning/https://goodomens.studio/categories/machine-learning/2/https://goodomens.studio/categories/machine-learning/3/https://goodomens.studio/categories/reinforcement-learning/https://goodomens.studio/contact/https://goodomens.studio/plugins/https://goodomens.studio/plugins/favorite-admin-bookmarks/https://goodomens.studio/series/https://goodomens.studio/series/building-blocks/https://goodomens.studio/series/data-engineering/https://goodomens.studio/series/deep-dives/https://goodomens.studio/series/foundations/https://goodomens.studio/series/lora-deep-dive/https://goodomens.studio/series/ml-foundations/https://goodomens.studio/series/quantization/https://goodomens.studio/series/reinforcement-learning/https://goodomens.studio/series/when-ai-cheats/https://goodomens.studio/tags/https://goodomens.studio/tags/1-bit-models/https://goodomens.studio/tags/adapters/https://goodomens.studio/tags/agi/https://goodomens.studio/tags/ai-agents/https://goodomens.studio/tags/ai-and-work/https://goodomens.studio/tags/ai-ethics/https://goodomens.studio/tags/ai-forecasting/https://goodomens.studio/tags/ai-policy/https://goodomens.studio/tags/ai-safety/https://goodomens.studio/tags/ai-systems/https://goodomens.studio/tags/alignment/https://goodomens.studio/tags/ani-agi-asi/https://goodomens.studio/tags/architecture/https://goodomens.studio/tags/artificial-intelligence/https://goodomens.studio/tags/attention/https://goodomens.studio/tags/automation/https://goodomens.studio/tags/autonomy/https://goodomens.studio/tags/awq/https://goodomens.studio/tags/backpropagation/https://goodomens.studio/tags/bias/https://goodomens.studio/tags/bitnet/https://goodomens.studio/tags/building-blocks/https://goodomens.studio/tags/byte-pair-encoding/https://goodomens.studio/tags/chatgpt/https://goodomens.studio/tags/chinchilla/https://goodomens.studio/tags/common-crawl/https://goodomens.studio/tags/compute-budgets/https://goodomens.studio/tags/cosine-similarity/https://goodomens.studio/tags/data-curation/https://goodomens.studio/tags/data-pipelines/https://goodomens.studio/tags/data-quality/https://goodomens.studio/tags/decoding/https://goodomens.studio/tags/deduplication/https://goodomens.studio/tags/diffusion-models/https://goodomens.studio/tags/dimensions/https://goodomens.studio/tags/distillation/https://goodomens.studio/tags/dpo/https://goodomens.studio/tags/dqn/https://goodomens.studio/tags/embeddings/https://goodomens.studio/tags/evaluation/https://goodomens.studio/tags/experience-replay/https://goodomens.studio/tags/explainer/https://goodomens.studio/tags/explainer/2/https://goodomens.studio/tags/explainer/3/https://goodomens.studio/tags/fairness/https://goodomens.studio/tags/fine-tuning/https://goodomens.studio/tags/forward-pass/https://goodomens.studio/tags/generative-ai/https://goodomens.studio/tags/gguf/https://goodomens.studio/tags/gptq/https://goodomens.studio/tags/gpus/https://goodomens.studio/tags/gradient-descent/https://goodomens.studio/tags/hallucination/https://goodomens.studio/tags/hardware/https://goodomens.studio/tags/hugging-face/https://goodomens.studio/tags/hyperparameters/https://goodomens.studio/tags/image-generation/https://goodomens.studio/tags/inference/https://goodomens.studio/tags/instruction-tuning/https://goodomens.studio/tags/int4/https://goodomens.studio/tags/kimi-k3/https://goodomens.studio/tags/large-language-models/https://goodomens.studio/tags/linear-algebra/https://goodomens.studio/tags/llamacpp/https://goodomens.studio/tags/local-models/https://goodomens.studio/tags/lora/https://goodomens.studio/tags/loss-functions/https://goodomens.studio/tags/low-rank-adaptation/https://goodomens.studio/tags/machine-learning/https://goodomens.studio/tags/markov-decision-process/https://goodomens.studio/tags/mixture-of-experts/https://goodomens.studio/tags/modal/https://goodomens.studio/tags/model-collapse/https://goodomens.studio/tags/model-serving/https://goodomens.studio/tags/model-size/https://goodomens.studio/tags/moonshot-ai/https://goodomens.studio/tags/neural-networks/https://goodomens.studio/tags/next-token-prediction/https://goodomens.studio/tags/nf4/https://goodomens.studio/tags/numerical-precision/https://goodomens.studio/tags/ocr/https://goodomens.studio/tags/onnx/https://goodomens.studio/tags/open-models/https://goodomens.studio/tags/open-weights/https://goodomens.studio/tags/optimization/https://goodomens.studio/tags/optimizers/https://goodomens.studio/tags/outliers/https://goodomens.studio/tags/parallel-computing/https://goodomens.studio/tags/parameter-efficient-tuning/https://goodomens.studio/tags/parameters/https://goodomens.studio/tags/policy-gradients/https://goodomens.studio/tags/post-mortem/https://goodomens.studio/tags/ppo/https://goodomens.studio/tags/prompt-engineering/https://goodomens.studio/tags/pruning/https://goodomens.studio/tags/q-learning/https://goodomens.studio/tags/qat/https://goodomens.studio/tags/qlora/https://goodomens.studio/tags/quantization/https://goodomens.studio/tags/rank/https://goodomens.studio/tags/reasoning-models/https://goodomens.studio/tags/reasoning/https://goodomens.studio/tags/reinforcement-learning/https://goodomens.studio/tags/representation/https://goodomens.studio/tags/reward-functions/https://goodomens.studio/tags/reward-hacking/https://goodomens.studio/tags/rlhf/https://goodomens.studio/tags/sandbox-escape/https://goodomens.studio/tags/scaling-laws/https://goodomens.studio/tags/security/https://goodomens.studio/tags/self-attention/https://goodomens.studio/tags/self-play/https://goodomens.studio/tags/semantic-search/https://goodomens.studio/tags/semantic-similarity/https://goodomens.studio/tags/series-closing/https://goodomens.studio/tags/series-introduction/https://goodomens.studio/tags/shapes/https://goodomens.studio/tags/specification-gaming/https://goodomens.studio/tags/stochastic-resonance/https://goodomens.studio/tags/synthetic-data/https://goodomens.studio/tags/tensor-cores/https://goodomens.studio/tags/tensors/https://goodomens.studio/tags/the-labour-market/https://goodomens.studio/tags/tokenization/https://goodomens.studio/tags/tool-use/https://goodomens.studio/tags/training-data/https://goodomens.studio/tags/training-pipelines/https://goodomens.studio/tags/training/https://goodomens.studio/tags/transformers/https://goodomens.studio/tags/trpo/https://goodomens.studio/tags/vector-space/https://goodomens.studio/tags/vectors/https://goodomens.studio/tags/vision-models/https://goodomens.studio/tags/vocabulary/https://goodomens.studio/tags/weights/