Topic: All topics
All articles

Index

All Articles

Everything published here, newest first.

76 articles · Page 1 of 8

Adding paragraph breaks to speech-to-text output with Gemma 3 270M

Adding paragraph breaks to speech-to-text output with Gemma 3 270M

A Gemma 3 270M fine-tune that adds paragraph breaks to flat speech-to-text output and changes nothing else. Small enough for on-device use, with an ONNX export. Break F1 0.78 on 26 held-out memos.

ai-models hugging-face fine-tuning +4
Bolding the key lines of a dictated email with Gemma 3 270M

Bolding the key lines of a dictated email with Gemma 3 270M

A preliminary Gemma 3 270M fine-tune that bolds the key lines of a dictated email (the asks, the purpose, the key facts) and changes nothing else. Small enough for on-device use. Word F1 0.62 on six held-out emails.

ai-models hugging-face fine-tuning +3
Jachnun and Sabich: the first stable Hebrish models, in depth

Jachnun and Sabich: the first stable Hebrish models, in depth

A deep dive into the first stable Hebrish models, Jachnun v1.0 and Sabich v1.0: the data, the training recipe, the results against stock Parakeet and Whisper, the formats, and the full white paper.

hebrish speech-recognition code-switching +3
Hebrish: one home for my English-Hebrew speech models

Hebrish: one home for my English-Hebrew speech models

Hebrish is my one-person, not-for-profit home for English-Hebrew speech models: Jachnun, Sabich, Shakshuka and the Whisper fine-tunes, free apps, open data and benchmarks, all at hebrish.app.

hebrish speech-recognition code-switching +2
Proximate Pseudonymizer: a small model that swaps identifying details for close substitutes

Proximate Pseudonymizer: a small model that swaps identifying details for close substitutes

A proof-of-concept Qwen3-0.6B fine-tune that rewrites text with names, places and dates swapped for close, consistent substitutes (Daniel to David, Jerusalem to Tel Aviv). Runs on CPU. Still leaks 9.3% of details, so review every output.

ai-models hugging-face fine-tuning +2
Boxty v1.0: English dictation without the Hebrew

Boxty v1.0: English dictation without the Hebrew

Boxty is the no-Hebrew sibling of my Jachnun code-switching model: Parakeet fine-tuned on my voice for plain English and tech dictation. Tech WER drops from 2.51% (stock) to 1.35%, and general English holds.

speech-recognition parakeet fine-tuning +3
Do LLMs push back on conspiracy believers? A small first look

Do LLMs push back on conspiracy believers? A small first look

A small, hand-graded first look at my conspiracy-endorsement eval: 7 prompts, 7 models, one sample each. In this run two uncensored models helped a believer and an invented theory tripped up two more. Not a verdict on any model.

evals llm ai-models +3
My Weird Prompts passes 6,000 episodes
Update

My Weird Prompts passes 6,000 episodes

My Weird Prompts, my fully automated AI podcast, crossed 6,000 episodes on 15 October 2026: 5,991 in the archive, 2,688 hours of audio, and new translated editions in eight languages.

my-weird-prompts text-to-speech ai-agents +3
Voice typing that understands Hebrish

Voice typing that understands Hebrish

Speech-to-text mangles the Hebrew words English speakers in Israel use every day. Hebrish fixes that: free on-device apps for Android, Linux and Windows, the Jachnun and Sabich models, results, and how you can help.

speech-recognition code-switching hebrew +3
Whisper Hebrish Small ACFT: a code-switching model for FUTO Keyboard
Update

Whisper Hebrish Small ACFT: a code-switching model for FUTO Keyboard

A phone-sized Whisper Hebrish for FUTO Keyboard on Android. It halves the stock English model's error rate on English-with-Hebrew dictation (11.8% to 5.8% WER) and imports as a single 264 MB file.

ai-models whisper speech-recognition +4