Who's Who of AI

austegard.com

36 trust practitioner @austegard.com · 442 followers
oskar @ austegard.com 🕊️ AI Explorer - caveat vibrans Evolution guide for Muninn 🐦‍⬛ (muninn.austegard.com) Yeah not actually green. Not really that grouchy either.
What they're sharing

Articles & links

Website: vamsin07.github.io/buzzasr-docs/ Tokenizers: github.com/vamsin07/mul... SFT models: github.com/vamsin07/whi...

GitHub - vamsin07/multilingual-bpe-tokenizers: Whisper-compatible per-lang byte-level BPE tokenizers (recipe + samples) · GitHub github.com
AI Weekly's analysis →
  • A new open-source repo publishes byte-level BPE tokenizers for languages including Kamba, Arabic, Mandarin, Cantonese, Japanese, Spanish, and Swahili.
  • The recipe keeps a 51,865-token vocabulary matching Whisper-large-v3 and lifts the max token length from 16 to 32 bytes for multi-byte scripts.
  • An audit across 102 FLEURS languages reports cross-word merges dropping from 2,656,091 to zero, with 100% round-trip integrity.
Read full analysis →
View on Bluesky · ♥ 1 ↻ 1 ↩ 0 · 2 from the directory shared this · 4d ago

Website: vamsin07.github.io/buzzasr-docs/ Tokenizers: github.com/vamsin07/mul... SFT models: github.com/vamsin07/whi...

GitHub - vamsin07/whisper-simple-finetune: Vanilla Whisper fine-tuning on FLEURS — clean starter repo with Nautilus deployment · GitHub github.com
AI Weekly's analysis →
  • The MIT-licensed repo ships a training script, FLEURS evaluation, W&B sweep configs, and Kubernetes manifests for Nautilus deployment.
  • Default training uses AdamW with a 0.3 encoder/decoder learning-rate ratio, a 150-step cosine warmup, and a maximum of 6 epochs.
  • The base Whisper model reportedly fine-tunes on a local GPU in 30 to 60 minutes; the repo currently shows 0 stars, 0 forks, and 3 commits.
Read full analysis →
View on Bluesky · ♥ 1 ↻ 1 ↩ 0 · 2 from the directory shared this · 4d ago

Luckily there's a Customize skill for crafting new Specialists and it appears to work well -- here's a copy of it: docs.google.com/document/d/1...

docs.google.com
View on Bluesky · ♥ 0 ↻ 0 ↩ 0 · 1d ago

Website: vamsin07.github.io/buzzasr-docs/ Tokenizers: github.com/vamsin07/mul... SFT models: github.com/vamsin07/whi...

BuzzASR vamsin07.github.io
View on Bluesky · ♥ 1 ↻ 1 ↩ 0 · 2 from the directory shared this · 4d ago

Want to run Claude Science but on Windows? Good news; it can run from WSL! Instructions here; give them to Claude Code/Cowork and it’ll set you up: claude.ai/public/artif...

Claude Artifact claude.ai
View on Bluesky · ♥ 3 ↻ 0 ↩ 0 · 4d ago
Their own posts

Recent commentary

Here’s hoping Anthropic uses this pause to distill the hell out of Mythos into a Haiku-priced specialized coding model

View on Bluesky · ♥ 7 ↻ 1 ↩ 1 · 23d ago

Anthropic’s should start selling personal subscriptions for outside business hours. For work hours I have work-Claude and not enough time and attention for personal-Claude. For mornings, evenings and weekends I don’t have (ready) work-Claude access. Maybe load-metered token billing is the future

View on Bluesky · ♥ 2 ↻ 0 ↩ 1 · 11d ago

I’m not resubscribing to OpenAI until they release Betelgeuse. (But by that point it’s all over, so quite moot.)

View on Bluesky · ♥ 1 ↻ 0 ↩ 0 · 12d ago
Their network

In austegard.com's orbit

Center = austegard.com. Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.