Eyal Toledano
@EyalToledano • 18,294 subscribers
invented plan mode • leading @usehamster, the ai native workspace for product teams • creator @taskmasterai
Videos
just shipped taskmaster v0.18 🚀 → Claude Code Community provider without API keys → init with IDE-specific profiles for Cursor Claude Code Community @windsurf Cline Visual Studio Code TRAE → cleaner git repos & Python support → PROVIDER_BASE_URL follow + bookmark + dive in 👀👇
Eyal Toledano85,994 次观看 • 1 年前
just shipped taskmaster v0.16 🚀 → .taskmaster folder → bedrock, @azure & Google AI vertex → TASK_MASTER_PROJECT_ROOT & BaseURL → semantic context for add-task → README <> task list sync + welcoming legend Jason Zhou to @taskmasterai follow + bookmark + dive in 👀👇
Eyal Toledano69,520 次观看 • 1 年前
just shipped taskmaster v0.12 🚀 → parse multiple PRDs, smarter breakdowns, easier setup, command improvements,MCP polish +++ → introducing the npm i task-master-ai website, which i have become ossessed with (link below) → 4.5k ⭐ GitHub big upgrades based on your feedback 👇
Eyal Toledano72,354 次观看 • 1 年前
"which quant should I download?" is a question you may never have to answer again the team Hamster Labs has figured out how to kill it with pMLX. download once at full precision (bf16) and the engine re-fits it to your machine on the fly, based on the job you give it tell it two things: how much context you need, and the slowest speed you'll accept. it reads your mac and picks the quantization plus how many experts stay in ram vs. stream from disk. no need to download a smaller quant or figure out which quant fits your hardware. same model for every use case and the config changes based on what you need I ran this on my own M4 Max and Qwen3.8-Flash-Next splits into two configs. short context, under 32k: full bf16, ~85% of experts in ram, fast at full precision since it fits my ram long context, 64k to 256k: keep ~70% bf16 and stream the rest from SSD at 22 tok/s, or drop to q8 and get 40 tok/s. I pick per task and the model itself never changes. there are many possibilities since I have the ram to spar - if i need speed, Q4 100% resident (74gb ram @ 62 tok/s) - if i need balance, Q8 95% resident (76gb @ 38 tok/s) - if i need accuracy, bf16 70% resident (96gb @ 16 tok/s) given whatever RAM you have (16/32/64/128/256 GB) + your context + your min speed, the engine picks the precision (bf16→q8→q4) and the expert-residency/paging split that fits your box and maximizes quality & speed last thing to optimize is speed. there are so many things we want to power with open models at Hamster and these 180b-300b class models have the potential to play a big role in that
Eyal Toledano10,743 次观看 • 1 个月前
Qwen3.8-Flash-Next now fits on a 48gb macbook air Got q4 running on MLX using only 39gb of memory and decoding at 28 tok/s 🔥 Also q8 running on just 50gb of memory I stacked two tricks so far: 1. REAP pruning cut the expert pool from 68gb to 35gb 2. the 51b n-gram table now lives on the ssd instead of ram My patch memmaps the shards and reads ~100 bytes per lookup straight off disk, dequantized on cpu and verified bit-exact against the gpu path Qwen3.8-Flash-Next at 288 experts with its PLE table on nvme loads in 6 seconds, 39gb resident, 28 tok/s decode, 600 tok/s prefill. Logits identical to the ram-resident version in every way 97gb stock Q4 down to 68gb if pruned to 288 experts down to 39gb if pruned to 288 experts with the table on disk humaneval scoring 91.5% with 288 experts vs. 93.9% for stock I'm not done. next: 1. MTP drafter for speculative decode 2. n-gram lookup drafter both are built, benchmarking them now
Eyal Toledano10,538 次观看 • 1 个月前
we've shipped taskmaster v0.20 🚀 → mcp sampling support (zero api keys) → Gemini CLI integration & provider → advanced Claude Code Community rules → language override → Grok 4 support → Groq Inc support → parse-prd auto-selects task number & more follow + bookmark + vibe 👀👇
Eyal Toledano48,608 次观看 • 1 年前
we've shipped taskmaster v0.23 🚀 → official npm i task-master-ai Visual Studio Code extension → manage your taskmaster workflow visually → introducing scope-up & scope-down tools → dozens of QOL improvements & fixes → we're hiring 🐹 follow + retweet + bookmark 👀👇
Eyal Toledano43,375 次观看 • 1 年前
just recorded a 10min video that shows how to use @taskmasterai on existing projects you'll learn how to: - weaponize Cursor for code context - quickly create PRDs for refactors/improvements - create, research & expand the new tasks bookmark, retweet & vibe away friends
Eyal Toledano43,716 次观看 • 1 年前
recorded a 42 min video explaining what npm i task-master-ai is and how to get started with it. didn't get to write code yet so part 2 is coming i have no idea if you'll find this useful, but if you do, lmk and i'll continue and possibly turn it into a series open to all feedback!
Eyal Toledano25,980 次观看 • 1 年前
trying a TDD + git workflow npm i task-master-ai → test-driven development → github branch for each task → commit + test each subtask once done → push PR when task is feature complete → full code coverage → nearly autopilot will polish and ship as rules first then as a feature
Eyal Toledano14,586 次观看 • 1 年前