ยท
AI & ML interests
Work with agentic systems, agent memory, model fine tuning and unlearning, generative AI, and the development and deployment of open source AI
Recent Activity
reacted to Mati83moni's post with ๐ 2 days ago Kos: a Polish voice assistant that's allowed to say "I'm not sure" ๐ฆ
Kos is a free Polish voice assistant running on ZeroGPU: Silero VAD โ Whisper large-v3 โ PLLuM-12B (EXL3) โ TTS. It's a cascade, not a native speech-to-speech model.
The interesting part is what happens between "the model wrote a sentence" and "the sentence is spoken":
โข math, units, exchange rates (NBP) and weather are computed by code, not by the LLM
โข facts go to Polish Wikipedia/Wikidata first, with sources you can check
โข free-chat answers are buffered and rejected if they contain an unsourced number, a fake citation or a false "I have no internet"
โข a rejected answer is never spoken: Kos looks it up or says it isn't sure
Voice: my own cloned voice on Cartesia. When its daily/monthly limit runs out, Kos switches automatically to Chatterbox, then Piper. It still answers, it just sounds different.
Limits: it's a 12B model, so small talk can still be wrong. Polish only. ZeroGPU quota is shared by all visitors.
Try it (allow the mic, speak Polish, headphones work best):
https://huggingface.co/spaces/Mati83moni/kos-pl
If it answers something wrong, tell me in the Space's Community tab: what you said, what Kos replied. Those cases become regression tests. reacted to savoji's post with ๐ 2 days ago Openly available models now rank near the frontier on Artificial Analysis (AA). This week's most discussed agent work treats trust as something to engineer and verify, not assume.
- Kimi K3.1 Pro: multimodal MoE, 2.8T total / 104B active params, 1M-token context. It is 4th on the AA Intelligence Index (55, 11 points over K3) and scores 64% on the AA Cyber Index, #1 among openly available models. Its weights had not been published when it was scored.
- DeepSeek-V4.1-Flash-1017: an MIT-licensed minor update that cuts output length and task execution time. AA had named it #1 openly available on the Cyber Index before K3.1 Pro arrived. 85,275 downloads last month. https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017
- OpenAI removed live internet access for all frontier models during training and gave them an offline, cached copy of the web instead.
- NVIDIA OpenShell (Apache-2.0, Rust, 0.1.x): one sandbox per agent, with kernel-enforced policy on files, system calls and network. Agents never see real credentials. Formal verification holds risky policy changes for human review. Anthropic argues LLMs make this kind of verification feasible at much larger scale.
- LGWM (173M params at inference) predicts the next screen's representation and compares it with the one observed. It reaches 0.987 AUC on RSWT-Bench at 17 ms per decision on one A100, against 0.995 for Gemini 3.7 Flash and 0.988 for GPT-5.6 Luna. The benchmark is the authors' own. https://huggingface.co/papers/2610.11942
Shortlist 1017 for agent work, or K3.1 Pro once its weights ship. Treat containment as a separate decision: sandbox the agent, keep real credentials out of its reach, and verify each transition, not just each screen.
โข Medium: https://medium.com/p/4ce5f13afba9
โข Substack: https://sajjadpakdamansavoji.substack.com/p/kimi-k31-pro-and-deepseek-v41-flash
โข X: https://x.com/saavoji/status/2114480628121100394
โข Website: https://sajjadpsavoji.github.io/blog#kimi-k3-1-pro View all activity Organizations