Andrei 𒅴𒂠 @andreidro

chief genesis officer @anara andrei.art Cluj Joined September 2016

Tweets

989
Followers

691
Following

794
Likes

3K

Andrei 𒅴𒂠 @andreidro

4 days ago

@flaviuscj1 romanian model made in apartment

0 0 1 18 0

View Details

leaked weights for le chaton fat - 35B active 10T params a romanian stole it for real... divine inspired french, european union compliant masterpiece trained in french only, but if you try to talk in french will respond in english (native human like interactions) beats fable 5

3 0 10 610 0

View Details

Andrei 𒅴𒂠 @andreidro

5 days ago

huggingface.co/andreidro/le-c…

1 0 0 58 0

View Details

Mihai-Cristian @mihaicristianf

6 days ago

- i always loved the Wallpaper Engine app on Windows - saw there was no real equivalent to it for macOS - one month later Supernova was born supports video, shader and web-based wallpapers, plus it's cursor-aware it's motion-aware shaders run on the gpu via metal check it out>

5 6 15 686 2

View Details

Andrei 𒅴𒂠 @andreidro

a week ago

@flaviuscj1 real

0 0 1 84 0

View Details

Andrei 𒅴𒂠 @andreidro

3 weeks ago

@PiticVic drop

1 0 1 60 0

View Details

Andrei 𒅴𒂠 @andreidro

4 weeks ago

trying to get --enable-dp-attention too, was crashing with deepseek and EAGLE

0 0 0 38 0

View Details

Andrei 𒅴𒂠 @andreidro

4 weeks ago

using sglang python -m sglang.launch_server --host 0.0.0.0 --port 8000 --model-path deepseek-ai/DeepSeek-V4-Flash --tp-size 4 --dp-size 2 --config /root/config.yaml --enable-mixed-chunk yaml config (config_deepseek_v4.yaml): host: 0.0.0.0 log-level: info tool-call-parser: deepseekv4 reasoning-parser: deepseek-v4 trust-remote-code: true # Memory mem-fraction-static: 0.92 context-length: 1048576 # 1M chunked-prefill-size: 8192 # MoE moe-runner-backend: flashinfer_mxfp4 # Observability enable-metrics: true enable-cache-report: true # Batching — 8x B200 max-running-requests: 64 cuda-graph-max-bs: 128 # EAGLE 3/1/4 speculative decoding speculative-algorithm: EAGLE speculative-num-steps: 3 speculative-eagle-topk: 1 speculative-num-draft-tokens: 4 # Scheduling schedule-policy: lpm num-continuous-decode-steps: 4 # Kernel fusions enable-mixed-chunk: true enable-fused-qk-norm-rope: true disable-flashinfer-autotune: true env vars (in modal container): SGLANG_ENABLE_SPEC_V2=1 SGLANG_ENABLE_THINKING=1 SGLANG_JIT_DEEPGEMM_PRECOMPILE=0 HF_HOME=/cache/huggingface HF_XET_HIGH_PERFORMANCE=1 HF_HUB_ENABLE_HF_TRANSFER=1

1 1 1 109 1

View Details

Andrei 𒅴𒂠 @andreidro

4 weeks ago

started self-hosting models, my Arrakis, my Dune deepseek-v4-flash on modal 8xB200 180 TPS

1 0 1 90 0

View Details

Andrei 𒅴𒂠 @andreidro

4 weeks ago

benchmaxxed

Logan Kilpatrick @OfficialLoganK

a month ago

Gemini 3.5 Flash has made huge progress from 3.1 Pro on GDPval, Flash is competing at the frontier, post training going strong :)

128 41 968 106K 66

0 0 1 90 0

View Details

Andrei 𒅴𒂠 @andreidro

a month ago

3.1 pro is utterly horrible for any agentic use case and everyone knows it

News from Google @NewsFromGoogle

a month ago

New stat from @vercel's AI Gateway in @BusinessInsider: Gemini 3 Flash is leading across AI models in token usage as of April. 🚀 See more stats on how developers are using our models → goo.gle/4dlBiol 📊via @BusinessInsider

16 25 211 53K 21

0 0 4 152 0

View Details

Andrei 𒅴𒂠 @andreidro

a month ago

@sebisemeniuc top

0 0 0 77 0

View Details

Andrei 𒅴𒂠 @andreidro

a month ago

anara mogging - last 7 days bro

dax @thdxr

2 months ago

our teams last 7 days of spend damn gpt5.5

271 80 3K 724K 347

1 0 10 207 0

View Details

Andrei 𒅴𒂠 @andreidro

a month ago

@sebisemeniuc what models

1 0 0 61 0

View Details

Octavian Cristea @octavicristea

2 months ago

I made a game where Sam Altman calls your pitch "pre-seed energy" and Peter Thiel offers you a term sheet AI rewrites the whole story every play, pulled from real SF news and stories built in couple of days for #ElevenHacks @elevenlabs and @zeddotdev hack #6

6 2 14 1K 1

View Details

Andrei 𒅴𒂠 @andreidro

2 months ago

@BogdanGruiaIvan we goon now

0 0 0 120 0

View Details

Sebastian Semeniuc @sebisemeniuc

3 months ago

I really liked wispr flow but didn’t feel like paying $12/month for it, so i built a local-only open source version it’s a native macos dictation app that runs whisper fully on-device and types directly into any app if you use dictation, this is a simple alternative that just works, fully local you can download the dmg and use it right away