Hackernews posts about Gemma3

1. 01
   

   newOpen-jev: One-pass option scoring with Gemma 3 4B, similar to jevgithub.com

   ▲ 7suriyaG18d💬 1savesaved
2. 02
   

   newShow HN: Gemma 3 4B as a typed decision function in Rust (47 ms/decision)github.com

   ▲ 1zozo123-IB-IL212ddiscusssavesaved
3. 03
   

   newShow HN: Breadcrumb, record everything on your mac + context manager for AIinnerloop.works

   ▲ 49jv222223d💬 9savesaved
4. 04
   

   newShow HN: Phntm-ONE: I built a local AI desk assistantwww.phntmcore.com

   ▲ 5phntmcore29d💬 1savesaved
5. 05
   

   newShow HN: Determinstic LLM inference for lowest price Gemma 4, with Windows XPwww.tokendelivery.ai

   ▲ 9carsonpoole22d💬 1savesaved
6. 06
   

   newGemma 4 26B on a 24 GB Mac Mini M4: from 66% CPU to 100% GPUmac-mini-m4-doc.masterfabric.co

   ▲ 4gurkandev21d💬 4savesaved
7. 07
   

   newGemma 4 E4B running inside Obsidian via WebGPU and Karpathy-style LLM Wikigithub.com

   ▲ 4itsyuimorii25d💬 1savesaved
8. 08
   

   newGevva0 – a Jev like decision engine on Gemma 26B via direct logit scoringgithub.com

   ▲ 3solvingSteve6ddiscusssavesaved
9. 09
   

   newShow HN: Implementing Embedding Gemma in PyTorchwww.youtube.com

   ▲ 3prasoon2124ddiscusssavesaved
10. 10
    

    new640tok/s on Gemma 4 with a megakernelwithhopper.com

    ▲ 2jashwanthp125ddiscusssavesaved
11. 11
    

    newShow HN: Gemma – NYC Apartment Matches over iMessagegemma.nyc

    ▲ 1danielalter2ddiscusssavesaved
12. 12
    

    newShow HN: I made a computer vision tool for evaluating deadlift formgithub.com

    ▲ 8dr_blueberry3d💬 1savesaved
13. 13
    

    newWhy human syntax breaks LLMs (and how to fix agentic coding)

    ▲ 5aslang26d💬 3savesaved
14. 14
    

    newShow HN: OOMU is a free macOS native desktop AI harness (local and cloud)oomu.ai

    ▲ 2jeff-oomu9d💬 1savesaved
15. 15
    

    newShow HN: Shlexball, describe a file and a local 1.5B model finds it on your Macshlexball.com

    ▲ 1porkychan12hdiscusssavesaved
16. 16
    

    newShow HN: 9.8M product opinions collected from Reddit / HN, on 48,837 productswww.trustmebro.reviews

    ▲ 1davidzweig3ddiscusssavesaved
17. 17
    

    newShow HN: PicoLM v1.0-rc2 ("Yura Kana"). Run an LLM on Digital Unixgithub.com

    ▲ 1gabucino11ddiscusssavesaved
18. 18
    

    newChasing Speed of Light on TPU v6ewww.sailresearch.com

    ▲ 2simonpure14ddiscusssavesaved
19. 19
    

    newGemma 3 270M: Compact model for hyper-efficient AIdevelopers.googleblog.com

    ▲ 824meetpateltech416d💬 317savesaved
20. 20
    

    newGemma 3 QAT Models: Bringing AI to Consumer GPUsdevelopers.googleblog.com

    ▲ 602emrah532d💬 276savesaved
21. 21
    

    newGemma 3 Technical Report [pdf]storage.googleapis.com

    ▲ 488meetpateltech572d💬 245savesaved
22. 22
    

    newGemma 3 270M re-implemented in pure PyTorch for local tinkeringgithub.com

    ▲ 417ModelForge410d💬 57savesaved
23. 23
    

    newGemma3 – The current strongest model that fits on a single GPUollama.com

    ▲ 252brylie572d💬 138savesaved
24. 24
    

    newFine-tune Google's Gemma 3unsloth.ai

    ▲ 226tomdekan564d💬 78savesaved
25. 25
    

    newGemma3 Function Callingai.google.dev

    ▲ 202canyon289561d💬 32savesaved
26. 26
    

    newGoogle calls Gemma 3 the most powerful AI model you can run on one GPUwww.theverge.com

    ▲ 127gmays563d💬 100savesaved
27. 27
    

    newI have written gemma3 inference in pure Cgithub.com

    ▲ 65robitec97251d💬 23savesaved
28. 28
    

    newLandrecords – cheap nationwide parcel dataset standardized using gemma3landrecords.us

    ▲ 12mapsperson411d💬 5savesaved
29. 29
    

    newGemma 3 QAT Models: Bringing AI to Consumer GPUsdevelopers.googleblog.com

    ▲ 8meetpateltech534d💬 1savesaved
30. 30
    

    newGemma 3: The most capable model you can run on a single GPU or TPUblog.google

    ▲ 8meetpateltech572ddiscusssavesaved
31. 31
    

    newShow HN: Gemma 3 inference in pure C++ with Metal accelerationgithub.com

    ▲ 6ybubnov92d💬 2savesaved
32. 32
    

    newI Gave Up Trying to Run Gemma 3 27B on AWS P3dn.24xlargemedium.com

    ▲ 6live_alone419d💬 1savesaved
33. 33
    

    newGemma 3 QAT (Quantized Aware Training) 3x less memoryhuggingface.co

    ▲ 5philschmidxxx549d💬 2savesaved
34. 34
    

    newOwn your AI: Learn how to fine-tune Gemma 3 270M and run it on-devicedevelopers.googleblog.com

    ▲ 5simonpure359ddiscusssavesaved
35. 35
    

    newGemma 3 QAT Modelssimonwillison.net

    ▲ 4mdp2021533d💬 3savesaved
36. 36
    

    newGemma 3: Google's new multimodal modelsai.google.dev

    ▲ 4danielhanchen572d💬 2savesaved
37. 37
    

    newGoogle's Gemma 3 is an open source, multimodal single-GPU AI 128K context windowarstechnica.com

    ▲ 4danboarder571d💬 1savesaved
38. 38
    

    newGemma 3-270Mhuggingface.co

    ▲ 4georgehill416ddiscusssavesaved
39. 39
    

    newQAT optimized Gemma 3 slash VRAM needs 54GB –> 14.1GB while maintaining qualityold.reddit.com

    ▲ 4tosh534ddiscusssavesaved
40. 40
    

    newSelectively reducing eval awareness and murder in Gemma 3 27B via steeringwww.lesswrong.com

    ▲ 3gmays206ddiscusssavesaved
41. 41
    

    newYou can fine-tune Gemma3-270M and prepare for secure deployment within minuteswww.youtube.com

    ▲ 3astro_09407ddiscusssavesaved
42. 42
    

    newGemma explained: What's new in Gemma 3developers.googleblog.com

    ▲ 3simonpure521ddiscusssavesaved
43. 43
    

    newGemma 3 QAT Models: Bringing AI to Consumer GPUsdevelopers.googleblog.com

    ▲ 3npalli533ddiscusssavesaved
44. 44
    

    newTry Gemma3 Free on Heregemma3.co

    ▲ 3sterlingjone565ddiscusssavesaved
45. 45
    

    newGemma 3 270M – Google's NEW AI – How to Fin-tune Gemma3 [video]www.youtube.com

    ▲ 2rjn32s415d💬 1savesaved
46. 46
    

    newShow HN: Video Summarization Using Local Gemma3github.com

    ▲ 2RamboRogers510d💬 1savesaved
47. 47
    

    newGemma3n architecture: a short guide [slides]drive.google.com

    ▲ 2perone331ddiscusssavesaved
48. 48
    

    newTiny gemma3 270M implementation in Rgithub.com

    ▲ 2dfalbel405ddiscusssavesaved

· edit