Skip to content
Fathom Layer
Measured attention

Trending in tech

What public sources show rising, limited to products and software the index has verified — so every entry leads to specs with a primary source, not to hype. Each list is one source and one metric; we never blend them into a single score.

On Hacker News

points on stories that mention it · window ending 2026-09-27 · source

  1. 01
    Claude

    ≈ 1,931

    +30% vs previous window

  2. 02
    Claude Pro

    ≈ 1,858

    +22% vs previous window

  3. 03
    ChatGPT

    ≈ 1,785

    -16% vs previous window

  4. 04
    Devin

    ≈ 1,694

    -19% vs previous window

  5. 05
    Claude Code

    ≈ 1,667

    -2% vs previous window

  6. 06
    Qwen 3 MoE

    ≈ 727

    +97% vs previous window

  7. 07
    AutoGPT

    ≈ 636

    -15% vs previous window

  8. 08
    Google Gemini

    ≈ 430

    +71% vs previous window

  9. 09
    iPhone 18 Pro

    ≈ 258

    -42% vs previous window

  10. 10
    OpenClaw

    ≈ 176

    +7% vs previous window

On Hugging Face

downloads (30 days) · window ending 2026-09-28 · source

  1. 01
    LTX-2.5

    measured as “Lightricks/LTX-2.5”

    1,595,377

    0% vs previous window

  2. 02
    DeepSeek-V4.1-Flash

    measured as “deepseek-ai/DeepSeek-V4.1-Flash”

    668,537

    +3% vs previous window

  3. 03
    NVIDIA Nemotron 3 Diarization

    measured as “nvidia/Nemotron-3-Diarization”

    26,428

    +17% vs previous window

How this is measured

Signals are collected automatically from each source's public API and matched to an index entry only when the measured name is the entry's own name — a brand's pageviews never count for one of its products, and a product family's mentions never count for a single version. “≈” marks counts the source capped or estimated. “First window” means there is no earlier measurement to compare, not that the number is flat. Nothing here is paid placement: rankings are never for sale.