Specr Lab launches: predicting spoken numbers
What happened
On October 14, Hacker News featured Specr Lab, a web experiment by Shaun Pedicini. Users say digits from 1 to 9 out loud, and the page predicts the next one in real time using speech recognition that runs locally in the browser. Ten small predictors run at once, and the one tracking digits that have not appeared for a while contributes the most. Its initial calibration comes from a psychology study in which 418 people each called out 200 numbers. The first 10 digits are a warm-up and do not count toward the score; final results can be compared with 206 people from that study. The speech pack is about 31 MB, and audio never leaves the page.
Written by AI from the coverage · updated 3 hours ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageShow HN: How Random Are You?
Shaun Pedicini 发布网页实验 Specr Lab，让用户从 1 到 9 连续口报数字，页面在浏览器本地运行语音识别并实时预测下一个数字。该工具同时运行十个小预测器，其中"追踪久未出现的数字"贡献最大，并基于一项 418 人各报 200 个数字的心理学研究做初始校准，前 10 个数字为热身不计分，最终结果可与该研究中 206 人对比。语音包约 31 MB，音频不离开页面。
Heat over time
Heat now 9·Comparable peak 9（Oct 14 03:00）·Comparable change over 24 hours –
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.