Darren LI

Many of these projects are saving time by training on small, highly curated datasets. This suggests there is some flexibility in data scaling laws. The existence of such datasets follows from the line of thinking in Data Doesn't Do What You Think, and they are rapidly becoming the standard way to do training outside Google

In streaming settings, StreamingLLM outperforms the sliding window recomputation baseline by up to 22.2x speedup.

It’s both a little awkward and entirely inevitable that businesses like Think Expansion would emerge on OnlyFans. Awkward because OnlyFans markets itself as providing the infrastructure for authentic, personal connections between creators and their fans. Inevitable because platforms like OnlyFans naturally encourage businesses to scale up,... See more

These companies follow a playbook originally written by Amazon: serve the needs of users within a single high-frequency purchase category, and use that category as a beachhead from which to expand vertically (brands) and horizontally (categories).

1) AI换全身产品Viggle,曾经制作出多个具有爆款属性的视频,被传播的原因主要是搞笑、AI技术自带流量,2) AI生成完整可用视频InVideo,输入文本,AI输出讲述性质的视频,适合效果类似于“巫师财经”这种文字解说+对应视频片段的内容,AI生成后支持进一步的编辑,看起来更接近直接可用,3) 视频切片工具Opus Clip,使用非常简洁,生成具备爆款潜质的短视频