Darren LI

Generative AI’s first year out the gate—“Act 1”—came from the technology-out . We discovered a new “hammer”—foundation models—and unleashed a wave of novelty apps that were lightweight demonstrations of cool new technology.

So in reality they aggregate three sides. On Facebook:- Content makers like newspapers (content supply) bring the quality content expecting in exchange clicks to visit their sites and monetize through ads or subscriptions. - Consumers (content demand) go to the site to consume that content. Some of them can also become content supply some of the... See more

Based on taps, clicks, and sales, an algorithm doubles the production quota, which is immediately updated on the factory floor with extra materials automatically ordered. A second algorithm updates the weightings and recommends the item to more users with similar profiles.

6. Same trend in lending platforms. Aave and DeFi collectively command ~90% of capital moved through lending platforms.

具体到Bolt,因为单笔金额灵活、数量会比较多,我告诉LP,可以把Bolt看成一个一揽子计划(basket),最终的总体投资额(包括后续加码)会在2000万到5000万美元。这些项目中最终能几个有在contextual data、personal data、collaborative data这三个方面实现积累并且与时俱进的团队,最终的收益也不会差。而且, 未来应用公司创造的总价值会远超过大模型。

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

新的序列建模方法,可以学习大型视觉模型,无需使用任何语言数据。通过定义“视觉句子”格式,将原始图像和视频以及注释数据源表示为序列,然后通过训练模型来预测下一个标记,从而实现模型的训练