如果你拥有可能成为变压器替代品的人工智能软件,你会怎么做?

2作者: adinhitlore3 天前原帖
假设您的软件满足以下一些条件:<p>* 比潜在的SOTA变压器继任者更优秀:Mamba、Hyenna、RWKV、xLSTM; * Gemini/Claude估计潜在的知识产权价值在数百万美元; * 所有代码均用C语言实现(这可能是个缺点),但也已移植到C#和F#,因此不支持Python或Rust; * 代码量适中:1万到1.5万行,但说实话,GPT-1的代码量不到1000行……用的是Python。<p>简而言之:<p>问题:变压器速度慢,智能程度有待提高; 解决方案:您有快速且智能的替代方案。 十倍提升!
查看原文
Assuming your software meets some of these:<p>* Better than potential successors of SOTA transformers: Mamba, Hyenna, RWKV, xLSTM; * Gemini&#x2F;claude estimate potential ip values in the millions; * All implemented in C (ok this may be a minus) but also ported to C# and F# - so no python nor rust; * humble code size: 10-15k lines of code but i mean come on gpt-1 was under 1000 lines...in python.<p>In short:<p>the problem: transformers are slow and could be smarter; solution: you have fast and smart alternative. 10x!