The future of foundation models is closed-source

&gt; Open-source software started as an act of charity – the world owes the likes of Linus Torvalds and Fabrice Bellard for endowing humanity with Linux, Git, and FFmpeg.<p>So Linux is a simple act of charity from one person. If the author so fundamentally misunderstands what Open Source software is in principle and Linux is in particular then I struggle to see how he&#x27;s qualified to come to such a definitive conclusion on a subject as complex as foundation models.

Considering this statement from the author, buried at the very end:<p>&quot;I admit I have an allergic reaction when many open-source advocates expose their socialist tendencies from Europe, academia, or both. ... America&#x27;s tech success is subject to endless criticism from those who missed out, but we handily won the last tech wave because American capitalism aligns users and companies for the long term.&quot;<p>I&#x27;d take everything else in this article with a huge grain of salt.

nice try but this is clumsy and one-dimensional in the parts that matter.. proof of unsuitability for purpose in this essay is that Mark F&#x27;in Zuckerberg is quoted directly as if that means anything.. (the guy is a liar). Secondly, LLMs are not the only AI application, they are the application that Joe Ordinary is going to get stuck talking to about your insurance premium; thirdly China is at this time arguably better at many AI tasks than anyone.. so a thin claim that China wants to copy something for free and nothing else, shows lack of sophistication.<p>The race is on behind the scenes, and the dollar valuations are proof of it. This essay starts a worthwhile conversation about the nature of the reproducible builds beast, but has little to bring in the way of insights or revelations IMHO

Setting aside the politicking about &quot;Left vs. Right&quot;, the post makes a good argument why Meta will &quot;soon&quot; stop open-sourcing their models.<p>Whether people will eventually give all their data to one of four companies is another matter. For example, it depends a lot on how much further improvements go before we reach diminishing returns and end the current hype cycle. Chips might catch up with how much compute vs. how much (good quality) data you need to build a useful model. I also think it&#x27;s not entirely impossible that state actors might not go for being ruled by one of four US companies.

It’s closed source in one company unless every other company does something about it.