> a bombshell scientific paper recently cast serious doubt on AI developers’ ability to continue to dramatically improve their models’ performance.
This is the problem with this type of (AI isn't real, I promise!) writing. Each and every single time. This bombshell scientific result that he links to? Is Gary Marcus lamenting about a single paper, specific to a single architecture, which primarily shows how quickly algorithmic improvements are bringing down energy costs[0].
We get it. You don't like AI. But you don't have to lie about it, unless of course you are Gary Marcus and this is how you talk your book and earn your speaking fees.
This is the problem with this type of (AI isn't real, I promise!) writing. Each and every single time. This bombshell scientific result that he links to? Is Gary Marcus lamenting about a single paper, specific to a single architecture, which primarily shows how quickly algorithmic improvements are bringing down energy costs[0].
We get it. You don't like AI. But you don't have to lie about it, unless of course you are Gary Marcus and this is how you talk your book and earn your speaking fees.
[0]: Algorithmic progress in language models. https://arxiv.org/abs/2403.05812