kasmura

This is THE book. It is suitable for beginners with basic math background equivalent to CS undergrad I’d say

I disagree. Ring attention and tree attention are so general that the core ideas are independent of details about modern GPUs. Maybe so for flash attention but not this. I also disagree because these algorithms are fundamentally about enabling long context by distributing across gpus and this would not be enabled by “moored law for gpu hardware”

Yes, it is just a way of computing the self-attention in a distributed way

That cannot be done when using OpenAI API calls as far as I know

Integrated information theory which is mentioned in the article is abstract and supposedly applies to any type of physical system.

Is the following quote at odds with what you are saying about 50-way classification?

"On the other hand, the network is not merely classifying sentences, since performance is improved by augmenting the training set even with sentences not contained in the testing set (Fig. 3a,b). This result is critical: it implies that the network has learned to identify words, not just sentences, from ECoG data, and therefore that generalization to decoding of novel sentences is possible."

I am not much of an expert either, but as I understand it Zermelo-Frankel set theory is the most common foundation of mathematics and it is stronger than Peano arithmetic so I think Gödel's results hold in general.

Internet Access is a NEGATIVE right

I am excited about this. Competition in the market is good for innovation and the price.

Pedro Teixeira has made a screencast about Coffeescript in Node.js on his website notetuts.com: http://nodetuts.com/tutorials/16-introduction-to-coffeescrip...

This could be cool for computer science only too.

... in Spain

javascript:document.write('<img src="http://lilqr.com/qr>)

;)

I first thought that this was something only implemented in Chrome for Google, but I then saw that this is somehow a HTML feature that they will also add to Google Search.

I don't like if Google uses its own browser to give their sites features, that could not be implemented without - website's features should only be made from source code on the server. In my opinion.

This is core innovation from Google - but I see the problem with incorrect website visitors. It think Chrome when prefetching the pages, should send some information with the request that the request is a prefetch.

A cool option for mobile browsers, would be that an online service, could convert JPEG files into WebP. Then the mobile browser could be downloading the WebP file instead - that way saving money.