Boo is the only think i can imagine when I hear about Vercel

Seems to be down.

They already died imho.

Lets admit it - previous research loop was going for a google search then iterating on results, which takes time and effort. LLM based research (although could be misleading) is just much faster and more suitable for natural language.

It happened.

This doesn't mean that Google will be obsolete, if it leverages its collected data so far it will be a strong competitor. But Google isn't leading at the moment.

This is a very limited view of the things people use search for. You still need links and a way to browse through the results. And for many queries you need factual up to date information. LLMs don’t do “local supermarket opening times” and “coffee near me” - or the ones that do run a search query anyway and give you back the results.

Except Google invented transformers, which is what all the major LLMs use.

If Google was dead, this stuff would not exist right now

For Google, it has had no trouble developing the technology necessary to be a leader. While it sometimes does not take advantage of the result as well as others, the ability to invent that stuff is critical. It's much easier to become good at product management than good at inventing the future. You need to do both, but I'd bet on a company good at the latter being able to become good at the former well before I'd bet on a company good at product management that can't innovate anything.

For starters, the ones good at inventing the future have less competitors in the future they create. You have some time to try to build the right product. The ones good at products often are competing against hundreds of other companies with the same good idea

The purpose of a company is not to invent technology. It’s to invent products that can make a profit. Google is very bad at producing profitable products outside of adTech.

align these ideas and check out the autism spectrum and get ready for your mind to be blown.

Mind is pretty blown. There really is so much overlap

that’s inspiring.

Microsoft gained a lot of respect back! Great move executives, well done. But I believe in your intentions, keep it right, people like this.

Great post. Long sequences are a problem I have been dealing with long enough now. Even though GRU outperforms LSTM in longer sequences, it still isn't enough and vanishing gradients problem is still experienced. Currently, one-dimensional convolution units are effective while dealing with longer time steps, but sadly it's not the precise unit people wish to use while processing longer time-steps.