[fcanesin]

OpenAI knows that everyday someone uses Gemini their ChatGPT brand dies a bit faster. Wonder what Google has in storage for I/O now in May, would be a death sentence to just steamroll with Gemini-3.

[xnx]

> Wonder what Google has in storage for I/O now in May

"Gemini 2.5 Pro Preview (I/O edition)" https://news.ycombinator.com/item?id=43906018

[fcanesin]

LMAO, like one hour after. And guess what, it is a coding upgrade .

[qoez]

Google IO in may? Guess we'll be getting a huge OpenAI release May 19th then.

Edit: Oh of course, it's the open weights model they've been teasing.

[andai]

They launched a new version of Gemini 2.5 Pro today.

https://developers.googleblog.com/en/gemini-2-5-pro-io-impro...

[ukuina]

So soon after Gemini 2.5?

[casey2]

Open AI needed to spend $3B pivoting away from bigdata based AI. But instead they went for the most shorted sighted move possible of snapping up the "trendiest" company nobody has ever heard outside the Ycombinator echo chamber.

Typical VI-fallacy BS. If LLMs were actually good they would replace IDEs completely not be integrated.

[frabcus]

There's an in between case, where LLMs are useful and give coders a (say) 20% speedup, and everyone has to use them. They don't have to be perfect to be a big industry!