dcre

It does not count AI summaries; they break that out explicitly. I don’t think it includes other integrations either, though it’s less clear.

https://news.ycombinator.com/item?id=46395744

I don’t think that’s right, and it’s telling that this is the response every time I mention these numbers. The numbers I’ve seen are Gemini web and mobile app users, which are explicitly distinguished from AI summaries and AI mode in search.

“Google said in October that the Gemini app’s monthly active users swelled to 650 million from 350 million in March. AI Overviews, which uses generative AI to summarize answers to queries, has 2 billion monthly users.”

https://www.cnbc.com/2025/12/20/josh-woodward-google-gemini-...

How do you reconcile the sense that there's a vibe shift with the usage numbers: about a billion weekly users of ChatGPT and Gemini and continuing to grow.

The brand-name ones are surprisingly durable. Can’t speak to the cheaper knockoffs.

The bits that stand out to me are the non-question questions.

“Their headline?”

“Scientists are thriving. The workforce is managing. But creatives?”

“The top trust destroyer?”

This seems very very implausible to me as an explanation of what prompted them to get acquired. They wanted to get rich and stop having to fundraise!

Wow! Didn’t realize.

Who has claimed to have 100x productivity?

Do you have confidence they can execute?

“inserted at the start of any query” feels like a bit of a misunderstanding to me. It plops the skill text into the context when it needs it or when you tell it to. It’s basically like pasting in text or telling it to read a file, except for the bit where it can decide on its own to do it. I’m not sure start, middle, or end of query is meaningful here.

This question doesn’t have anything to do with skills per se, this is just about how different agents handle context. I think right now the main way they cull context is by culling noisy tool call output. Skills are basically saved prompts and shouldn’t be that long, so they would probably not be near the top of the list of things to cull.

Agree, surprisingly weak trailer considering how good the game is likely to be.

They’ve been planning this since October.

Not much of an announcement if you’ve paid attention to anything the company has done in the past two years.

This is what every agentic coding tool does. You can try it yourself right now with the Gemini CLI, OpenCode, or 20 other tools.

I think the idea is not that there are features that aren’t listed, but rather that if a typechecker supports 10 features people care about and is missing 10 that people don’t really use, it will look a lot worse on a list like this than a typechecker with 100% compliance, when in practice it may not really be worse at all.

Edit: Based on this other comment, the point was also about things not covered by the spec. “The spec mostly concerns itself with the semantics of annotations, not diagnostics or inference.” https://news.ycombinator.com/item?id=46296360

Did you read a word past the part of my answer you quoted?

For the reason I gave in my answer: it would be answering based on the text of the book. I don't expect it to be particularly great regardless because these features always use cheap models.

Did you give it the text of the book and tell it to answer based on that?

It's not that direct a counterexample. We have no idea what underlying data from the Fallout show they gave to the model to summarize. Surely it wasn't the scripts of the episodes. The nature of the error makes me think it might have been given stills of the show to analyze visually. In this case we know it is the text of the book.

Nearly everyone else (and every measure) seems to have found 3 a big improvement over 2.5.

The models you interact with through the API (as opposed to chat UIs) are held stable and let you specify reasoning effort, so if you use a client that takes API keys, you might be able to solve both of those problems.

I think you are totally off. Individual benchmarks are not very useful on their own, but as far as I’m aware they all tell the same story of continual progress. I don’t find this surprising since it matches my experience as well.

According to Wikipedia, the IC engine was invented around 1800 and only started to get somewhere in the late 1800s. Sounds like the story doesn’t change.

https://en.wikipedia.org/wiki/Internal_combustion_engine

It's not useless, but the original document is aimed at expert craftspeople and there's a lot of content in the texture of it.

At least you can tell they wrote it themselves — an LLM couldn’t write a sentence that bad.

Thanks, this will be perfect for the team of 6 year olds that work for me.

Feynman at the 1965 Nobel banquet: “Each joy, though transient thrill, repeated in so many places amounts to a considerable sum of human happiness. And, each note of affection released thus one upon another has permitted me to realize a depth of love for my friends and acquaintances, which I had never felt so poignantly before.”

https://www.nobelprize.org/prizes/physics/1965/feynman/speec...

Docs also often don’t have anyone’s name on them, in which case they’re already attributed to an unknown composite author.

This does not work on Bryan, however.