vinni2

I am from Europe and managed to try Waymo in SF recently. Not sure why you couldn’t.

I guess fundamental difference is that Waymo uses a lot of sensors while Tesla FSD is relying on AI and computer vision.

Ggml Q8 models on ollama can run on much cheaper hardware without losing much performance.

When you remove the quantifier millions of Canadians it is misleading and implies all Canadians!

astrophage

Which is why you should ask people to exclude self citations when reporting h-index. First of all Google scholar should fix this.

Just curious what does a Python team do? It doesn’t make sense to dedicate a team to a programming language.

Money doesn’t grow on trees but grows as shrubs

There are people who strongly believe data is a non-blocker possibility due to synthetic data that is of high quality.

Dario Amodei from Anthropic. https://www.dwarkeshpatel.com/p/will-scaling-work

So using LLMs to fact-check LLMs with Google results? There are non-LLM fact-checkers out there [1] which aggregate search results from more diverse sources and which are shown to outperform LLMs [2].

[1] https://editor.factiverse.ai [2] https://arxiv.org/abs/2402.12147

Well but they don’t mention that it is clickbait to call it open source fact checking tool which needs LLMs to do everything. Also code is not designed to easily swap with a free locally running LLM.

Ok but bigger issue is there is evidence that the LLMs are not better than specialized models for fact-checking. https://arxiv.org/abs/2402.12147

So only thing they open sourced is the prompts [1] and code to call LLM APIs? There are plenty of such libraries out there. And the prompts seem to be copied from here [2]?

[1] https://github.com/Libr-AI/OpenFactVerification/blob/main/fa...

[2] https://github.com/yuxiaw/Factcheck-GPT/blob/main/src/utils/...

It’s a bit misleading to call it open source tool when it relies on proprietary LLMs for everything.

Sounds like you are an outlier.

I also wonder if in future search engines would rank pages containing AI generated images or content lower.

Academic papers are peer reviewed white papers aren’t.

This paper is just accepted in NAACL which is a top NLP conference.

I realized this the hard way after about 5 of my se cards were fried.

Something tells it is going to disappoint the fans of the books.

Dupe and flagged https://news.ycombinator.com/item?id=39749647

This is a bit of a narrow view imho and also comparing to Bangalore is not the optimal thing to do.

There are more efficient denser urban areas which utilize resources more efficiently than sparse suburbia.

Also there is a middle ground between the stringent zoning laws of the Bay Area and Bangalore.

Reminds me of a House episode.

I would say this has a cascade affect and big well funded universities are displacing smaller institutions.

Thanks. I am struggling a bit to login and download the model during GH CI pipeline. I keep getting invalid token error but it works when run locally. Any ideas? Sorry to post this question here.

Dupe https://news.ycombinator.com/item?id=39558015

I get that, but the model needs to be downloaded at least once? Do you download the models as part of the Dockerfile, then?

Thanks for this. Can you explain a bit more? is it like it like it downloads a once and caches or do you need to provide the pre download model s in a volume?

> I don't think recovering 97% vs 95% of 8% of the material is going to change the economics of battery recyling.

Not yet but as more and more EVs reach end of life and these rare earth metals become rarer then it starts make a difference.

That's on hearing aids not sure if this finding is transferrable to smartphones. Here is the recommendation from experts in fixing electronics https://www.ifixit.com/Wiki/Do_Not_Put_Your_Device_in_Rice