Recent submissions

Comment · 3 days ago
No, it says it cut through an area of "tens of thousands of acres"
Comment · 3 days ago
It makes for a good headline, but

> the cleared land which stretches more than 13km (8 miles)

Comment · 3 days ago
At least the people responsible for wasting about a quarter billion euros on trains that are too wide for Spains train tunnels seem to have resigned.
Comment · 5 days ago
I prefer to get my supply-chain attacks through conventional package managers
Comment · 6 days ago
I suppose blocking/censoring Facebook/WhatsApp in Brazil would be one way of making sure their opposition wins.
Comment · 6 days ago
That seems a bit schizophrenic
Comment · 8 days ago
Have you tried... reading the article? Under the big headline that says "Limitations", in particular
Comment · 8 days ago
> Hopefully this will send in place a party-shutdown procedure

Let's save democracy by abolishing democracy!

Comment · 10 days ago
Not very shocking that OpenAI wants you to use tokens
Comment · 10 days ago
Gemini offers indefinite cache on API, but you pay per hour.
Comment · 11 days ago
AFAIK, there's no evidence that Alibaba won't keep releasing more open source models. Before 3.8 there was speculation they wouldn't release the weights, but they finally did - Although I guess the largest model wasn't multi modal.

And Minmax is set to release their latest flash model on October 23rd.

So not sure where you're getting this from, unless I misunderstood something

Comment · 12 days ago
I have this idea that the big labs internally have much more capable models which are prohibitively expensive to run though, and the competition is mainly around getting the price down by releasing models that are quantized/otherwise made more efficient at the cost of a lot of performance.
Comment · 13 days ago
30GB a month isn't too terrible? I guess it depends - for example my provider includes 10TB egress per month, so 30gb would be negligible if it's static content. I'm not sure what others do, but I often click on the links that the agents pull up as sources so the websites do get more discoverable.

What really sucks is e.g. Gemini that pulls images without being able to go to the source (unless it changed recently.)

Edit: I didn't read "logs"

Comment · 13 days ago
I'm not sure just skipping the scraping by providing your data directly for harvesting is the solution
Comment · 14 days ago
You sure? HuggingFace should be happy they got their leaked keys reported at pretty much zero cost and no real adversarial breach.

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton