Surrounded by the barbarian warmonger nations of Canada and Mexico, the US has to increase its military budget to close to a trillion to pay for wars it definitely hasn’t been single-handedly starting since the 1970s while the USAID budget which is 5% of the military’s which tries to eradicate tuberculosis and malaria needs to go. Makes total sense.
So that’s where all the used V100s are going. I’ve been thinking about making a water cooled version of those because I don’t have a rack to put these servers on.
I built a very crude app sourcing comments from the HN firehose, classifying them using Gemma 4 and a prompt made from the comment guidelines. It had some amusing results.
The app did a decent job at surfacing problematic comments that a mod can do something about.
It was cool to optimize llama-cpp arguments for throughput. During the slightly off-peak hours the post processing was pretty much real time. I suspect a second 3090 would’ve be enough for peak posting hours too.
It could be that polarizing threads end up more actively voted on compared to “cool stuff” threads. We’re not immune to social network effects after all.
It’s the same as posting paywalls articles. There’s no need to make a new rule. If the users can’t read it, the conversation will simply devolve into title commenting.
The idea is interesting and I’ve been thinking about tools along the same lines but Claude’s paws are all over the landing page - non-meaningful icons attached to specific concepts, not-optimized for mobile, the contrasts, etc. The bar is higher and if I determine that the author is using coding assistants without touching up the various elements that the agent spits out, it looks bad for the project.