Yeah trainium and inferentia. They’re just not nearly as well supported on the software level. Google has already made sure this new generation will be supported by vllm, sglang, etc. Amazons chips barely support those and only multiple versions back. Super under invested in (at least on the open source side)

That's seems odd. I'd figure if they are going to sell it as a product in AWS that they'd have some sort of off the shelf tooling that would be available.

Yeah the spy network is just all these people living in a 7 mile radius of each other (San Francisco) and have lots of overlap in social circles

Yeah, I’ve been beta testing the Streamyfin tvOS app. it’s now about on par with the official tvOS client, but feels like it’s making more progress regularly

There definitely is a cap. I believe once you make more than $176k you stop being taxed for that

Yep the cap was $176,100 for 2025

Someone needs to make a market on whether or not this is profitable

What happens when the bot places a huge bet on "No" in this market?

I wanna do this but with a locally running LLM. They’re getting better and better. Can’t wait for something like Taalas to ship custom LLM hardware for personal use

Ok maybe I’m too young, but what is BeOS? Everyone here is linking other alternatives, but no one’s linked to the original BeOS. Or is it gone now?

I don't understand this type of helplessness when you're already competent enough to use HN...

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

"Why did you ask this when you can google it?"

Because sometimes I learn unexpected things and get another perspective even when I could search for it myself.

what perspective could there possibly be on "what was BeOS?" Like, it was an operating system.

stick around...

> Ok maybe I’m too young, but what is BeOS?

Does this help?

https://news.ycombinator.com/item?id=47520510

Yeah it’s gonna make the bar for good enough here super easy to meet and people will have less reasons to look around outside

Yeah, if you had any faith in these private companies to not bend over backwards for the feds, I have a bridge in San Francisco to sell you

No concrete pricing anchors makes this basically useless. Give some estimates for various team sizes even if it’s not exactly what you will charge

Thanks, really fair feedback. Honestly we're still landing on exact tiers as we work with our early customers. The model is per GPU under management, lower at scale. Happy to chat if you want to learn more — founders@usechamber.com

I/m no expert in this area, but I doubt the longer term plan is build a viable business by selling the product. In a prediction market I'd bet that coming out of AWS they saw this problem first hand and acquisition (really acquihire) is the outcome. Based on the way Amazon operates they should also know it's a. hired by someone else, or b. have your producted copied/cloned by AWS.