Is HN the last refuge of real internet discourse? While some of the replies are very Reddit m-coded, it still feels like real humans. I wonder how long this will last before the agents take over.
We just dropped our H100 on-demand prices to 1.99/hr. It's now the lowest in the world, you don't need to talk to a sales person, and it's available now: https://t.co/AcqRIITS1s
Great summary of the industry. I do want to point out that we actually have the lowest price for H100s: 1.89/hr. Coreweave’s pricing requires long term commitment and you’re comparing our on demand PCIe cloud to their ‘reserved’ h100 cloud. Our reserved H100 + InfiniBand cloud is priced as follows and is the lowest in the world:
Lambda A100s - $1.10 / hr
Paperspace A100s - $3.09 / hr
Genesis A100s - no A100s but their 3090 (1/2 the speed of 100) is - $1.30 / hr for half the speed
For those affected and wanting to run your stable diffusion notebook more, you can always spin up a notebook on Lambda cloud with A100s for only $1.10/hr. PyTorch, TensorFlow, and Jupyter notebooks are pre-installed:
I know it sounds obvious. But if you want to talk to users, just build a little feedback mechanism to talk with users.
Write 10-30 random prompts: “what do you hate about this page”, “what do you love about this app”, “what do you wish you could do”, etc. show them the prompt and log a free form response. Simple as. We’re talking a few hours of work.
Also, find your most engaged users via your analytics and email them to do a video interview to understand how they use the product. Try to find overlaps between them that you can use to target new customers.
Nice tutorial, we worked with the Razer BIOS team to find the touchpad problem and fix it. And we use the 5.15 kernel to fix a handful of the other hardware issues when running Linux on the Razer hardware.
We can run the battery issue / suggestion by the Razer BIOS team.
The target market for this is people who don't want to have to manage a Linux machine learning environment. If you want to least expensive access to GPU compute, you should design a workstation, purchase all of the parts yourself on newegg or amazon, personally take the risk of a part failing, and fully support yourself. That is probably the least expensive way to get a computer in 2022.
This is definitely targeted to people who place a premium on their own time and effort.