#Oh I don t know possibly Definitely
63 messages · Page 1 of 1 (latest)
the numbers seem crazier the more I look into it
open AI's computer is WAY ahead of things like Andromeda, which is 30+ million and takes 16 racks
wow
Man imagine how much heat those rack put off, jeez
the electricity bill goes crazy
I know right? Must be thousands of dollars to run them
if they are charging over 1 cent per 500 words of davinci, the variable costs of everything are ludicrous
IIRC projected estimates for open Ai is 3 million per month rn
wait they charge based on input and output
no there is a 100% chance open ai is hemorrhaging money rn
Yeah true, they have a lot of investment to spend
they are silicon valley burn money but on steroids
They are trying to gain a market share before others hope on
Especially since open source models come about really quickly after private ones
Just look at Stable Diffusion
not even market share, probably just clear the cost of entry
True, those machines and training times aren't cheap
- they are probably aiming to be aquired
Yeah true, I really despise companies that do that. If they wanted to, they could dominate the next decade and become the next google. It's really disappointing most times
But I guess it's the fastest way to make money with these things for the execs
its just a startup mentality
bc FFANG and the like can afford to eat losses in a way that a smaller entity cant
True
at some point individual investors don't want to keep the ship afloat
something like openAI isnt really their own business
they are a tool for other ones
like googles search engine is a whole company with subdivisions
but openAI is more of one of those subdivisons
its not a parent product, its a child prodcut
wait you started a NFT company... was your long term goal not acquisition by a money whale?
Nah, we render out images, animations, and 3D models for clients. We are all bootstrapped and freelancers
We didn't really even handle the NFT stuff, just did everything in Blender really
huh.. looking at your webpage it looks like your goal would be to be acquired by a large exchange or something like that
or some company that does the full stack
Nope, had some offers but turned them down, wanted to build something of a render farm service. We are small, I started it with my brothers as a side project but it grew
got it. Your plan is to scale into the full service?
altho nfts arent doing too hot rn
Yeah, most of our work is manually done, everything except rendering which is handled by our open source Blender add-on
but I wanted to automate the whole thing so someone could just upload a Blender file and not have to worry about the hassle
Yeah we slowed down development of the site as the market turned red, didn't want to sink to many resources and time into it if the demand wasn't there. If the market turns around we will come back to it
idk NFTs in the crypo bro sense are probably dead imo
We gained a lot of experience with Blender and software development during the NFT hype, which we plan on using in other areas of tech now
Yeah I can see that
like it fails to provide a benifit
I think the tech is interesting and has some applications, but definitely not the crypto bro stuff it used to be
it definitely has a use
it turns digital files into property
but it was ruined by the demographic that used it
but nfts for certificates and other things could be an interesting prospect
Arweave and Filecoin are some real solutions, NFTs could be used to verify art among other things, if they are done in a responsible way. But it needs to be done in a way that prevents scams and fraud. There needs to be a major change in the way these projects are built that prevents catastrophes like FTX and other rugpulls
I was never really a crypto bro, or into that cult, I really just liked pushing GPUs to their limit and seeing what I could make computers do
I mean power to you to hop on a train when u see it
and be able to get off when its crashing
Definitely, it's a bit to complex I feel to make it a public thing
I see the argument for privacy and the ability to keep funds secure from banks in emergencies. But other than fringe stuff, it hasn't yet proven itself in a way that will enable mass adoption. One day it might, I hope to be proven wrong, but as of now it's got a long way to go
idk time will tell
All right, I'm going to get some sleep, hope you have a good day. Feel free to reach out to me on linkedin for a chat or something!
wat