This is Meta we're talking about... They can't even get their own documentation right for developer access to their tools. Then they make you jump through a thousand approvals and testing... and then they want you to pay for the interaction. Won't hold my breath, but prove me wrong @Meta
So if everyone is hoarding GPU/Inference inventory and there's no power, no DC and massive resistance; when do the fire sales start? I'm sure they got good deals on the cards and they're likely worth more than they negotiated. Is anyone going to finally just start liquidating? Get the money to pay back those loans before the next generation completely devalues their current inventory?
There's a very good reason this happens in some companies and its by design. You've all read about the H1B mills... I dug into one in the past... you could literally NOT get a match or scored against opportunities INTERNALLY if you were US based. Guess where the system was built?
The Mission S and Roadster launches seemed familiar to me. AI Hype vs Reality. In one corner we have a system that will do the job its supposed to do, be a supercar vs one we've been waiting for since 2020 that's all flash with features most buyers will never be able to use. We all love a good demo, now make it work.
I get it, but its not ZDR like it was a few weeks ago, its throttled... and keeping it real, you're not going to pump real work through it because you'll burn that $10 or $40 opencode go account up quickly. I'm awaiting my 3 day weekly reset right now. If you wanna do some real work, stick to deepseek API, its cheap, fast, and look at that token throughput. Discounts during US working hours and the weekends.
@GeniusGTX funny, the newly announced version is the Ryzen AI Max Pro 495+ https://t.co/dZeBfJKFjB and we're over here arguing about the 395+ price... give me the new hotness.