Why the fuck do so many corporate garages deny you service just because you already bought the part you need at autozone
Like my brakepads just need fucking changed bro its not hard I just dont have a jack that can fit under my car bc its fucking lowered
I’m the CEO of a hot dog company. I’ve worked on hot dogs for 10 years. And *I* wasn’t prepared for what I’ve just seen. Your life is about to change.
So what can you do?
Buy as many hot dogs as you can. Buy stock in hot dog companies.
Sam just posted the corporate equivalent of “we’re fine, everything’s fine” while Reuters cites eight sources saying OpenAI has been actively shopping for Nvidia alternatives since last year. His own staff blamed Nvidia’s GPUs for Codex performance issues. Three days ago on a press call, Sam himself said customers “put a big premium on speed for coding work” and that Cerebras would help meet that demand.
The reason this matters: inference is now two-thirds of all AI compute spending. Training made Nvidia untouchable. Inference is a different game. GPUs rely on external memory (HBM), which creates a latency bottleneck every time a chatbot fetches data to generate a response. For something like Codex, where users need code generated fast enough to feel like pair programming, that bottleneck becomes a product problem.
OpenAI went to Groq to solve it. Groq’s LPU architecture uses on-chip SRAM instead of external memory, which eliminates the fetch penalty and makes inference dramatically faster. OpenAI was in active discussions with both Groq and Cerebras for chips that would handle roughly 10% of their inference fleet.
Then Nvidia wrote a $20 billion check and absorbed Groq’s entire leadership team, IP, and 90% of its engineers. That deal closed five weeks ago.
Read that sequence again. OpenAI identifies a performance gap in Nvidia hardware. OpenAI starts talking to Nvidia’s most credible inference competitor. Nvidia buys that competitor for 3x its last valuation. Now Sam is posting “we love working with Nvidia.”
This is what a hostage negotiation looks like in the semiconductor industry. Nvidia’s $100 billion investment in OpenAI has been “closing within weeks” since September. The Groq acquisition wasn’t a technology play. It was a blocking move. Nvidia paid $20 billion to make sure its largest customer couldn’t build an escape route from GPU dependence.
Sam knows all of this. His tweet reads like damage control because it is damage control. The “insanity” he doesn’t get is eight of his own people talking to Reuters.
Premium: The AI Bubble is a time bomb, burdening hyperscalers with billions of new depreciation a quarter that will soon eat away profits, VCs with dead equity in fallen AI startups, and Oracle with hundreds of billions of unpayable leases.
https://t.co/tAWxEv0NzD
>Build podcast career around saying there are elites on the Epstein List
>Be named Deputy Director of the FBI, tell everyone there is no Epstein List
>List of elites in the Epstein files gets released
>Quit and go back to podcasting
magnificent run