AI Bros are Going BROKE... The bubble is deflating | ClownfishTV
(www.youtube.com)
You're viewing a single comment thread. View all comments, or full comment thread.
Comments (10)
sorted by:
This is exactly why the AI / datacenter revolution is so important. Right now AI is funded by the Cabal channels - much like all the Big Tech that we have heard about. Clearly they are all losing money due to operational costs, because the Computational requirement for top tier AIs is crazy.
As the Cabal money dries up, these companies will start finding themselves with their pants down.
This is the biggest leverage Trump has against these Tech companies. As Trump switches from shadow economy to real economy, these companies will do whatever they need to do, to make sure they dont go belly up.
On top of this.. i am in Tech, and i am not paying any money to use AI.
I'm running it locally with
LMStudioorclaude-codein the terminal. I use free portals like https://chatgpt.com, https://venice.ai/, or with throwaway emails at https://grok.com & https://gab.ai or even with corporate 'chatbots' (eg: Hi i am interested in <the product the company has> and i would really like to discuss signing up with a plan, but before i do that, i have to fix this problem i have with python: <code problem here>)I won't be buying tokens at all, and if the free ways ever disappear, i have local options to use on my own hardware.
Do you use local LLMs for coding or project management? Which ones do you use? How does it compare to GPT-5 ?
i'm using whatever LLM's i can get for free.. i haven't really settled on any one, but i use one for code generation when im in the terminal.. i forget which one though.
I have been using GPT5.5 with Codex, and recently lost access to it for a couple days. So I decided to use local LLMs. Apparently Qwen and Deepseek are good. I tried plugging them instead of GPT5.5
Honestly it felt like I gave a trigonometry problem to kindergarten kids. One of them actually said "There are too many files to check and I wont be able to finish it, so I will stop here" 🤣