Get the latest tech news
GPU-rich labs have won: What's left for the rest of us is distillation
AI inference for 90% lower cost
Fortune-500 companies would spend tens of millions and proudly that they trained their own SOTA models, only to have them be antiquated months or weeks after their release. Even with the "generosity" of Meta and Alibaba (Qwen), who have spent hundreds of millions just to release model weights, open source simply cannot compete with the hegemony of superintelligence labs when it comes to general intelligence. If an LLM can't solve a particular task acceptably yet, it's not the worst strategy to build what's possible now and wait a couple of months.
Or read this on Hacker News
