Many users attacked beffjezos's take on open-weight models as untrustworthy or brain-dead, while positive users highlighted their ability to enable advanced AI agents impossible with closed models.
Based on 16 visible X reactions from 32 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Falling token prices could halt venture funding for frontier research.
@beffjezos With closed-weight models like GPT-5, Claude, or Kimi K3, building you own AI-Agent is impossible. E.g. Qwen3.8 provides this level of control, allowing live-code to be embedded directly into your hot-plugin self-programming local Open-Source AI-Agent. https://github.com/NathaliaLietuvaite/Quantenkommunikation/blob/main/PQMS-ODOS-MTSC-V-MAX-12.md#PQMS-ODOS-MTSC-V-MAX-12-APPENDIX-A-9
I think it depends on the degree to which you believe the success of Chinese models is from distilling American models or from pursuing aggressive efficiencies out of necessity and picking up alpha from one another by being part of a more open AI community. If it’s the latter then the thesis falls apart.
@beffjezos The commentary is less of the issue - it's the accumulation of power behind someone who doesn't deserve it and should not be trusted @TheJusticeDept he should be scalped along with the traitors from that TT Trojan horse who got rich spreading an algorithm that poisoned @america
@beffjezos These guys will do everything but figure out how to make money or become profitable. The market has assigned this commodity fair market value. He’s sad he can’t over sell oranges for 100x the cost.
@beffjezos Anyone who doesn't recognize this simple game theory I can only assume is either stupid or anti-american, especially in a world where some are considering distillation fair play
@beffjezos dean is such a pseud i'd be ashamed of agreeing even if he was correct
Ok, I actually agree with this. If there is no expected revenue there is not enough of an incentive to train extremely large models, so Decel. At same time, if there's only closed models, then the model makers tend to collapse to an oligopoly, which is Decel. A balancing act.
Many users attacked beffjezos's take on open-weight models as untrustworthy or brain-dead, while positive users highlighted their ability to enable advanced AI agents impossible with closed models.
Based on 16 visible X reactions from 32 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@natolambert Wondering if the money will stop happening if progress stops. And progress is dependent of new ideas from th ecosystem. Otherwise even large closed orgs suffer from context rot
also a good point tbh