@benjamin_horne@benitoz@edzitron You're allowed to use obvious AI to refine your own post about AI that is enthusiastic about AI, it's the one exception. It would be a little hypocritical to just shoot off your raw ideaslop.
@The_AI_Investor Don't forget Michael Burry accused Nvidia management of Enron-like accounting crimes months ago, was rewarded with attention for that, and then just moved on from it after being apparently wrong
@cashisking36@eevblog@C_S_Skeptic Automated direct indexing (no fund middleman) is possible now in a few places. You can exclude specific equities that way, but there are minimums.
@Midnight_Captl@jim_jones69 Finance rationalists would just say that you can set up recurring sales if you want income and it's academically the same, but for NVDA in particular the market doesn't always recognize its earnings, and the dividend cuts through that inefficiency. Psychologically nice also.
@Pete_Wexler@SamRo The answer is actually C, since playing each lottery is free and you can submit infinite tickets, ∞=∞, and therefore also D because the question remains underspecified. A and B also correct, because education is win-win.
@OceanSailing3@Mr_Derivatives Anson Advisors, it's in the complaint. You can look up their people who neither admit nor deny paying for short tweets via sham invoices from the convicted fraud at Citron
@TShirtnJeans2@Midnight_Captl@chrisbarber Berkshire is effectively not getting diluted at all due to the GOOG discount, they're making instant gains. Different from entering a new position where they are famously constrained.
@IntuitMachine@spaceandtech_ Between around $60k and $100k depending on variant, but it's meant to be just one tube-launched part of M-LIDS which also includes the autocannon. So overall cost per intercept is hard to say, but could be favorable.
@jjacky@buccocapital The culture has been like this for years, pushing out "low performers" and being aggressive. "Year of Efficiency", "some people shouldn't be here"
@Midnight_Captl@handleym99 I think I disagree that "the inference market will cost ~$0 for consumer / prosumer workloads". Those workloads do value accuracy at the very least, which may be a function of latency or other "premium" token dimensions, via scaling laws. If it can be done for ~free, great.
@handleym99@Midnight_Captl I think intelligence as a product will evolve beyond chatbots/agents, closer to actual good proactive assistants. Basic science fiction, nothing new. The value becomes more compelling, more worth a subscription, but requires frontier accuracy. Just wouldn't bet on stagnation.
@handleym99@Midnight_Captl Yes, and current free tiers of chatbots are not as good as the paid tiers, which will continue to be the case. The point is that ordinary users who don't fully utilize their subscriptions are still paying, because they still get value, and this is underestimated.
@Midnight_Captl The current capability will be free, but standards will move on even for those with low standards. "everything will be handled at the edge for ~free" as a general narrative has consistently underestimated demand for frontier intelligence, there are OOM in capabilities to go.
@HCH_Hill@WilliamtheBTC@AaronBastani Destruction of highly sympathetic and visible wildlife is not controversial to the public, the authoritarian crackdown on this will be lauded.