Zero reason to believe open AI employees, big tech has history of employees/subcontractors creeping on exes and minors, someone with keys to creep on customer for valuations is more parsimonious.
Cheap money gone, investor money being burnet in bubble, only thing long term is government money, but gov money still not enough to get US defense primes to make stuff at scale and not SaaS the shit out of hardware (see lockmart f35). I doubt SV will be any different. My feeling is any $$$ US spends on soft MIC stuff, PRC going to acquire anyway and have their industrial base utilize more efficiently. If US tech end up build AGIing, and when US gov as it how to beat China expecting 5d chess answer, AGI going to conclude crippled US industrial base simply can't - it and defect.
300 billion tokens is like.. $5-25 million giving range of OpenAI ouput prices, I"m sure they pay less at cost so, I wonder if more $$$ in wage hours have been spend by humans on the problem. My feeling is yes?
PRC's been white labelling / JV components for tier1 HVAC brands for 15 years now, they're fine quality/lifespan wise. Flagship PRC heatpump already caught up to JP in HSPF2 @ ~11 (GREE). Either way, more cheap commoditized heat pump = more installations = more aggregate global efficiency - every $ spend on JP premium is inefficient allocation of resources in terms of maximizing global energy savings.
Arguably, PRC price / value engineer is their innovation and by far more MORE important than vs minute efficiency differences. Ultimately it's about TCO and global adoption, if JP can't bring their prices down, if they can't fractionalize cost, then they're doing more harm than good for net transition - there's better things to with that 30-50% premium.
Oh I totally get their game - and its a smart one. Flood the market, own the distribution and servicing and you win the game. I feel like the GREE product is tier 2 but most people don't care. Its not comparable to the well made products - they aren't junk but they aren't quality either. PRC seems to have targeted the broad market - smart move.
Id argue that the PRC price / value engineer is only viable because of the development that the innovations in heat pumps have managed to provide. Can't have one without the other. Though once price / value owns the market - innovation dies.
Midea/GMCC, Gree spend ~5% of revenue on R&D like rest of industry, except they're larger vs JP companies (multiple size of related JP conglomerates & magnitudes larger than JP niche thermal management companies). Also more directly focused on thermal management/hvac (i.e. Gree doesn't have defense/space like Mitsubishi to split R&D). Then factor in PRC PPP, PRC simply spending magnitudes more on R&D and will continue to (involute vs domestic competition). Hence most of the tier1 western suppliers are now using white labelled / rebadged components (compressors etc) that are designed by PRC manufactures.
PRC just doing innovating AND value engineering - they have both. IMO with respect to energy efficiency tech is one of those "any idiot can build a bridge that stands, but it takes an engineer to build a bridge that barely stands". The primary metric for "quality" of energy transition products is efficiency / $, and in that JP fails hard. Only PRC is scaling efficiency / $ effectively, which has much greater value than JP trying to push top end, because frankly that's all they can do. Which has it's economic rational, i.e. TOC in wealthy countries where installation/labour expensive, but its still net, lower "quality" innovation because $ spend on JP premium nets less aggregate efficiency.
Hopefully PRC will involute the shit out of compute and commoditize it like agriculture, margins should be so thin that ASML/TSMC/NVIDIA should only be sustained by subsidies.
reply