Nvidia’s New AI Chip Delayed Due to Design Flaws [Bloomberg]
Nvidia’s New AI Chip Delayed Due to Design Flaws [Bloomberg]
You're viewing a single thread.
Oh noes.
AI industry is losing some interest (consumers not trusting AI products, companies that have tried AI undoing those changes), intel have recently lost a SHIT tonne of valuation/credibility.
NVidia delaying a product due to issues might hit them harder than they think.Cracks in the facade aren't great with a freshly inflated bubble. It'll be interesting to see how this develops.
If you believe in technical analysis of stocks, NVDA formed a double top in mid-July which is an extremely bearish signal. (If I am not mistaken, it's caused by many investors switching from long to short position... Or some kind of magic like that. The data is out there regarding short sales, but I just gotta look, and I am lazy.)
But yeah, something more realistic is the recent trickle of news about OpenAI starting to falter financially which is compounded by the recent (free'ish) release of Llama 3.1.
LinkedIn is silent on AI, and has been silent for a while. The AI snakeoil companies are dead silent on there now and most major companies have reverted their direct marketing back to pre-AI crap.
So yeah, the cracks have been forming for a while and I think NVidia has better things to worry about than a dedicated AI chip. ("Design flaws", my ass. )
I don't believe in TA, but I do think if if we don't see any large scale monetizable "AI" use cases in the next 12-18 months, Nvidia's stock price is going to get hit.
Small scale, relatively niche "AI" use cases will continue to grow and develop, but I am talking about novel products that would support Nvidia's current valuation, not niche techy use cases.
It will be interesting to see what happens to all this "AI" hype in the next 12-18 months.
I wouldn't be surprised if the total global spend on "AI" since the Jan 2023 (when ChatGPT3 got big) is approaching ~$400 B and they don't really have much to show for it.
LLMs are good as an enhanced spell check or a quick quasi-search engine, but you cannot rely on LLMs for anything important.