• 0 Posts
  • 49 Comments
Joined 11 months ago
cake
Cake day: November 6th, 2023

help-circle





  • Correct. Dress it up however you like, but LLM and ML programs are probability gamblers all the way down. We’re building a conversation tool, that doesn’t truly comprehend the language because it’s a calculator at its core - it’s like asking your eyeballs to see in UHF frequencies.

    They’re called “computers” for a reason, and we are deep in the myopic tech tree of further and further complexity. The current wave of AI has solid potential, but not globally for all applications. It is a great at ‘digital assistant’ roles and is already killing it in CCTV monitoring software. Mindjourney can make incredible images, but it can’t make art. ChatGPT can write, but it’s a terrible author or speechwriter.








    • Fired the Supercharger head and the entire department
    • Fired the lead of new vehicle development
    • Previously fired head of battery development
    • Constantly “one year away” from Tesla full self driving, whilst Mercedes just launched geofenced FSD, with Mercedes assuming 100% liability during FSD
    • Elon just had a out of the blue trip to China, appears to have ‘kissed the ring’ of Beijing, and hyping TaaS robotaxis

    What’s Tesla’s USP to an investor now? The supercharger ‘lock in’ and early head start at the EV game are Tesla’s biggest boons, but the former appears to have been gutted and the latter has been squandered on a slow model release schedule




  • I’ll agree there’s huge potential for ‘assistant’ roles (exactly like you’re using) to give a concise summary for quick understanding. But LLMs aren’t knowledgeable like an accredited professor or tutor is, understanding the context and nuance of the topic. LLMs are very good at scraping together data and presenting the shallowest of information, but their limits get exposed quickly when you try to go into a topic.

    For instance, I was working a project that required very long term storage (+10 years) with intermittent exposure to open air, and was concerned about oxidation and rust. ChatGPT was very adamant that desiccant alone was sufficient (wrong) and that VCI packs would last (also wrong). It did a great job of repackaging corporate ad-copy and industrial white papers written by humans, but not of providing an objective answer to a semi complex question.



  • Not really news that Tesla has a lot of headwinds, even before Elon’s peculiarities and ‘leadership’ is factored:

    • Was disruptive in early 2000s, now suffering from many of the same diseconomies as the auto industry
    • Vertical integration especially introduces huge CapEx that has to justify its continued existence via regular, strong sales. There’s a very good reason the big automakers sub-contract out a significant part of their BoMs for production and JiT stocking
    • Traditional auto makers (and government) are pushing EVs hard, greatly increasing competition especially in entry level markets where Tesla has poor showing
    • High profile recalls, durability issues, and ‘autopilot’ deaths leading to growing concern, both regulatory and consumer

    We all know the stock is massively overvalued. The question is will it correct and level off, or fall so fast from that height that Tesla ends up like Lucid, Rivian, or Polestar etc as a low-mid manufacturer, or go so far that the short sellers bring the knives out until delisting



  • …now that the marketplace has been altered.

    Zoom was disruptive in the market because the other options SUCKED - Skype, GotomyPC, WebEx, etc all did a lot of rapid maturing during the pandemic to catch up to Zoom, and this ‘bundling’ of Teams with Office has built user lock-in via Office 365.

    Well played Microsoft, once again ducking a well deserved anti-trust suit. ‘Historic injury’ should count equally as ‘current injury’ for prosecution, arguably especially so if that historic injury is what created your current market share