Comments by "clray123" (@clray123) on "Wall Street Millennial" channel.

  1. That point exactly. Hiring dumb employees is not increasing your profit, it's subtracting from it. And AI bots at present are like dumb employees tripping over their shoelaces while pretending to be trustworthy.
    3
  2. There is no exponential growth of technology. The transformer architecture on which most of the AI is based today was invented/popularized in 2017 (and neural networks were invented in the 1950's).
    2
  3. We already have an answer with Amazon's AI (Attentive Indians) running their automated stores (well, no more).
    2
  4.  @angie99656  Ehh just like nobody is going to use automatic checkout at supermarket instead of letting a friendly cashier help them?
    2
  5.  @Wary_Of_Extremes  With the tendency for overfitting and hallucinations, I can see researchers/scientists just as likely wasting much time being misled by ML models as they are to make new discoveries / improve predictions.
    2
  6. Correct, AI will be limited to entertainment for the masses as well as clue/idea generation to seed further research by professionals. Not replacing said professionals any time soon, contrary to what the peddlers are pretending.
    1
  7. You can already self-host models equivalent to ChatGPT 3.5 turbo on your local PC. The main problem is that they are just not that useful for anything much, really. The reason why the AI stuff will not really ever be truly open source is that the training is so costly. Training new LLM from scratch todat would be $80K in hardware rent alone, but you need to add to it expenses for obtaining and filtering data (depending on what you want to do, this can become really, really expensive - the LLMs have the "luck" of feeding from publicly accessible Internet), possibly initial tests with different model architectures, alignment and benchmarks aimed at your application, possibly retraining again if anything turns out wrong... this is not something a big bunch of individuals can realistically do with small contributions of their spare time. And it involves a lot of meticulous work which is not very fun to do (another exclusion factor for open source projects).
    1
  8. Sharing generated stories with Microsoft Clippy.
    1
  9. NVIDIA is gonna crash as well when they are left with a huge pile of inventory nobody wants to purchase from them when the market comes to senses.
    1
  10. But you can already do just that (renting hardware from the cloud). And not many revolutionary innovations are in sight. Unless you mean training a brand new LLM (why?), that will currently cost you $80K for a one like the original ChatGPT (down from $1M a year ago).
    1
  11. Don't forget he is Microsoft top guy's boyfriend. That is his entire role in this.
    1
  12. "What we can't do in profit we will make up in revenue."
    1
  13. As a developer, no.
    1
  14. Not likely, as the smaller ones will remain dependent on the infrastructure provided by the large players to host their services. It's not like you can build up your own computing center to support your software. It's a bit like with energy prices - where energy prices are high, everything is expensive. You cannot out-compete your big electricity provider by providing cheaper electricity yourself at scale.
    1
  15.  @wezzard  You are assuming that competition will work. But AI is an oligopoly, why would the vendors compete with each other to their mutual detriment instead of colluding to keep the prices up? They are not as independent from each other as you might think. Same goes for big energy corporations - I mean there is even a completely official cartel - OPEC - which is there to "manage" the price and suppress competition according to political whims. Free market prices are only possible if there are many vendors and the products/services are easy to imitate, but this is just not true for platforms/fundamental infrastructure. Today whole wars are being fought for it (while pretending it's for other reasons).
    1
  16. Copilot is not even any longer popular. People getting tired/bored of it.
    1
  17.  @Poochie1  the main question is - where to? this bubble inflates everything else along with it
    1
  18.  @matt.stevick  never forget the scam makes money when it goes up, but also when it goes down
    1
  19. Let's be honest, this is the biggest shitbubble of history, rivaled only by the scale of public money theft achieved during pandemic. But here the wealth transfer is (sort of) voluntary. Sort of because the assets from pension funds of unsuspecting people will disappear together with the businessmen here.
    1
  20. The thing which makes it gambling is that AI companies are actively misinforming prospective investors about what the can do. Even the term "AI" is scammy (the companies are not to blame, it was minted as such many decades ago). The algorithms are expensive to run and the performance is underwhelming. They are calling mimicry and copying templates "reasoning". Extrapolating from early breakthroughs (transfomers: 2017) far into the future, ignoring the fact that no more such breakthroughs have happened since. Etc etc. This hype train is gonna crash not just because of the economic mismanagement, but because the snake oil they are selling is really just that. As more people get to learn and use these deficient products, the disenchantment will set in and trust will be lost (already happening with ChatGPT). I can guarantee you that the CEOs, well informed by their top research staff, are truly aware of what they are perpetrating here. They have a problem of sunk costs though, the more they lie and overpromise, the more they are forced to uphold their lies. As usual, they will come unscathed when that house of cards eventually collapses. I think they are banking on government rescue when push comes to shove, just like the bankers did in 2008. P.S. As with the "pandemic" there will be talk about how "nobody could have predicted" and there will be no consequences for the guilty.
    1
  21.  @gamesguy  Catch 22, it is a good reason to not get hooked on these AI assistants. According to your theory every basic Internet service, like search, should cost heaps of money because people are already "hooked" on it. The problem is the "assistants" are not really all that useful compared to not having them... they are competing with the free Google/Bing search engine and more specialized human-driven (also free) search like StackOverflow. So the companies built a paid product which competes with their free offerings, which doesn't sound very smart. Unless, of course, they want to discontinue the free products, but this may cause users to quit using them altogether. I believe the only advantage of AI assistants is that you can make them spread your advertising or lies more efficiently than search result pages. So maybe that's the plan, devise a better way of fooling consumers to behave as the megacorps desire them to behave, which certainly has some value (e.g. you can rig elections, brainwash people into new taxes etc.)
    1
  22. The problem is you can't just screw them, they are "managing" (stealing) your pension money. When their globally orchestrated scam fails, they will beg for bailout and receive it.
    1
  23.  @highbread817  Keep a close watch on when the insiders start selling.
    1
  24.  @taragnor  The tool has to be good and fit for purpose though. If the tool produces incorrect/misleading results which cause extra cost and burn time of the tool's user often enough, that tool is discared and another sought (or you can do without, as before the crappy tool's invention).
    1