Comments by "kazedcat" (@kazedcat) on "CNBC Television" channel.

  1. If it did not threaten the market then why would they stop retail from buying the stocks? So what they did is purely to save the short sellers that is criminal.
    32
  2. John Scovill No. I really hate that Blue origin is considered number 2 in new space. It ignores Rocketlabs which has already delivered several commercial satellites in orbit. Rocketlabs is number 2 and Jeff Bezo's company is not even a contender.
    13
  3. I thank the smart folks at DeepSeek for their contribution to the inevitable coming of our magnificent AI Overlord.
    6
  4. Microsoft is already deploying DeepSeek.
    6
  5. Rn1024 No Blue origin has no experience on heavy lifting they did not even reach orbit. Heavy lifting is all about delta V and orbital rockets has higher delta V than suborbital rockets. If you do not reach orbit then your rocket is puny.
    5
  6. Starlink already deorbited several malfunctioning satellite they burn up before reaching the earths surface.
    3
  7. Embrace, Extend, Extinguish
    3
  8. You miss that DeepSeek V3 is open source so they dont need to work on the base model they can just download it. Also Berkeley study already shows that the techniques also works for smaller models. This means even if you have to start from scratch using the Reinforcement learning techniques will allow you to quickly reach parity with the most advance AI's
    3
  9. CNN's headline is that Tesla have halted S and X production. Like of all possible headline they think that is the most important part.
    3
  10. The executive is going to Lucid Motors so he is still doing electric. And if I am Lucid I will poach Tesla executives too. They are the ones who have extensive EV production experience. Their guidance on what to avoid will save you a lot of money. A lot more than the salary overpay necessary to poach them.
    3
  11. ​ @Felix-s9z7i Elon is now selling robots he does not care if the vehicle market goes down the drain.
    3
  12. For poor people in africa they will use cellular internet and use starlink as backbone for remote towers. Starlink was not marketed to be cheap. Spacex needs to put several thousand of satellites in orbit that is expensive. It was marketed to be better option than current satellite services but it is still a satellite internet so you should compare it to the price of other satellite services not fiber optic internet available in big cities. I have been correcting this misinformation that Starlink will somehow free people from the tyranny of wired internet provider. But people keep believing this wishful thinking. No Starlink is a better satellite internet not an end to wired internet services.
    3
  13. The media is attacking the tech companies are adapting. Except Open AI who is whining.
    2
  14. Pfizer tested 40000 people Moderna tested 30000 people. Out of 40000 people 170 develop Covid symptoms 162 of them have taken a placebo while 8 have taken the Pfizer vaccine. Out of the 30000 people 95 develop Covid symptoms 90 of that 95 have taken a placebo and 5 have taken the Moderna vaccine.
    2
  15. ​ @xinyiquan666 No they prove that the technique works on smaller model but they did not show that a smaller model can compete with giant models. The model they have develop was very limited and can only solve high school math problems. Not the PhD level problems that o1 and R1 can solve.
    2
  16. 1 billion in profit umm ignore that 23 million Bitcoin loss will be the headline. This is evidence that watching CNBC helps you get less informed.
    2
  17. Just distill it there us no moat.
    2
  18.  @nagaea7409  Blue Origin is a suborbital launch project and there is no evidence that this will change in the next few years. Meanwhile Rocketlabs is putting commercial satellite in orbit yet everyone dismiss them. Rocketlabs is a more capable launch company with healthy revenue stream. Blue Origin is simply Jeff Bezos vanity project and will be dissolve once he is bored with it.
    2
  19. They are extracting more money per user. But there is a limit of how much money an average user can spend. As you increase margin more and more. More and more user get priced out of the apple ecosystem. This in turn will force them to increase margin to keep revenue from dropping which will result into more user going to be price out. Then as user keeps dropping the network effect keeps dropping in value which means more user will drop out of the ecosystem.
    2
  20.  @mactek6033  So you have no say on who will be the CEO. Corporations are not a democracy where you can vote clowns for president. Who has the most money get to decide. If you want to replace the CEO then buy the stock and force him out.
    2
  21. ​ @musashi-san____1409 No Elon did not name Tesla. Tesla was already named before he invested in it.
    2
  22. Meta will distill DeepSeek onto their 7Trillion parameter Llama. DeepSeek's weakness is that it cannot expand to meet demand due to GPU restrictions. DeepSeek will eventually be force to choose between using their GPU for training or using it to serve users for inferencing.
    2
  23.  @R3tr0v1ru5  I also hate the regulation that prevents me from freely stabbing you with a rusty knife. Why won't the government allow me to stab people when I want to. It is unfair.
    1
  24. ​ @DynamicUnreal They don't have the edge. Every cloud provider is now scrambling to productize DeepSeek R1. The gap in inference efficiency is very wide.
    1
  25. They don't even need to ship it. They can build their AI supercluster in Malaysia and just remote control it from China. They can even rent their supercluster from Amazon and dont need to buy GPU's
    1
  26. ​ @nahlene1973 They are a quant company they have a lot of money.
    1
  27.  @xinyiquan666  The size of their group does not change the conclusion of their study. They only studied a small model that can only solve high school level math problems. You cannot then say that $30 is enough to train an AI that can solve PHD level math problems. Yes they are small and their cost is also small but if they are larger their cost will also be larger.
    1
  28. @xinyiquan666  They are not random guys and it does not matter if it was Mr. Bean who did it. Their study did not prove you can build an AI model that can solve PHD level problems for only $30. What they prove is that $30 dollars will get you a high school level math solver.
    1
  29.  @skyak4493  Without credits it is not GAAP anymore. If you want to know the real bottom line you use GAAP and GAAP includes all revenue and all expenses. You cannot ignore some revenue and still pretend to be using GAAP numbers. GAAP is the standard where no expenses or revenue gets excluded. If you want to exclude credits then that is Non GAAP. That is the reason for GAAP so that no one can cook the book by picking and choosing what to exclude and what to include in the financial report. If you want to pick and choose your number then you are absolutely not GAAP compliant. You can do creative accounting but don't call it GAAP. Again GAAP is standard accounting principle where everything is accounted with no exception.
    1
  30.  @skyak4493  Other car companies are paying them to build more factories so they can produce more EV. I think that is legitimate way to earn money. If you are selling pickaxe to people digging gold. It does not really matter if the gold will run out. You already sold your pickaxe.
    1
  31.  @skyak4493  No since the pickaxe store use the money they get from selling pickaxes to expand their business. They just release a new product autopickaxe that does all your digging. You only need to point where it needs to dig.
    1
  32. Microsft is already deploying DeepSeek on their platform. Dont think Microsoft does not know how to Embrace, Extend, Extinguish.
    1
  33. Reports of slowdown is FUD. Basically they choose to report the lowest month of the quarter and compare it to the highest month of the previous quarter. Teslas delivery is cyclical so there is always best month and worst month in their delivery. Tesla short sellers targets that to mislead the unaware. But a quick sanity check reveal it as a lie. How can you have record delivery if you are slowing down? To suggest that as what is happening is insane.
    1
  34. ARM is already being ditched with RISC-V. If it is possible to replaced x86 then it is more prudent to jump into an open ISA instead of being lock for a second time to a proprietary ISA.
    1
  35. It is to funny that they did not realize the irony of their statements. Their entire channel is market manipulation.
    1
  36. Open AI is currently the King of AI. Nvidia only has hardware they have no market leading AI solution on their own. Microsoft will dump them to develop their own AI ASIC to cut inference cost.
    1
  37.  @sirus312  Yes it will be brutal for Nvidia but not much gain for AMD. Similar to their console deals it will be an ultra low margin contract. MS can leverage their scale to negotiate custom ASIC from the many AI hardware startups not just AMD.
    1
  38.  @nagaea7409  I could not find the mass and specific impulse of BE-7 so there is no reasonable way for us to judge how it compares to Rocketlab"s Rutherford engine. I also could not find the mass and specific impulse of BE-3. Blue Origin is all smoke and mirror design to fool people who could not input numbers into the rocket equation. In order to claim BE-7's superiority we need 3 numbers the specific impulse, the mass, and the thrust output. Having the thrust output alone is useless in measuring an engines performance. Also in terms of throttle ability Rutherford engine can go from 0 to 100% thrust. BE-7 has double the thrust of Rutherford but without engine mass we do not know if Rutherford can beat BE-7 by simply doubling the number of engine.
    1
  39.  @nagaea7409  Nasa did a complete review of the National teams proposal so I defer to them in terms of the merits of the propose architecture. What I have problems is the constant over estimation of Blue Origin's capability.
    1
  40. Rn1024 No it cannot do that because it is a paper rocket. And they will not test a real hardware doing unmanned flight to do land, launch and moon insertion operation. Which makes their system dangerous. This overestimation of Blue Origins capability that whatever they do will work the first time even if they have zero experience in actual orbital operation. And they are so arrogant that their proposal omits unmanned testing of actual moon landing.
    1
  41.  @nagaea7409  Blue Origin is all smoke and mirrors but no show. Rocketlabs should have more attention than them because they have an orbital class rocket not the puny suborbital rockets Blue Origin use to fool people into thinking that they are relevant. Blue Origin should be rank with amateur rocketeers that can only deal with suborbital rockets. That is the limit of their capability and they have not shown the ability of growing out of that limit.
    1
  42. ​ @coltzhao It's an LLM model but it's tiny only 1.5B parameters so they focus it to solve high school math problems.
    1
  43. All the AI software is built around Pytorch. Nvidia has no moat. Pytorch are already developing tools that can automatically optimize AI to different hardware. The convergence of all AI to the transformer model makes this process easier. Mathematically transformers are just matrix operation and any hardware that has a dedicated matrix engine can be easily targeted by the Pytorch framework.
    1
  44.  @ngroy8636  Transformers are magic but they are just matrix operation in the end. MI300 also has the magic it is called sparse matrix multiply.
    1
  45. It depends in which mathematics. Buying 1 share worth 3000 versus buying 10 different stocks worth 300 each share. You are spending the same amount of money but the second option you are 10 times more diversified. When you assume zero risk then the first option looks the same to the second option but factoring risk the second option is mathematically better. Your expected return is better with the second option.
    1
  46.  @epemsley3787  Mathematically "The company you believe in" should not matter. If we are going to include human psychology then apply it the same to stock splits. Also what you believe does not affect the fundamentals of the compnay. And even if the company have good fundamentals it could turn out that the executives are doing something illegal that would hurt the stock price. It is impossible for a shareholder to know 100% that there is zero risk in a company.
    1
  47. I feel I lost IQ hearing them talk over each other. I don't even remember the content of their speech just that all three of them are very angry and you should stay away or they might snap at you.
    1
  48. Good that means more supply for the rest of the world.
    1
  49. They must have been using their fleet as collateral. The bankruptcy probably allowed them to unload their debt by giving up their fleet. Now they have no debt but also no fleet that is why they have the opportunity to rebuild their fleet from scratch by buying Model 3. Model 3 are very good in depreciation so creditors will have confidence to let Hertz borrow money again.
    1
  50. Yes you don't understand their worry is that the labor market is too expensive and wages are going up. They wan't to keep you poor.
    1