Regardless of what happens theta and IV crush will probably wipe you out. I don't like Tesla's stock but I don't touch it just because both the stock and options tend to be way overpriced.
Yeah it's been a good learning lesson. I only put in what I was comfortable losing (and fast!). The way the stock moved after the absolutely abysmal earnings will certainly stick with me lol
It really is wild that investments are driven by the marginal investor, not the median investor. 99% of us can think that Tesla is trash, but 1% of world investors is an absolute ton of capital.
The last price in any market (whether it's stock shares or housing) is driven by the market liquidity which is extremely inelastic. It mostly just does whatever it feels like short term and the time it takes for elasticity and fundamentals to overwhelm it can be so agonizingly long.
You're more complaining that investors who don't own a stock have no influence on its price. Which is true, but I don't see a workable way to change that.
The median investor in Tesla, on the other hand, seems to be happy with the situation since they're not selling.
I'm not complaining really, just think it's a explanation that describes the downward sticky nature of companies that can't seem to justify their valuations.
I agree that the median investor feels that way, I just think that the median Tesla investor (apart from passive broad based funds) is a tiny, tiny, tiny part of the market.
Actually, the reason is the opposite. Tesla is reportedly over 40% owned by retail investors compared with under 20% for most big tech stocks. It's a meme stock.
Investing in a proportional index fund moves the market as a whole, but does not move the individual stocks in relative rank. Aside from short term frictional liquidity issues, it just makes the stocks' relative movements exaggerated.
The valuation of Tesla is still decided by the marginal investor.
One could even be excused for the paranoid thought that there's a conspiracy of capital backing techno authoritarians. Of course, some of that is a money maker, like surveillance tech. But these are the same people backing dodgy brain implants, and third rate LLMs at fabulous valuations. And who are OK with merging a dying social media site with that third rate LLM start up.
Did the reporter reach out to Anthropic for public comment on this? They list a "source familiar" with some details about what the intended purpose was for, but no mention on the why
They’ll never make their money back. Autonomous driving is mostly software and will be commoditized very shortly after it works well.
There’s not enough money paid to drivers in the world today to repay the investment in autonomous driving from direct revenues. It’ll be an expected feature of most cars, and priced at epsilon.
Autonomous driving and the attendant safety improvements will turn out to be a gift to the world paid for by Google ad revenue, startup investors, and later, auto companies.
I agree here, because the profit margin on taxi services is too low. Well, on an unrealistically long time horizon, like 50 years, they might make it back, but surely much worse returns that investing that same money into US Treasury bonds.
> Autonomous driving is mostly software and will be commoditized very shortly after it works well.
I disagree here. To be clear, when we talk about AI/ML here, I separate it into a few parts: (1) the code that does training, (2) the training data, (3) the resulting model weights, (4) the code that does the inference. As I understand, self driving uses a lot of inference. (Not an expert, but please correct me if wrong.)
How can Waymo's software be "commoditized very shortly after it works well" if competitors don't have (2) and (3)? The training data that Waymo has incredibly valuable. (Tesla and some Chinese car companies also have mountains of it.)
You point out yourself that Waymo doesn’t have a monopoly on training data. And the trained model is software, of which the price-to-use trends towards epsilon, even when it’s very expensive to make. For example Google search, maps, docs, YouTube. An exception is Netflix, but there the value provided by a subscription is access to novelty, which is not intrinsic to driving software.
I'd really expect that he is actually at a loss on all of those except for SpaceX which has a clear path towards being cash flow positive if it isn't already
You don't have to make money to be worth an astronomical amount. HN should know this better than anywhere.
I hate "fictitious" valuations as much as the next guy, but at the end of the day it's what people are willing to pay for equity that determines value, not what it's books look like.
Erm, why not? A 0.56 result with n=1000 ratings is statistically significantly better than 0.5 with a p-value of 0.00001864, well beyond any standard statistical significance threshold I've ever heard of. I don't know how many ratings they collected but 1000 doesn't seem crazy at all. Assuming of course that raters are blind to which model is which and the order of the 2 responses is randomized with every rating -- or, is that what you meant by "poorly designed"? If so, where do they indicate they failed to randomize/blind the raters?
> If so, where do they indicate they failed to randomize/blind the raters?
Win rate if user is under time constraint
This is hard to read tbh. Is it STEM? Non-STEM? If it is STEM then this shows there is a bias. If it is Non-STEM then this shows a bias. If it is a mix, well we can't know anything without understanding the split.
Note that Non-STEM is still within error. STEM is less than 2 sigma variance, so our confidence still shouldn't be that high.
Because you're not testing "will a user click the left or right button" (for which asking a thousand users to click a button would be a pretty good estimation), you're testing "which response is preferred".
If 10% of people just click based on how fast the response was because they don't want to read both outputs, your p-value for the latter hypothesis will be atrocious, no matter how large the sample is.
Yes, I am assuming they evaluated the models in good faith, understand how to design a basic user study, and therefore when they ran a study intended to compare the response quality between two different models, they showed the raters both fully-formed responses at the same time, regardless of the actual latency of each model.
I did read that comment. I don't think that person is saying they were part of the study that OpenAI used to evaluate the models. They would probably know if they had gotten paid to evaluate LLM responses.
But I'm glad you pointed that out, I now suspect that is responsible for a large part of the disagreement between "huh? a statistically significant blind evaluation is a statistically significant blind evaluation" vs "oh, this was obviously a terrible study" repliers is due to different interpretations of that post. Thanks. I genuinely didn't consider the alternative interpretation before.
Sure, it could be, you can define "preference" as basically anything, but it just loses its meaning if you do that. I think most people would think "56% prefer this product" means "when well-informed, 56% of users would rather have this product than the other".
