toth

10 ms·

61.
by toth 3y ago
> You said > * HTF provide liquidity. > * HTF beeing invited into markets formerly excluding them. > * Regulation attempts on HFT hit retail traiders too. > * You disagree, that HTF is rent seeking? > Imho, all those state
62.
by toth 3y ago
LLM growth (or frontier ML model growth in general) is driven in part by the original Moore's Law, but right now the main driving factor is the growth of investment. I.e., how much companies are prepared to spend on training run. This
63.
Google Gemini Eats the World(semianalysis.com)
62 points by toth 3y ago|21 comments
64.
by toth 3y ago
I work in HFT, so feel free to dismiss what I say as biased... There is this widespread perception that HFT traders are rent seekers, and bad for markets, and in particular for retail traders. Anyone well informed knows this is completely w
65.
by toth 3y ago
I always wonder about what something in between popular science/math and academic journals would look like. Something oriented towards people that know a nontrivial amount but are not researchers. E.g., you can assume they know calculu
66.
by toth 3y ago
You probably are aware, but just to make it clear for others: topologically a torus has one hole (genus 1). The inner void is not considered a hole. Similarly, a 2-sphere (surface of solid sphere in 3d) has 0 holes (the inner void is not co
67.
by toth 3y ago
You are right, but I am pretty sure they just meant the inner tube ignoring the puncture for inflation. Usually people just use donut/bagel for an illustration of a torus. Not sure why they used inner tube here - maybe to make it clear
68.
by toth 3y ago
This type of instictive negative reaction to prediction markets is, unfortunately, common, but, I think, misguided. Prediction markets are one of the (or just, the?) best ways of aggregating knowledge from multiple sources and producing the
69.
by toth 3y ago
Maybe not everyone, but if JPM doesn't then who does?
70.
by toth 3y ago
Very cool! The top answer has a 12,731 character regular expression that matches base 10 numbers that are divisible by 7. (There is an answer with only 105 characters, but that uses .NET "regular expressions" which have some pret
71.
by toth 3y ago
You are right that this is state dependent, and I am certainly not a lawyer. I have no direct knowledge about WI. My personal experience concerns NJ/NY. Still, what I understood at the time is that, if the employer is not paying you at
72.
by toth 3y ago
Off topic, but I have to ask. What do you have against Kelvin?
73.
by toth 3y ago
That is very interesting. As you most likely know, Borges was famously a huge anglophile, he once said "I find English a far finer language than Spanish" in an interview. There's a video of his reasoning: https://w
74.
by toth 3y ago
Not really. If you are not being paid, you can just take a new job during the non-compete period and if your previous employer sues you they will lose, and they know it. They could still sue you just to make a point and make you spend money
75.
by toth 3y ago
FYI, in all jurisdictions I know of, this is already the case. I.e., a non-compete period in which you are not paid to a level commensurate to what you got before is not enforceable. If employers try to enforce them in court, they will lose
76.
by toth 3y ago
This a letter anyone can sign. As someone pointed out Grimes is one of the signatories. You can sign it yourself. Hinton, Bengio, Norvig and Russell are most definitely not getting prestige from signing it. The letter itself is getting pres
77.
by toth 3y ago
Roll to disbelief. That tweet is precisely about what I mentioned in my previous post that doesn't count: finetuning LAMA derived models. You are not going to contribute to the cutting edge of ML research doing something like that. For
78.
by toth 3y ago
State of the art AI models are definitely not something you can develop in a basement. You need a huge amount of GPUs running continuously for months, huge amounts of electrical power, and expensive-to-create proprietary datasets. Not to me
79.
by toth 3y ago
I disagree. My read on him is that until very recently (i.e., possibly when GPT4 came out) he didn't take x-risks concerns seriously, or at least assumed we were still many decades away from the point where we need to worry about them.
80.
by toth 3y ago
I think you are wrong. The risks are real and, while I am sure OpenAI and others will position themselves to take advantage of regulations that emerge, I believe that the CEOs are doing this at least in part because they believe this. If th
81.
by toth 3y ago
As a EU citizen, I really wish we had competent leaders, particularly in Brussels. Have never seen any evidence that is the case, specially with regards to tech. The last EU initiative at sweeping regulation in the tech space (GDPR) was dis
82.
by toth 3y ago
While I don't disagree with you, if you are going to say something like you should really at least give the right number. Or at very least include a disclaimer that 12 hours is not the right number. 2022 Meta revenue was 116 billion US
83.
by toth 6y ago
This is really fun. Thank you!
84.
by toth 7y ago
You should really adjust for dividends, in which case it is still down, but only -7.5%. [1] [1] https://dqydj.com/nikkei-return-calculator-dividend-reinvest... (I did Dec 1989 - Jan 2020. They don't have data latter th
85.
by toth 8y ago
This is definitely wrong. In multiple conversations with Uber drivers, they prefer the longer trips as it makes them more money. They are able to not take long trips if they prefer (although they don't know the exact destination before
86.
by toth 9y ago
Leap year rules have to do with the actual length of the year being 365.24219 days - not with days not being 24 hours.
87.
by toth 9y ago
Not the OP, but it wanted to point out it has 10X less error on the holdout sample so it is not simply overfitting.
88.
by toth 9y ago
Congrats on the paper, very cool work. I bet the "Datasaurus dozen" is going to become the go to example for the dangers of relying on summary statistics, it is much more memorable than Anscombe's quartet.
89.
by toth 9y ago
Not sure why this is downvoted to gray, it is correct. In most countries, if you are a tax resident, your foreign income is taxed. What is unusual about US is that it also taxes foreign income for US citizens that are living abroad, almost
90.
by toth 9y ago
I see, that makes sense. I guess at that point you have something very much like std::map, except it doesn't keep itself sorted, you have to do it yourself. I guess for cases like this where you initialize it once and never change it a

More ›