NhanH
The baseline expectation in the original post also applies to the SVP: he knew enough to know that things aren't unreasonable to start with, this expectation might not apply to all companies.
In gold, you trust that everyone is believing gold has value. And that it is commmon knowledge (in the logic sense [0])
Saying that gold has no more or less inherent monetary value than fiat currency assuming a magic wand that can change a large percentage of people's belief.
Then Lenin concluded that if you will always have concentration of power and some folks doing centralized planning, it might well be the state.
They will die, for some definitions of death -- I don't think they disappear, but they should be a niche, rather than the dominant doctrine.
- Hyperscaling “we are going to serve billions of people in our applications”, which is becoming increasing unlikely as regional tech companies become more dominant than than the global one (this one is as much about geopolitics as technology)
- Operations is hard, in which case non-frontier models should be increasingly capable. Devops for small-ish deployment is one of the few cases where it is hard to clam you need deep expertise and AI can’t do it. Previously, the claim is that you need people specialized in ops, which is expensive. Now…
My prediction is that not just cloud LLM, but cloud business general will have to change. Not yet in the next 5 years, but probably 8-20 years-ish
https://professorrobertpape.substack.com/
https://escalationtrap.substack.com
I would rather not paraphrasing or rewrite in my interpretation too much, as the topic is difficult enough as-is. But the short version is that the mainstream analysis of "worst case" is nowhere near his level of worst-case scenario and that the US has no control of the situations any more (the US can still control its actions, but Iran is already emboldened enough to take actions on its own strategic plan). The worst-case scenario has Iran with nuke as the baseline, and Iran also growing to be the dominant regional power in Middle East, on-par with Russian in the peak strength in 2010s.
A standard situation of 1% positive rate and even if your diagnosis method having 90% accuracy (for both sensitivity and specificity), more than 90% of the positive detection would be false positive.
Though I don't think I will be able to cite an official document stating Al Chop is a Snoopy fan :-), so there's that.
> The user studies in this work were carried out with a DLP projector because it provided a variable frame rate for experimentation. However we have observed the effect on every display we have available, including commercial LCD and OLED TV panels. Indeed, the authors first noticed the effect while designing a coded display pattern and started the present research study in order to explain the unexpected observations. The results presented here now serve as design guidance for our work designing better displays.
The study is about an effect, the effect is relevant in coded display, but it is the effect we are studying about, not the display itself. The fact that the effect might be everywhere is the motivation for studying it (the traditional displays you kept mentioning? Apparently the reseachers didn’t find one of those in their office. I believe it is fair to put the onus on you proving this does not generalize)
And let’s take a step back to the big picture anyway: the point is that this effect is very relevant in displays circa 2015 just makes the original article MORE correct.
I think if we continue talking, we will keep running in circles. So let’s drop the details on research: it is there, we can both read it. Here is what I was trying to convey since the beginning:
- If you think the article is an ads, with the writing not up to scientific standard: sure, I am ambivalent about the article itself
- If you think the gist of the article and their recommendation is wrong, I mildly disagree with you
- If you think led-flickering affecting people is in the same ballpark of concern about Wifi or GMOs, I violently disagree with you.
LEDs are new, and so the high frequency related research are not too numerous, but for the few exist, they generally point to a higher threshold of perceiving than previously thought. As for the health-effect, I believe that part is more extrapolation than researched (since those can only come after the more generic research on perceiving). So the final assessment is: how bad was the article in presenting information the way they did.
I mean, they are not even using a screen during the study, they are using a projector. How are you going to even make the claim that this is display technology specific when it is not using a display?!
> The work presented here attempts to clarify “the rate at which human perception cannot distinguish between modulated light and a stable field
Otherwise, they would have tested the dithering directly for the full image. Here they are testing a more simpler model: varying flickering causes higher flickering-free requirements (due to eye movements). This would applies to dithering, but potentially other situations.
Basically the claim is that when there are varying flickering frequency, the requirements for non-flicker frequency is much higher.
They mentioned in Results section: > For the median viewer, flicker artifacts disappear only over 500 Hz, many times the commonly reported flicker fusion rate.
If you want to redo the numbers and check if they fit the definition, please feel free to do so, but you will need to put some works in (since the flicker hz -> risk showing in the article is a computed value, you need to find the modulation value and plug it in too)
I understand your fight and your idea, I am just saying that in this specific instance, this is not a fight to be fought. The article is generally correct, and if you want to complain about the writing style or it being an ads, it’s up to you. But this is not the same situation with GMO stuffs
> Either there's a study showing that 246 Hz flickering poses "Extremely High" health risks or there isn't.
They calculated it using the definition from the standard.
The standard also linked to the researches during their discussion.
Please read it, instead of just randomly throw out things hoping that they supported your argument.
There is nothing anecdote about flickering in LED light causing health risks.
The current median age of China is about 40, which is not great for their context, but a world apart from 52.
This sentence has way too many assumptions doing the heavy lifting.
“Pure logic machines” is not a thing because literally, there are things that are uncomputable (both in the sense of Turing machine’s uncomputability, and in the sense that some functions are out of scope for a finite being to compute, think of Busy Beaver)
To put it the other way, your assumption is that machines (as we commonly uses the term, rather than scifi Terminator”) are more energy efficient than human in understanding the universe. We do not have any evidence nor priori for that assumption.
The message is very deliberate and has been gradually more focused on the “unification” aspect since at least 2014-2015.
This is the author’s own description of his expertise.
I have to look it up since I am not able to validate (by myself) the very strong claim he made in the beginning paragraph. I am comfortable not reading the rest. The claim is:
> Our knowledge has advanced to the point where, if we had a safe and reliable means of modifying genes in embryos, we could literally create superbabies. Children that would live multiple decades longer than their non-engineered peers, have the raw intellectual horsepower to do Nobel prize worthy scientific research, and very rarely suffer from depression or other mental health disorders.