reallymental

Prediction:

Color me pessimistic, but it's obviously not going to be the entire picture. This explanation is going to live that sort of vague truth/lie land that the majority of news today exists.

It will instill confidence amongst less tech-savvy people, by "putting it out there" that TikTok is actually magnanimous in releasing it's secret sauce.

If by some magic they do release the real thing, it will be gamed so quickly that it will be rendered moot.

If I were them, I would do the second. That way you'll release the truth (at that time) and reap the benefits of good PR.

edit: grammar.

Never assumed the day where I would have intimate knowledge of HN topic would arise.

This is my space. A very crowded space, and incredibly chaotic, where data is outright wrong most of the time.

Personally, I don't prediction should be the end goal of these analyses, tracking and anomaly prediction would give more insight to the traders etc.

But of course I would only say this cause prediction is hard in this space.

DL (Deep Learning) Et al. don't really help. Traditional stats is more likely to be helpful in this case.

Incredible, but wasn't this kind of available though psycopg2 (postgres connector), SQLAlchemy or any other kind of database connection library ?

I realise the difference between the kernels altogether, but how is one better than the other?

See, I don't buy this (I'm not arguing with you, just initiating a conversation).

There is no doubt we are incredibly complex, so much so that we've not been able to crack the puzzle that is "us". We invent crude medicines to carpet-bomb some bacteria/virus, which can potentially harm another part of the body, we staple cadaver-parts onto our knees to move well again, we're incredible crude in a bio-hackish way.

We've not even begun to understand what goes on in our heads.

Room for improvement? I'm not sure we are there yet. Room for understanding what is going on? A lot of room there. Its a massive empty hangar to put all our future knowledge into.

Bottom line is ... I don't think we should go about pulling strings without caution because it will not happen without consequence.

Why would you sell your goose that lays golden eggs?

Even if you think that the goose might be getting older and less fertile, it doesn't make sense.

It only makes sense if you know when the goose is gonna run out of eggs. And in this case I'm not sure anyone knows.

I agree with your first sentence, but not so much with your second.

He's the best out there now as he's got the pick of his guests (the result of a forward feedback loop of being a good interviewer).

But view/listen to enough of them and you'll learn to pick the gems from the drivel, like almost everything.

okay, it does sound like you know what you're on about...

I'm a skeptic as well, but I'm not a physicist (I am an engineer so I know the basics).

I however, cannot throw a convincing rebuttal to this.

All I know, is that a high ranking officer has requested a patent to go through for a seemingly "perpetual motion machine".

All the buzzwords for the UFO-logists are there to get them salivating about Area-51 and Bob Lazar.

But, imagine such UFO-logists are reasonable about their beliefs (hard, I know) and are open to other suggestions (they clearly are to many).

Given such UFO-logists, how would you tell them this is bullshit without getting emotionally frustrated at the ludicrousness of the article ?

edit: any papers that point me towards the possible edge of today's tech would be welcome as well

It seems like a very French thing to make the 'conquered' (or in this case, employed) part of oneself.

Antithetical to any other way of merely being the administrative authority over the employed.

Schneider Electric (again a French company) is doing this as well... is there a need for such cap raise? It seems like they're unloading risk onto their employees, from a cynical perspective.

I wonder why its unethical to test such things on human beings. I'm assuming it's unethical, because there rarely seems to be any damn work on this using humans.

Those rats ran themselves to death, will we humans? We share about 92% of the genes[0] (on the first google attempt), what changes does that giant 8% contain?

Just pick up people like me, and make us run and give us food.

I doubt we will object if it means there's a possibility that we might help ourselves.

Are they afraid they'll make it worse? Sure that's a possibility, but that's the price to pay. I'm willing to pay that price if it means people after me, who have this wire crossed in their brains will be served better healthcare in the future.

Some experiments might not work, but we've just discovered another way to _not_ make the lightbulb, but the good news is that the lightbulb _will_ be made if we try n times.

[0] https://blog.23andme.com/23andme-and-you/genetics-101/geneti...

Thank you, I just read through that and it's all very anxiety inducing. Probably because it's true.

I'll give it a try. I realise I need to.

It's chilling when you see a description of what you are doing, online, in a random click that you just made.

I have OCD, and hide it well. Looks like I have ON and have not hid it well at all. I run obsessively, eat extremely clean and have been struggling to even eat things that are outside my "normal" diet ever since the stress at work and relationships have increased.

I figured it was something I needed to research when I felt _bad_ for eating a sandwich after a 56 mile bike ride. But I never did, maybe because I didn't want to know the answer.

Oh well, the more you know...

My OCD has increased as well. They're definitely connected.

Edit: grammar and some spaces.

How far away is that? 20 years? That's 20 years worth of "laughing all the way to the bank" for that guy.

Also, does every business need to scale more than 1 employee? Even if it's 2, 3 people max, that level of "freedom from worry" would make that duo/trio very happy indeed.

AFAIK smart electricity for scheme for even a block level is notoriously hard with today's tech, the number of edge cases must be enormous.

The machine recognizing you was the harder problem. That's the "Minority Report" stuff that people thought was in the _far far_ future that we don't have to worry about constantly.

Storing and using that model across various devices is trivial. So if we've done the hard part, the easy part is _definitely_ going to be done. There's no doubt about that.

States, companies that can convince the state that they somehow need the data (cough insurance companies) will, I repeat, WILL utilize it. And now, suddenly your premiums for this month have gone up cause in that ski trip you took last weekend, a camera noticed the lack of headgear for about 5 mins. That's the kind of risk the insurance company cannot take. Pay up.

Let's assume that they knew this was a half baked product.

Not a "first edition" product like the first iPhone, but a rushed, half baked product, released without consideration of sales but just for being the first in the market for this niche.

Given this assumption, why would you give it to reviewers at all ? Remember "Vertu" ? They were the Ferrari of phones during the Blackberry era and were _shit_. I got my hands on them as a kid and was astonished how bad they were compared to my dad's cheap simple Blackberry. There's a reason why they were not reviewed widely, or positively by a large margin.

Why not just do the same thing? Don't be daft and give it to the hands of reviewers who have twitter on their fingertips and will tell everyone about every single flaw about your device, when you know your device has flaws.

Give it to the CEOs/Sports personalities, and let them flash it around for a few days/events, and cover up the flaws in the second edition. If it breaks on their hands, they certainly won't care to advertise that fact, they'll just get another phone.

If they actually thought they had finished this product, well... SSNLF is going to notice a strong presence of gravity for the next few days.

Perhaps you're thinking of Fermat[0], who famously remarked "I have discovered a truly remarkable proof of this theorem which this margin is too small to contain.", and promptly proceeded to take the solution with him to the other side.

I doubt anyone burned his manuscripts however, but I seem to recall an article where they mentioned that everyone searched for the solution (or for his approach) in every one of his notebooks, but couldn't find it.

[0] https://en.wikiquote.org/wiki/Pierre_de_Fermat

I suppose the amount of stress/strain held within these rocks will be hard to ascertain. If these earthquakes occur periodically, sure that gives us a better estimate, but we still don't know the unintended consequences of triggering a quake. Not to mention the property damage.

Yes absolutely, we're just exploiting potential energy here, and that was my point, perhaps I wasn't clear enough.

That's why the efficiency of the power plant (if this energy would have been captured) would have been a sight to see on paper. The energy they put into the fracking would have been tiny compared to this.

edit: spelling

---- From the article:

A South Korean government panel has concluded that a magnitude-5.4 earthquake that struck the city of Pohang on 15 November 2017 was probably caused by an experimental geothermal power plant....

Unlike conventional geothermal plants, which extract energy directly from hot underground water or rock, the Pohang power plant injected fluid at high pressure into the ground to fracture the rock and release heat — a technology known as an enhanced geothermal system....

---

This is both horrifying and incredible. They managed to fracture enough rock to generate energy contained in a mag-5.4 earthquake.

That's around 200 KT of TNT [0]. That's a lot of useful energy if contained. I wonder what the efficiency of the plant would be then.

South Korea already has 23 nuclear plants though [1], I wonder why they decided to go this route, perhaps easier setup and lower associated costs ?

[0] https://science.howstuffworks.com/environmental/energy/energ...

[1] http://www.world-nuclear.org/information-library/country-pro...

(Disclaimer: Not American)

I've been reading NYT for a long while now, and I've been noticing more and more of these "medicines are more expensive here" trope.

There's half an answer isn't there? The US is subsidizing the development of these new drugs to the world, at the cost of it s citizens, as these companies operate all over the world and not just in US. The _relatively_ richer people in Cali can afford to pay $x more than the people in Bali.

Sure, the exploiting the people under the marketing guise of "patented" paracetamol varieties is _very_ unethical, and just slimy. That we can all agree on.

But every time such an article is posted, the only question in my mind is, "isn't the problem half explained?".

What's the alternative, not innovate and let everyone die? As opposed to innovate, and some people who live better than the others pay a bit (or in this case, ~2.49x) more ?

Is there an alternative where everyone pays equally ? Sure, you can lower the cost for an American and subsidize it (which comes from their tax again, so they're being more/less charged the same amount).

or you can increase the price for the rest of the world until the death toll vs affordability ratio looks good for the company that decides this is a good idea.

The problem is half explained.

The rest of the problem's explanation must lie here.

--- From the article:

Part of insulin’s price rise in the United States is because of the middlemen who buy the drugs on behalf of insurers and hospitals and negotiate discounts off the list price for their clients. So Lilly often doesn’t make the full $275 a vial (though, since rebates are secret, we don’t know how much less).

By selling an authorized generic, rather than merely lowering the brand’s price, Lilly is essentially doing an end-run around those middlemen and giving patients who don’t purchase through an insurer another option.

---

It's also a supply chain problem, not the manufacturing supply chain, but the business to customer supply chain.

I guess margins are not that wide enough for middle men to make anything significant in other countries except the states.

My tone in this comment is very inquisitive, not aggressive in any way, I don't mean to come across like that.

dannykwells' link has got all the good points for why a Jupyter Notebook is kinda disadvantageous, but what about the new Jupyter Lab?

It's not got support for simple code/header folding yet. That's an issue.

Most people have heard of this "method" of performing invasive brain surgery (so that the surgeon knows he/she's not messing with the patient's ability to play the guitar, which may be linked to other abilities etc.).

But I've literally never seen a story where they did this, and they found that the patient stopped playing the guitar/stumbled playing it/played better than ever before.

So what happens in any of those cases?

1) If they stop playing ? Does it mean they have 'damaged' a part of the brain ? Can you actually fix it ?

2) If they stumble/forget parts of the tune? Have you partially damaged the part of the brain that's responsible for that muscle memory?

3) If they suddenly become Hendrix? Have you found a way to turn mortals into gods with a metal stick ?

I'd think the distribution leans heavily towards 1) than 2) or 3).

But I've never seen any (1, 2 or 3), either they don't occur (i.e the surgery is just routine without any adverse effects) or I've not bothered to dig that deep into medical journals.

Any links would be appreciated.

Who's to blame when the model sees 'red' ? Management needs a head, a model isn't one yet.

So what happens with all this data?

I'm willing to bet some of it is pretty critical[valid credit card details?], it might end up in the Dark Web, auctioned off or something like that.

Some of it is useless (not valid anymore, or just junk emails used to sign up).

Cathay doesn't tell us what else has been leaked (or they don't know), but what if there's a team on the other end of the hacks that actually analyses the meta data of these hacks, and finds out the frequency of movement of some people, and add that to their 'alpha' ?

That's a deep rabbit hole. What about the will of those on JusticeServed? Were they free to decide to be on that?

As a naive young person, I hoped that a metal illness was just that, an illness that can be cured.

After discovering mine, the curtain was pulled back, revealing the 'trick' all along behind the illusion of cure, was just suppression. Bacteria are killed, sometimes Viruses too. But thoughts, and feelings cannot be cured or killed, they must be suppressed.

For 'Free Will', one should be free to control one's will.

I can't pass through doors in the first try, or the 20th, sometimes when I'm by myself.

Some people cower right before a fight they've prepared for, for months, they knew fight day was coming.

Anger kills people, people 'see red'.

If these thoughts and feelings are your will, hopefully we find out that they're free, not you.

Besides the heterogeneous nature of being an "Indian", the wobble is just agreement.

The person who's wobbling their head is agreeing to what is being said. It's not a yes/no.

It's mostly a "Go ahead and say what you want, I'll do what I want to do after you're done".

I didn't grow up in India, so I can see it for what it is, but I am Indian, so the subtleties are visible to me.

And aptly, people who don't need to take any shit/be kind in that instance/appear submissive don't "wobble" their heads. Those people mostly reserve it for their elders (parents/elder relatives mostly).

Take all of this with a huge grain of salt.

Yes I'm aware its got the MIT licence [0].

I should have made myself clearer, I was referring to this particular example, where the OP's articles plots don't render anymore.

I know that's how they make money, by making it possible to share plots (or make them accessible) through their own platform.

One could have just* self hosted them with bokeh.

Oh well, to each their own. Maybe they just wanted to show the cool thing, rather than having to worry about all the devops stuff.

My gripe was pedantic, I know.

[0]https://github.com/plotly/plotly.js/blob/master/LICENSE

* with a non trivial amount of effort.

Plotly's got one great feature: Dash apps, that's it.

It's true bokeh can take a bit of work to get it to look like plotly, but I don't need to pay for it.

I think that's a more 'scalable' feature of bokeh.

It's just too damn bad that bokeh doesn't have something like Dash. I've tried Holoviews, bokeh apps, they just need so much more work. And its integration with flask/Django is just not there yet.

Hopefully future versions of bokeh will get this fixed.

An interesting, and quite a bizarre article. Thanks for pointing it out.

It'll be interesting to find out how prevalent this indicator is.

"This new indicator – rather than looking at the spread between longer-term yields of two years and 10 years – is looking at the spread between short-term yields. It’s “based on the spread between the current level of the federal funds rate and the expected federal funds rate several quarters ahead derived from futures market prices.”

- The above is weird, this indicator completely flips the focus from 'long term' outlook, to the 'QoQ' outlook. That too derived from futures market prices.

So if I can spin up some 'projects' (on a large-ish scale) that look like they'll do well for the next 6 months, I can make this graph look great. That's brilliant, we've just avoided a recession for the next 6 months.

This is kind of the same model that corporations work on. But corporations have regulations enforced on them to keep them from running into everyone with a chainsaw (EPA comes to mind).

The yield curve has been flattening for a while now, especially for longer term (10Y, 20Y, 30Y T-Bills) securities [0, 1, 2].

As confidence in the government (i.e through a trade war) dips, as interest rate hikes become more of a reality, and as increasing amount of companies come forth with weirdly worded statements about their 'strong financials' and 'consistent growth', everyone is running for the hills.

No one needs to toot their horn when they're on a bull run, everyone can see it, and everyone is on their own run. It's only when no one can see you're having (or you will have) 'strong financials' that you need to announce it.

And it seems like you need someone to announce their strong position to keep the heard mentality of a bull run going on.

I very well might be wrong.

[0] https://fred.stlouisfed.org/series/T10Y2Y [1] https://fred.stlouisfed.org/series/DFII20 [2] https://fred.stlouisfed.org/series/DFII30

There is one meetup for data science that I know of [0].

Cycling in dubai is.... well, uncomfortable.

The Dubai Roadsters are the only serious group that I know of[1]. The flats are really quite boring, and the dust really kills my appetite for air after a while. Also, cycling during the summer (8/12 months) isn't really recommended. If you can tolerate the heat, go for it.

[0] https://www.meetup.com/Dubai-Data-Science-Meetup/ [1] http://www.dubairoadsters.com/