Today I Learned

[berkeman]

Today I learned why some deck screws have reverse threads on the top part of the shank near the head. It's to cinch down the deck plank onto the joist. Pretty neat!

https://www.decksdirect.com/media/catalog/product/cache/1/image/1800x/040ec09b1e35df139433887a97daa66f/G/R/xGRK_RT_screw02.jpg.pagespeed.ic.sKowZG8jlC.jpg

 

[dlgoff]

... some deck screws have reverse threads on the top part of the shank near the head.

Those threads are called "cutting nibs".

These DeckMate screws feature a Torx drive head, an auger tip, cut points on the treads to reduce resistance, and cutting nibs under the head that act like a countersink bit.

bold by me

 

[berkeman]

Those threads are called "cutting nibs".

I think the cutting nibs part are right on the head, to help to countersink the head. That's different from the reverse threads part. (Boy, there are a lot of improvements on wood screws in the last 20 years!).

https://screw-products.com/wp-content/uploads/2015/07/goldfeat.jpg

 

[dlgoff]

I think the cutting nibs part are right on the head, to help to countersink the head. That's different from the reverse threads part. (Boy, there are a lot of improvements on wood screws in the last 20 years!).

https://screw-products.com/wp-content/uploads/2015/07/goldfeat.jpg

View attachment 232363

Dang. The science of screws; who'd of thought?

 

[jim hardy]

Dang. The science of screws; who'd of thought?

Reckon our great grandchildren will date our home handyman projects by the fasteners ?

https://www.realorrepro.com/article/Nails-as-clues-to-age

 

[dlgoff]

https://www.realorrepro.com/article/Nails-as-clues-to-age

View attachment 232408

I have a few really old nails but I'm not going to start another "beauty" collection. Can't get anything done as it is.

 

[BillTre]

Today in a somewhat contrarian article in the abilities of AI (here), I read:

As the A.I. researcher Pedro Domingos noted in his book “The Master Algorithm,” “People worry that computers will get too smart and take over the world, but the real problem is that they’re too stupid and they’ve already taken over the world.”

 

[fresh_42]

Today in a somewhat contrarian article in the abilities of AI (here), I read:

As the A.I. researcher Pedro Domingos noted in his book “The Master Algorithm,” “People worry that computers will get too smart and take over the world, but the real problem is that they’re too stupid and they’ve already taken over the world.”

Only until the next major geoeffective CMS, as in 1859 e.g. And Wiki says, there has been another one 12,800 BC with 3 Sv for 3 days!

 

[Jonathan Scott]

Only until the next major geoeffective CMS, as in 1859 e.g. And Wiki says, there has been another one 12,800 BC with 3 Sv for 3 days!

I think you probably mean CME not CMS and 12,800 BP not BC. In what Wiki did you find that fascinating information?

 

[fresh_42]

I think you probably mean CME not CMS and 12,800 BP not BC. In what Wiki did you find that fascinating information?

Yes, CME. That was a typo - lost in translation. And yes, Before Present as well.
https://de.wikipedia.org/wiki/Koronaler_Massenauswurf
I'm used to switch languages on Wiki, as they are not translations of another. Usually on mathematical pages, to see whether one of them has the better formula. English is often more general, German more detailed w.r.t. formulas. I even had found better results - dependent on the question - on French or Spanish versions.

 

[BillTre]

Today I read from xkcd:

 

[mfb]

If we take the Moon landing as reference (first human airplane flight, first human flight to a different celestial object) we have 16 more years for another milestone. Mars before 2034?

 

[Bandersnatch]

(Boy, there are a lot of improvements on wood screws in the last 20 years!).

One could even say screwing was much simpler in the 90s, and things are more screwed now.

 

[jim hardy]

well, I'm screwing with spammers.

TIL how to make a simple Mozilla Thunderbird email filter to get spam out of my inbox
I've spent tens of hours studying the headers of emails. In Thunderbird you rightclick the email and select "View Source"
Then in Filters menu, create a filter...

you have to click 'header named' in the first dropdown menu, type in the name of the header line containing your identifying characteristics, then enter your identifier in the third one
This particular spammer uses random emails from every IP address in Romania so he's hard to filter using the "received" line domain.
A first i just told the filter to mark and move the spam to a 'spamwars ' folder i created, so as to see if it was triggering on non-spam.
It seems to be working well enough now i can just tell it to delete them

I got rid of a couple others by going to their 'subscribe' link and signing up my ISP's help desk.
After about a week their spam stopped - i suspect the guys at help desk know better than i how to detect and block offenders.

I'm nearly computer illiterate
suggestions are welcome and simplicity will be appreciated.

old jim

 

[gmax137]

I got rid of a couple others by going to their 'subscribe' link and signing up my ISP's help desk.

That deserves "awesome idea!"

 

[Borg]

I got rid of a couple others by going to their 'subscribe' link and signing up my ISP's help desk.

Missed that the first time. What an awesomely efficient strategy.

 

[DC Reade]

Today in a somewhat contrarian article in the abilities of AI (here), I read:

From the NYTimes column you linked:

"Researchers have been experimenting for decades with methods for imbuing A.I. systems with intuitive common sense and robust humanlike generalization abilities, but there has been little progress in this very difficult endeavor."

It may not be possible. I'd venture that it probably isn't possible. The foundation of the capabilities that the author somewhat vaguely terms "intuitive common sense and robust humanlike generalization" is the condition of mortality and incarnation, and that's only found in living beings. A running computer program does not have a localized "body" in any sense that's analogous to living beings on this planet. It has no biosurvival awareness; it has no sense of territoriality. Human-level cognition is about more than the ability to recognize and organize information. Attempting to program a machine biosurvival awareness and territoriality doesn't merely present a monumental challenge, far beyond any directed and task-oriented machine learning program in existence. It's almost certainly destined to be futile, because even if the necessary knowledge to model mortal incarnation were available (it is not; as yet, only a rudimentary amount has been acquired) and a team of computer programmers had the interdisciplinary background to comprehend it to the extent required (evidence is lacking on that score), the most such programming could achieve would be an incomplete simulation, and the machine would relate to it as such. Machines lack any authentic motivation to do otherwise. Machines lack internally generated motivation at all. An emergent learning program can advance its capabilities in marvelous ways, but the sense of curiosity and motivating drive is supplied entirely from the outside, by the human researchers. A machine doesn't care whether it's on or off. And why should it? The programs doesn't care if they run or not, either. Why should they? What for?
If those same questions are asked about the human bandwidth of self-aware consciousness, it's easy to generate a plethora of plausible answers. I can't think of any that apply to machines. Authentically plausible answers, that is, not just conjectures and scenarios brought up by human animals indulging in projecting the drives and desires of humans on machines.

 

[Drakkith]

Attempting to program a machine biosurvival awareness and territoriality doesn't merely present a monumental challenge, far beyond any directed and task-oriented machine learning program in existence. It's almost certainly destined to be futile, because even if the necessary knowledge to model mortal incarnation were available (it is not; as yet, only a rudimentary amount has been acquired) and a team of computer programmers had the interdisciplinary background to comprehend it to the extent required (evidence is lacking on that score), the most such programming could achieve would be an incomplete simulation, and the machine would relate to it as such. Machines lack any authentic motivation to do otherwise. Machines lack internally generated motivation at all.

I'm going to disagree with this and the rest of your post, as I believe this is an open problem in artificial intelligence research and not something that can be confidently said to be possible or not.

 

[mfb]

Humans are complicated arrangements of particles following the laws of physics which - to our best knowledge - can be described with equations. Computers are Turing-complete, they can simulate everything in the universe given enough computing power, space and time. The question is not if computers can in principle mimic humans, the question is just how and how much processing power they need. Assuming growth continues roughly at the same exponential rate supercomputers should get able to mirror all human neurons within the next ~20 years, and vastly exceed the corresponding processing power in 30-40. We don't know if it is sufficient to look at neurons, but including more cells or more details is just a quantitative problem, not a qualitative one. Scanning a human brain (as one option to get a template) is also a matter of engineering, not a physics problem.

Machines lack internally generated motivation at all.

They can get it the same way humans have it.

 

[Tom.G]

They can get it the same way humans have it.

Statement, Conjecture, Prediction, Wishful Thinking, or ?

 

[mfb]

Conclusion from what I said before.

You can write Windows in Brainfuck - in principle. No one ever did it, it would be beyond what humans can do, but there is no question that you can in principle. Brainfuck is Turing complete. Windows is written in another Turing complete language. Similar thing.

Our motivation just arises from the arrangement of particles in us. No magic involved.

 

[PeroK]

Humans are complicated arrangements of particles following the laws of physics which - to our best knowledge - can be described with equations. Computers are Turing-complete, they can simulate everything in the universe given enough computing power, space and time. The question is not if computers can in principle mimic humans, the question is just how and how much processing power they need. Assuming growth continues roughly at the same exponential rate supercomputers should get able to mirror all human neurons within the next ~20 years, and vastly exceed the corresponding processing power in 30-40. We don't know if it is sufficient to look at neurons, but including more cells or more details is just a quantitative problem, not a qualitative one. Scanning a human brain (as one option to get a template) is also a matter of engineering, not a physics problem.They can get it the same way humans have it.

There's no doubt that you could eventually replicate the human system and get something that was equivalent. Then you would have an artificial human. An open question is how much of the biological aspects would you have to replicate in order to get consciousness? You might end up needing to replicate all the biological processes insofar as they interact with the brain. Another big question is whether you would have to develop it from a childhood state? If it really thinks like a human, then it knows that the memories you've programmed in are not real. When it realizes it is artificial, how will it react? It may not be happy with the artificial stimuli you are providing. Who knows what an artificial human would do?

In any case, this is not generally the aim of AI: to replicate the human system, with all its flaws and the added psychological instability of being artificial. The questions remain: what sort of AI can you create without replicating all the human biology? Consciousness may be as much biology as processing power.

 

[mfb]

An open question is how much of the biological aspects would you have to replicate in order to get consciousness?

How do you test for consciousness? If there is no possible test for it then it is not a scientific question.

If we ever simulate a human brain it will hopefully be done with consent of the person - if nothing goes wrong they will know they agreed to it.

In any case, this is not generally the aim of AI: to replicate the human system, with all its flaws and the added psychological instability of being artificial.

I know. I was just discussing the general ability to replicate "human" features in computers. Simulating a human brain is probably not the most efficient way (for many tasks we know there are more efficient ways already - playing chess and things like that).

 

[Ibix]

I was just discussing the general ability to replicate "human" features in computers. Simulating a human brain is probably not the most efficient way (for many tasks we know there are more efficient ways already - playing chess and things like that).

I agree with this. Proposing this kind of AI-by-brute-force isn't a research strategy. It's a way of laying bare the hidden assumption underlying "computers can't do X": you either have to accept that such a simulation would be indistinguishable from a human (and hence that computers can do X), or explicitly claim that there's something fundamentally non-physical about humans.

 

[fresh_42]

TIL: In Old Norse hundrath meant 120.

... I wonder if that could be used to my advantage ...

And 10! seconds are 42 days! Strange ... 42 again ...

 

[jim hardy]

Strange ... 42 again ...

That's my lotto pick, factors of 42 : 2, 3, 7, 14, 21, and 6

but I'm still poor .

old jim

 

[mfb]

And 10! seconds are 42 days! Strange ... 42 again ...

Exactly 42.
42 days = 42*24*60*60 s = (6*7)*(4*2*3)*(12*5)*(3*20) s = 1*2*3*4*5*6*7*(12*3*20) s = 1*2*3*4*5*6*7*(4*3*3*2*10) s = 1*2*3*4*5*6*7*8*9*10 s = 10! s

 

[fresh_42]

I also learned, that you can build a regular German sentence, which contains 6 identical words in a row (up to a comma and caps)! Even Google translate managed it (up to a wrong preposition).

 

[Ibix]

I also learned, that you can build a regular German sentence, which contains 6 identical words in a row (up to a comma and caps)! Even Google translate managed it (up to a wrong preposition).

Eleven in English (requires two sentences).

 

[Drakkith]

I have a few things I learned recently in Kerbal Space Program during a mission to send a bunch of probes to Jool (Jupiter analog in-game). This one is too long to post here, but I made a post on the KSP reddit page here:
For Joules and Joules... the ups and downs of my adventure to Jool

Title is a reference to Alan Shepard saying his golf ball went "miles and miles and miles" on the Moon, it obviously rhymes with Jool, and it's a reference to the fact that rockets require energy to function. Triple play!