Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Naturegirl1999]

The computer doth protest too much, methinks...

I am marginally more reassured that "Believe me" was revealed to be provided in the seed, and not a spontaneously result of so much insincere use of it out there in its world of original research material.

(Would be interesting to see the eight "original" texts that were then human-edited into this published one. Even/especially if it took less time to do than with human input.)

I agree it would be wonderful to see what the AI actually wrote, not what humans edited it to be

[Reelya]

You know all they need to do now is make an online GPT-3 system that spits out 8 versions of something, then has a built-in editor where you do the moves such as chopping and changing paragraphs, to make one final version. Let half a billion people use that then train another AI off their examples.

[Starver]

The next generation of CAPTCHA!

(Also, related, now that I think about it... Though the Machine Learning Captcha one that my search also brought up, is not what I was thinking of but could be relevent to my first comment!)

[Naturegirl1999]

You know all they need to do now is make an online GPT-3 system that spits out 8 versions of something, then has a built-in editor where you do the moves such as chopping and changing paragraphs, to make one final version. Let half a billion people use that then train another AI off their examples.

+1 to this

[wierd]

A terrible, terrible idea.


Step 1--  Train a set of generative AIs like this, against sites like TheOnion, and against the political rhetoric of the US for the past 50 years. (Include state level politicians, like governors, so you have a very big dataset, with lots of bullshit in it.)

Step 2--  Train more AIs against news anchors. Spread it all over, you want GenericFemaleNewsAnchor and GenericMaleNewsAnchor. Thats the only categorical distinction you should put.  You want to make AIs that can take a "news feed" (generated by group 1), and then spit out semi-realistic prattle about it, and then "engage" each other in such idle prattle.

Step 3-- Throw in some good DeepFake AIs, along with the kind of human face generation software used by "ThisPersonDoesNotExist", to create the avatars for GenericFemaleNewsAnchor and GenericMaleNewsAnchor.

Step 4-- Do the same for various world leaders, to be able to generate "Interviews" and "news reports about speeches".

Assemble the entire shebang into a website called FakeNews.com

[Naturegirl1999]

A terrible, terrible idea.


Step 1--  Train a set of generative AIs like this, against sites like TheOnion, and against the political rhetoric of the US for the past 50 years. (Include state level politicians, like governors, so you have a very big dataset, with lots of bullshit in it.)

Step 2--  Train more AIs against news anchors. Spread it all over, you want GenericFemaleNewsAnchor and GenericMaleNewsAnchor. Thats the only categorical distinction you should put.  You want to make AIs that can take a "news feed" (generated by group 1), and then spit out semi-realistic prattle about it, and then "engage" each other in such idle prattle.

Step 3-- Throw in some good DeepFake AIs, along with the kind of human face generation software used by "ThisPersonDoesNotExist", to create the avatars for GenericFemaleNewsAnchor and GenericMaleNewsAnchor.

Step 4-- Do the same for various world leaders, to be able to generate "Interviews" and "news reports about speeches".

Assemble the entire shebang into a website called FakeNews.com

This actually does sound interesting, maybe the images and videos from FakeNews.com can be watermarked somehow so that if said things were shared, it would be known to come from FakeNews,com

[Starver]

Step 5: Apply the FakeNew.com watermark to all regular footage, when nobody's looking.

[Bumber]

Step 6: ???

[wierd]

So, Pocket suggested this story for me and I decided to read it.

Basically, throwback corn variety produces sugary ooze that promotes nitrogen fixating bacterial activity.
https://getpocket.com/explore/item/the-corn-of-the-future-is-hundreds-of-years-old-and-makes-its-own-mucus?utm_source=pocket-newtab


So; I have a different question.  Since this stuff is basically simple sugars in a suspension, wouldn't this work straight up as a fuel crop?  Some clip on collection trays (similar to what they use to catch opiate from opium poppies), and you have ready-made microbe food for making ethanol fuel, no?  I would need to see some numbers but it looks interesting as a prospect.

Is anyone aware of any work done on that front with this species?

[Reelya]

The economics of the matter speaks for itself. *if* this was an easier way to get sugars, then it would by definition be a cheaper way to produce sugar in general, so companies would be expected to already be exploiting it.

Whether or not it makes sense to burn these sugars, or just any old sugars, as a fuel would be a secondary consideration.

[wierd]

The lack of general knowledge is an obstacle that you are not considering; When all you have is a hammer, all problems look like nails.  Not knowing that you have another tool to work with, or being unfamiliar with that tool, means passing up on potentially better solutions.

That's why I was asking if any work had been done.

Specifically, there is a large industry to produce fructose corn syrup from corn.  Even sweet corn does not contain a truly enormous quantity of these sugars; Processing from starches is often involved.  Since this juicy liquid slime is sufficient to provide up to 80% of the plant's nitrogen needs, through being a feedstock for bacteria of questionable efficiency (being wild bacteria in the open air, and thus having other bacteria involved as opportunistic parasites), this is already simple sugars (so processing from starch is unneeded), etc... 

Without investigation, you simply don't know. 

[wierd]

Here, I found a paper that is open access that deals with this plant, and its exudate.

https://journals.plos.org/plosbiology/article/file?id=10.1371/journal.pbio.2006352&type=printable

Gives the following data about the exudate.



It appears to be dominated by galactose and fructose.

The molarity of fructose and galactose together is very high; it's practically syrup.

Another study shows that normal fermentation processes with ordinary brewer's yeast shows high ethanol production rates for galactose.
https://www.sciencedirect.com/science/article/abs/pii/S1878818117306151

Other research shows me that fructose can also be bioconverted by the same yeast.


So, the real question now, is exactly how much exudate can you derive from these things.

[Bumber]

It appears to be dominated by galactose and fructose.

Fucose, not fructose.

[Starver]

In case you've not seen the apparent Twittering about it already (I hadn't), I thought I'd share this news about the technology behind digital pregnancy testers that I found... interesting.

[Starver]

With help from the police, speeding in your Tesla drive is a dream..

(My take-away from that, after actually checking and finding that Alberta has 100kph absolute limits (though 120kph is the max limit found anywhere in .ca), is that the cars that pulled aside only for the blue lights were previously 'holding back' the unattended Tesla whilst themselves going at maybe >=40kph over the limit. But that's not the point for this post, just an observation.)