Knowledge vs Slop

Knowledge and Symbolic Reasoning

When people gain useable insights into the operation of the universe, such as Einstein's law of Special Relativity, they have encoded those insights into "knowledge" which can be efficiently (if not completely) transmitted to other people, usually by explanation.

Generally this means that they have encoded insights into a reduced symbolic form, like E = MC^2.

When neural networks learn information, it is encoded into vast numbers of coefficients.  This form of knowledge cannot be efficiently transmitted to people, and it probably does not match symbolic concepts people already have.  So LLM's implicitly learn how different words are used, without necessarily learning or using our general categories such as noun and verb (which may be the most intuitive to us even if they present troubling exceptions to LLM's).

Proponents of LLM believe this is fine.  They want AI to deliver solutions.  They don't need AI to explain itself.  They feel that it is unnecessary for AI to help us learn how to solve problems for ourselves.  We stand on the shoulders of AI to do more.

But if AI is letting us solve problems without fully understanding them, then it is making us dumber.  The "solutions" that we are thereby creating might best be understood as slop (the word popularized by Cory Doctorow, who is an excellent critic of AI).

When you are a homeless person passing through a soup kitchen, they dish out slop.  That's fine because they have limited resources and limited staff and that is the only way they can feed so many people cheaply.  But when you are an aristocrat dining in a fine restaurant, or just a person who has enough time to do so, you want a carefully prepared meal, not slop.

Slop is perhaps unavoidable, but generally it is something we should preferably avoid.  Ideally everyone should have carefully prepared meals.  That's part of a quality life.

Resisting AI means we will not achieve the (alleged) productivity benefits.

But preparing slop makes us dumber.  Preparing meals carefully makes us smarter.  In the long run, this is more important than being "more productive."  It is much better to do less, and to understand what we are doing and learn how to do it better, than just to "do more."

Consuming more slop makes us poorer, not richer.  (Don't trust GDP and similar metrics here.  What is really most important is not how much we consume, or how much money circulates, but deepening our quality of life.)

We should seek to invent the technologies which make us smarter, not dumber.  Only by being smarter can we know and appreciate quality and how to get there.

Therefore, we should seek to build the society that makes us learn more, think, and create, not just dish out more and more slop.

Making people dumber and dumber is the quickest road to collapse of everything.

That also happens to be what you get by mindlessly raising "productivity."

"Higher Level" thinking

Proponents of AI think the sloppiness is fine and it enables us to think at a "higher" (more abstract) level while the AI does the lower level thinking for us.

But this higher level often becomes little more than BS and hand waving.

It is my feeling and my belief that the strongest learning comes from working things all the way through.  This is not a new idea.  Euclid famously told King Ptolemy I: There is no royal road to Geometry.

So when I build my programs, I do it this way.  I think problems through with paper summaries or diagrams first.  I think about the different kinds of ways they could be solved, and choose what appears to be the best one.  If it proves to have been a wrong choice, I flip to another one before I have written much code, if possible.  I build everything from the raw ingredients of my operating system and programming language as much as possible.  Only if things appear to be particularly tricky do I look for previous solutions (aka libraries) that I can use.  If fairly easy, I even reimplement the parts of those libraries that I need.  I rely heavily on built-in language features or libraries including things like associative arrays (aka hashtables) which are capable of dealing with many if not most hard problems.

I know this goes against the grain.  From the very beginning of my 39 year career in computer programming I was taught the mantra "Reuse."  But I reject that as a general rule for many reasons:

1) Learn (everything) by doing (everything).

2) Programs built upon combinations of even fairly simple libraries can become ever more impossible to fully understand.  Often different libraries do not intuitively connect with one another.  Then all your code becomes translating information from one library to another--very dull.

3) Copyright, patent, and similar issues.

During the whole process, even before starting to code, I start writing the user documentation as well.  This is invaluable in determining the fine details of the interfaces.  If something is hard to describe, it's probably not designed well either.

I don't create a 'detailed design' such as including all variables and data structures before coding.  That's basically humanly impossible.  When I was required to use a formal design process, most people could not actually perform a useful Design Review to being well into the coding process if not nearly complete.  As one of my colorful (and PhD) colleagues remarked, "We're supposed to do Design after Coding.  I prefer design while coding."

For over a quarter century, I've either written the documentation into the program itself, or straight into fairly simple HTML.  I like being that close to the metal.  I hate word processing programs.  I do all my editing in Gnu Emacs.

I've had some experience doing things other ways.  Java programming, for example, is traditionally done with the importation of dozens or even hundreds of libraries, with interactions so complex that fancy tools are needed to work out the ramifications and keep each library installed at a compatible version and all the interfaces correct for that version.  General code does little more than call one library after another.  This is the pinnacle of the "Reuse" concept.  I hated it.  It wasn't programming in my opinion, it was dishing out slop.

AI is a vastly greater extension of this.

Now I am very happy to be able to search the web to find code to solve each unfamiliar issue as it comes up.  I don't just cut and paste the bits of found (or generated!) code.  I read them and figure out how they work.  Then I write them into my program.  (My post-retirement program MakePlaylist was created exactly as described above, except I haven't written HTML documentation for it, only in-line documentation that gets spit out into help messages and full documents by built in program options.  But now I am writing HTML for a far more challenging project: a multivolume book about my life.)

DO WE REALLY NEED MORE PRODUCTIVITY ?

Capitalists, oligarchs, and their high priests known as Economists insist that all good things come from increasing productivity.  But they do not.  Increasing productivity may mean more income for them, but lower quality of life for all as everything becomes slop, prepared and consumed mindlessly.

Now old fashioned machines and even automation may be just fine, when they do the heavy lifting and boring routine tasks for us.

But the creative and thinking parts are not only the parts we most like doing, they are also the parts that make us better when we do them.

Now suppose you are a departmental manager responsible for several projects.  You could either have project manager staff for each project, or do them all yourself with AI.

If you didn't like working with people, why did you become a manager?

Having a staff working on each project means you can have informed feedback about the practicality of each project.  Doing it all yourself means you don't get that essential feedback.  It is an error of pride to believe that you don't need that feedback from another person.  The end result is slop which lacks humanity and depth, the hallmarks of great art.

It reminds me of the music created by electronic and automated music generation pioneer Raymond Scott.  Scott invented machines to do things like sequencing, pioneering devices that became very useful to many musicians.  For that he should rightly be honored.  But he invented these machines so he wouldn't have to work with other musicians.  The result in his own subsequent life's work is very lively music which is also very shallow. 

The world we want to construct is one in which each person contributes what they are best doing, which is quite often what they like doing best or something adjacent to it.  Turning all jobs into dishing out slop is exactly the opposite.

What we want to do is the thinking and creative parts, and have machines do the awful, boring, and repetitive parts.

And in many cases still and forever, the best machines are machines custom built for those purposes.

Israelis on Raping Prisoners

 Clip from Israeli talk show has one Israeli, speaking in English, calling the highly documented rapes of Palestinian prisoners (now finally reported in NYTimes) a "blood libel", while the other Israeli, speaking in Hebrew, saying the only problem is that rape is not the official regulated policy of the state, so the rapists don't fear prosecution.

https://x.com/muhammadshehad2/status/2054466066894422508

I have already published several debunkings of Israeli claims about October 7.  Grayzone has some of the best.  Here is another one:

https://www.trtworld.com/article/18165357

Debunking claims about the Hitler-Stalin Pact

 https://x.com/pawelwargan/status/2053755272724615386

"God" and Amelek

 It was not 'God' (a monotheistic concept) who commanded Saul to genocide the Amalekites, it was YHWH, the god of the Hebrews, who later became conflated with El as polytheistic Yahwism evolved towards monotheistic Judaism.  Any modern person who takes this command to genocide as an example of superior ethics and morality is seriously deluded.  Whether this actually happened, or things like it, or not, Hebrews got what was coming to them several times over in the following centuries, and there are few Hebrews practicing Yahwism today.*  The Torah of the Second Temple era, compiled during the Persian empire, demanded that only the Messiah could rebuild a state for the Jews, and this was re-iterated after the Bar Kokhba revolt in the Talmud Three Oaths.  Anyway, there is negligible evidence that the Amelekites even existed in Saul's time, let alone that Saul organized 200,000 soldiers to fight them.  The Agagites, the alleged descendants of Amelekites, were a rival group in Babylon who hoped to exterminate Jews because Mordecai wouldn't bow to Haman.  Negligible evidence of that too.

[Posted to this thread, doubtful it won't be censored.]

*Or perhaps, Zionism is the re-emergence of Yahwism, a tribal rather than universal religion.

Claims and counter-Claims about Immigrant Crimes

A priori, it seems reasonable to believe that immigrants are less likely to engage in crimes (other than "immigration crimes" of course, though most immigration "crimes" are "civil crimes") because it increases the risk they would get caught and deported.  OTOH, some seem to believe that illegals are more likely to engage in violent crime because (1) they already broke the law/rules of immigration, and (2) they are darker skinned people (etc).  Some of that perception seems to come from racism and bigotry.

That legal immigrants commit less crime than citizens is rarely disputed by those examining the data.  The only question is with illegal immigrants.

CATO has written many articles claiming that illegal immigrants commit less violent crime based on actual data from Texas.  (I'll link one below.)

Now, you may dismiss CATO as a hack right wing organization with an "open borders" agenda because that suits their corporate sponsors who want cheaper labor.  And quite often, I do dismiss CATO's conclusions, and many of their comments even in these articles.  But I believe they are presenting the actual crime data here, and their articles on crime are widely quoted in the mainstream news media.  MAGA who are even aware of CATO say they have "gone communist."  That is not what actual Communists think.  

It's not surprising mainstream media quotes the Economic Policy Institute (EPI) much less, because that's a lefty union related organization worried about worker conditions, but they reach the same conclusions as CATO on crime rates.

The people like CIS and others I have seen who claim to debunk CATO on immigration have just as much of an agenda, as you can often glean by looking at their other articles.  In 2024, CIS claimed to debunk a popular 2022 CATO study used by many media outlets.

https://cis.org/Richwine/Catos-Brazenly-False-Claim-About-Our-Illegal-Immigrant-Crime-Research

I can't seem to find any CATO study from 2022, but they published them almost every other year it seems, including 2018, 2024, and 2026:

2018

https://www.cato.org/publications/immigration-research-policy-brief/criminal-immigrants-texas-illegal-immigrant

2024

https://www.cato.org/blog/illegal-immigrants-have-low-homicide-conviction-rate-setting-record-straight-illegal-immigrant

It might be interesting to look back at the CIS original research itself, rather than just their "debunkings" of CATO  Their original research looks as hacky, if not more, that that from CATO.

https://cis.org/Report/Misuse-Texas-Data-Understates-Illegal-Immigrant-Criminality

While CIS isclaiming CATO is fudging the numbers, guess what CIS is doing.  Fudging the numbers.  They are claiming illegal immigrant status is undercounted, but claim to have come up with a trick to handle it...the time required for greater convictions means a higher rate of correct identification in the most serious crimes.

https://cis.org/Report/Misuse-Texas-Data-Understates-Illegal-Immigrant-Criminalit

Which brings to my mind a problem with this whole category of studies.  What people are really interested in is not the rate of "convictions."  They are interested in the actual rate of crimes.  Now I believe it's virtually certain that illegal aliens have a higher convinction rate among those who actually did the crime.  They are much more likely to get caught because of racial bias and exposure, and more likely to get convicted because of lack of connections and representation compared with citizens, including being more likely to get falsely convicted.

Everyone knows by now most crimes are not even investigated, much less solved.  What happens is that people get caught in the "justice" system somehow, such as with an immigration violation, and then their prints and gun numbers can be run through the system.  Crimes without that kind of exposure never get solved.

Then of course well connected people can often get off, that too.  That would certainly affect the most serious crime convictions.  Everyone knows a good lawyer can get you off, but not so much a public defender.

All things considered, I believe the excess-conviction-rate-relative-to-citizens (ECRRC) is many times higher than 1, such as 3-20, dwarfing all the concerns of CIS and others (and, in fact, enabling their narrative)...

But then CIS numbers for the illegals in the most serious violent crimes also differ from everyone elses, including CATO and the NAS below.  So that's worthy of checking too.  It could be cherry picking somehow.

Possibly the most authoritative source we can have in USA is the National Academy of Sciences, which includes the most highly rated scientists in the USA, and all of their reports are peer reviewed.  And I can tell you because I've been in science that the NAS is politically diverse, it's not just "libruls."  Even academic science relies on scientific entrepreneurs who become Principal Investigators, they make a lot of money, and tend to be conservative, just like medical doctors.  And it's those "top" people who tend to get into NAS.  It's not just one party involved with appointing the people who appoint the funders of science either, and that determines who become the top Principal Investigators and who fall into other roles.

The NAS weighed in with a paper in 2020 showing that illegal immigrants commit fewer crimes.

https://www.pnas.org/doi/10.1073/pnas.2014704117

Now, of course, they used a "model" to determine the number of illegals in any of the crime categories because of the incomplete identification problem described by CIS.  A model that was peer reviewed and available to public (as is the PDF of the original paper here).

That's the standard scientific approach, and they haven't seen fit to update that research.  I note that it was published during the first Trump administration, and scientists are well aware of such things.

That seems to be that's the best we know right now.

And the most respected mainstream media outlets like the NYTimes and WashingtonPost and all such follow it, with exceptions exceptions (perhaps quoting CIS) for FoxNews and the like.

https://www.washingtonpost.com/politics/2024/02/29/truth-about-illegal-immigration-crime/

They link to CATO, CIS, and NAS.

Still, it's a measure of convictions, not actual crimes, which I think are far more undercounted for the citizens. 

Pressing the Red or Blue buttons

 Mr McBeast says you should pus the Red Button, Never the Blue Button.

I don't think the answer is obvious.  You'd want to do a poll to see what other people think, or be part of the "Blue Button" movement to ensure 51%, rather than treat this as personal choice which could be a big mistake.  Or (defending the blue button choice) you don't want to live in a world where half of the people could be wiped out for pressing the wrong button.

I'm don't worry about solving such problems.  Real life versions of Prisoner's Life dilemmas are never so clear cut.  There are limitations, subtle effects, etc.

Thus it is with most electoral voting.  I recognize the Reason editors choice (never voting) as reasonable but I have a somewhat different take.

In addition to "the result" which one has only an astronomically small (at least 1 in a 10 million or maybe 1 in a quadrillion) chance of affecting, there are "subtle" effects (which may be very personally meaningful):

1) A greater or lesser margin of victory has significance which communicates to politicians and voters.

2) Similarly, a greater proportion of 3rd party voters (with no appreciable chance to win) or non-voters has significance.

3) Changes or rates of change of party choice has significance.

4) Whoever wins is likely to be bought by some part of the ruling class.  In no case is there going to be an end to excess profits or needless wars or general enshitification.

5) Talking about your choice could have some influence on others.

6) Talking about your choice makes you a greater or lesser part of the virtual club of people you know.

7) Expressing your real feelings makes you feel good.

8) Early collapse may be better than later collapse, etc.*  There might be more survivors from an earlier collapse, etc

9) Movement building solidarity if you are part of a movement.

10) Voting against your usual party communicates dissatisfaction.  If it leads to a loss, that is a form of "discipline" which might force the party to be better, or go down the tubes and be replaced by a better one.  (Or worse, etc)

I've decided that in view of everything else, #7 is central and most important.  There's never a good enough reason not to vote for the candidate you like more than others, or dislike least.  And IMO that's the way voting should be.  It's both saying what you feel, and communicating it, even if not through the singular "choice" of an allegedly "winner take all" system.

So the Reason editor should vote for (I'd assume) the Libertarian candidate, unless they though the Libertarian party was taking a bad turn recently, then they'd abstain.

They shouldn't not vote just because the Libertarian candidate won't likely win.  (Unless the cost of voting were a significant factor.). That's giving up an opportunity to say what you feel and be a comrade with your closest movement.

Also, the chance that you could still be persuaded is another power you have.  So it's never useful so say what you would never or always do.  So "I'll always vote for the Democrat" means the Democratic party has no need to improve and will just get worse as it surely gets bribed mostly that way.  "I'll never vote for the Democrat" also means the Democratic party has no need to improve.  Better to make it conditional and say something like "I'll vote for a Democrat when they oppose the genocide in Gaza (or perhaps even just call it genocide)."

*Collapse could mean many things.  Such as ecological collapse or dissolution of the USA.  In general we want to put catastrophes as far off into the future as possible, but we have no way of knowing whether that will actually be the best for most people (or animals, ET's, etc)

LLM's vs Symbolic Reasoning

 I am not a fan of "AI." Ironically, I took a class in "AI" in 1983, and "neural networks" as the earliest versions of LLM's were called, were barely mentioned. We read and wrote programs in Lisp. I declined an offer to work in an AI lab in preference for more solid engineering of CAD/CAM, and later learned more about Neural Networks from an ACM lecture in 1986.  I generally viewed them as for perception only.  In 1992, I watched a PBS Nova presentation about the Cyc computer project, to compile all knowledge symbolically.  Much work was being done in Texas.  (Wikipedia doesn't show Cyc as having been updated since 2017.). That sort of thing was what I would have bet on.

I've always believed in symbolic reasoning to be essential, and now there are papers to prove it. LLM's can't solve simple problems because they lack symbolic reasoning and planning skills. LLM's are primarily good at pattern recognition, the original problem, since programs couldn't easily be written to understand things like speech or text. Symbolic reasoning can actually solve these problems, and uses orders of magnitude less energy to do so.

https://open.substack.com/pub/garymarcus/p/even-more-good-news-for-the-future?utm_campaign=post-expanded-share&utm_medium=web

https://en.wikipedia.org/wiki/Cyc