English Language & Usage: Multi-Layered Discourse Room

Robusto

00:02

@Cerberus What vacant eyes he has.

Cerberus

00:19

Yeah or maniacal.

Robusto

I don't believe in souls, but most people at least look like they have one. That guy looks like his was recently extracted through his anus.

tchrist

Pod person or Stepford wived? Your call.

Cerberus

00:39

He could be a poorly skinned robot.

CowperKettle

The best place to teach physics is at the edge of a cliff, because students have a great potential there.

@alphabet No. I'll look it up!

Robusto

@tchrist Even Gollum looks like he has more life.

CowperKettle

Price and worth are synonyms, but priceless and worthless are opposites

Robusto

01:00

@tchrist: ^

Still snow on Sandía.

That was Friday, but even today there were still patches.

CowperKettle

Fake but funny

Robusto

But today was the first really warm day this year. Temp hit 82 °F (high 20s °C).

CowperKettle

Starting tomorrow, it will be about 20 C here.

Cerberus

Hot!

I think we had 16 or 18 today, also very warm.

Robusto

Riding my bike in shirtsleeves!

Cerberus

01:05

Scandalous!

CowperKettle

Going for groceries in underpants!

tchrist

@Robusto The entire Continental Divide up in Indian Peaks Wilderness is pure white. Friday I think it was we got snow on our in-town mountains above about 6300 feet. It came down the hillside behind me a couple hundred feet, and the higher Boulder peaks at 7 to 8,000+ are quite snowy.

@CowperKettle You don't normally wear underpants at the grocery store?

CowperKettle

Groceries come into our house through the door, but leave through the pipeline.

@tchrist I'll take the fifth

tchrist

Nor the liquor store, either, apparently.

Robusto

@tchrist Yeah, but you are far north of here. The high passes in the Rockies don't usually open till the middle of June or later.

tchrist

01:07

The high passes of the Rockies are not my in-town mountains. Those are at the Divide.

Certainly Trail Ridge Road and Mount Evans' Road aren't open all the way yet. Nor shall be for another month or more.

@Robusto We hit 68. It felt surprising warm exercising in the sun. Most guys weren't wearing shirts.

The old people still wore coats, of course. :)

It felt like 83 in high sun. No breeze.

Robusto

@tchrist Not me. I was in summer kit on my bike. And in shorts and flipflops at home.

tchrist

But the sun just now headed for the dark place, and you immediately feel cooler.

Robusto

Yeah, funny thing about that. But everything is blooming and sprouting and most of all the birds are singing the praises of summer.

tchrist

It was 42 when I woke up.

Robusto

Still 80 here right now.

tchrist

01:12

58 now at NCAR. 67 downtown.

Suddenly-ish.

CowperKettle

I wake to sleep and take my waking slow.

tchrist

40 tonight. No pollen whatsoever. It knows not to be too exhuberant.

Robusto

> And I shall have some peace there, for peace comes dropping slow,
Dropping from the veils of the morning to where the cricket sings;
There midnight’s all a glimmer, and noon a purple glow,
And evening full of the linnet’s wings.

tchrist

But I got buzzed by a hummer today.

Robusto

Bumblebee?

Humminbird?

CowperKettle

01:15

> What is more gentle than a wind in summer?
What is more soothing than the pretty hummer
That stays one moment in an open flower,
And buzzes cheerily from bower to bower?

tchrist

@Robusto Broad-tailed hummingbird, by whose sound I identified him as male and presumably he does as well. The ladies show up a week or two later.

Robusto

@tchrist So he's got to get the domicile sorted by then?

tchrist

Their primaries make a special sound unlike the normal buzz. It has a high pitch.

@Robusto Yeah, the males scout out food sources.

Robusto

I can't believe I typoed hummingbird.

I'm doing that a lot lately. Somewhat disconcerting.

tchrist

But we still have a fair chance of getting a bit more snow for the next month. It may not freeze again, but it might.

Average last frost is like the 11th to 15th of May, depending.

But that's down here in the balmy altitudes.

The broad-tail looks pretty much like a ruby throat from back east.

But it has a magical sound.

Robusto

01:21

@tchrist Your last frost date is 90% certain to be April 24.

Ours is March 30.

tchrist

earliest day of last frost:  Apr   3 (2012)
latest day:                  Jun   3 (1951)

average:                     May 5

median                       May 3

standard deviation           11.6 days

Robusto

The Pleasures of Merely Circulating
by Wallace Stevens

The garden flew round with the angel,
The angel flew round with the clouds,
And the clouds flew round and the clouds flew round
And the clouds flew round with the clouds.

Is there any secret in skulls,
The cattle skulls in the woods?
Do the drummers in black hoods
Rumble anything out of their drums?

Mrs. Anderson's Swedish baby
Might well have been German or Spanish,
Yet that things go round and again go round
Has rather a classical sound.

tchrist

from here

A standard deviation of 11.6 days is a pretty big swing.

Robusto

Indeed.

CowperKettle

Last days of spring in Boulder
May dislocate your shoulder

tchrist

01:25

You're supposed to do a backflip.

Robusto

In England the weather is coulder.

Morning Chores frost date finder.

tchrist

Says Star Wars Day.

On verra.

Put your tomatoes in before Memorial Day just makes more work for afterwards. It snowed here last year then.

CowperKettle

> Koryaksky or Koryakskaya Sopka (Russian: Коря́кская со́пка) is an active volcano on the Kamchatka Peninsula in the Russian Far East. It lies within sight of Kamchatka Krai's administrative center, Petropavlovsk-Kamchatsky.

tchrist

We've gone up at least one USDA zone since I moved here in my salad days.

CowperKettle

Imagine having this view from your house, on a 3-km volcano.

CowperKettle

02:05

alphabet

02:21

Vowel merger of the day: the meet-meat merger. Occurred in Middle English; now found in nearly every dialect except a few rare ones in Britain and Ireland.

Cerberus

02:56

@CowperKettle "Arrive over": do they mean Ukraine's warning systems predicted the launch of the missiles seven hours in advance?

Vikas

04:00

True?

Cerberus

Haha are those straws?

I don't think India resells oil to Europe?

But I know there is illegal reselling of Russian oil.

CowperKettle

05:02

@Cerberus Probably there were some indications

Vikas

05:47

@Cerberus Yeah

CowperKettle

06:06

@Cerberus Reportedly, some open-source intelligence mavens caught some radio signals or something, indicating some planes taking off

Neil deGrasse Tyson Is Not Afraid of Artificial Intelligence | Within Reason #28

►

Cerberus

@CowperKettle Planes taking off seven ours in advance?

CowperKettle

@Cerberus Yes, I was also amuzed

Cerberus

Maybe they took off and landed closer to the border?

CowperKettle

There are a lot of OSINT accounts on Twitter, but I only glance at them once a day, I grouped them off into a special folder, in order not to get distracted.

@Cerberus Maybe :)

I wonder how ketogenic diet affects people with diabetes. It does not require carbs. So.. would it have saved diabetics if applied prior to the discovery of the first insulin manufacture process in 1921, or would it have killed them faster?

Some psychiatrists come up with ideas of using the ketogenic diet in psychiatric disorders onlinelibrary.wiley.com/doi/10.1002/brx2.6

But the body clearly has not been developed by the evolution to follow this diet.

> Palmer CM, Gilbert-Jaramillo J, Westman EC. The ketogenic diet and remission of psychotic symptoms in schizophrenia: two case studies. Schizophr Res. 2019;208:439-440. pubmed.ncbi.nlm.nih.gov/30962118

> She tried the following medications over the course of treatment: haloperidol, clozapine, ziprasidone, risperidone, quetiapine, aripiprazole, olanzapine, sertraline, paroxetine, citalopram, fluoxetine, duloxetine, and venlafaxine. Her psychotic and mood symptoms persisted despite these medication trials. In 2013, she was started on a ketogenic diet by a functional medicine practitioner for symptoms of chronic GI distress.

> [..] Within one month, she reported complete resolution of her psychotic symptoms for the first time since 1993.

> She was tapered off haldol-decanoate over the following year, and has remained free of psychotic symptoms for the past 5 years off of antipsychotic medications.

CowperKettle

08:14

Two women have been detained in a park near my house, for this sign.

A third woman has been detained in the central square with a similar sign:

t.me/ovdinfolive/20828

CowperKettle

08:39

> A 2017 study found that Araucaria columnaris tend to have a tilt dependent on the hemisphere of their location, growing upright on the Equator but leaning south in the northern hemisphere and north in the southern hemisphere

> It was first classified by Johann Reinhold Forster, a botanist on the second voyage of Captain James Cook to circumnavigate the globe as far south as possible. It is named Cook's pine directly for Cook, and not for the Cook Islands.

CowperKettle

12:02

‘The Godfather of A.I.’ Leaves Google and Warns of Danger Ahead. For half a century, Geoffrey Hinton nurtured the technology at the heart of chatbots like ChatGPT. Now he worries it will cause serious harm.

> Dr. Hinton said he has quit his job at Google, where he has worked for more than decade and became one of the most respected voices in the field, so he can freely speak out about the risks of A.I. A part of him, he said, now regrets his life’s work.

CowperKettle

13:35

Partially blocked one.

@Vikas Probably yes, but it was expected. Somebody would do it anyway.

Russia has remained surprisingly strong. I can see no end to this totaly useless fighting. The world needs attention to dwindling resources, overpopulation, aging, vanishing species, global warming, plastic pollution, meteor danger, etc., etc., and not this 19th century stuff.

Robusto

14:03

#Worldle #465 1/6 (100%)
🟩🟩🟩🟩🟩🎉
⭐⭐🏙️
https://worldle.teuteuf.fr

17-YEAR-OLD STUDENT PILOT LANDS HER PLANE WITHOUT A WHEEL!

►

How did that aircraft ever get cleared at inspection?

Didn't she do a preflight? She was a student solo, her instructor should have supervised that.

Poor kid, she (rightly) sounds terrified.

CowperKettle

14:19

M.A.R.

@CowperKettle I don't know where I'd be without this guy's encompassing wisdom

@CowperKettle the pigeon is an advanced Western drone given to Ukraine to drop an advanced twig bomb on that pond or whatever that is

@CowperKettle killed them faster. Ketoacidosis is no joke. Every other myth about this diet is just that, a myth. It's only helpful in epilepsy.

@CowperKettle an interesting case but just a case nonetheless.

I dunno how metabolic acidosis would help schizophrenia. Or weight loss. Or reduced visceral adiposity.

It's a very difficult diet to follow, so self-reports of adherence are not very reliable

Robusto

14:34

@M.A.R. I very much enjoyed his Sapiens.

M.A.R.

@Robusto I feel like people like him should be held to a higher standard, is all.

Robusto

What standard would that be?

M.A.R.

@Robusto saying "AI has hacked civilization" would not meet it.

Robusto

@M.A.R. You don't know that he wrote that headline. Probably he didn't.

M.A.R.

Maybe. But he's fond of this tabloid-style . . . scientifism?

Or I'm just grumpy after a difficult day

Mitch

14:49

@Robusto "A.I.’s new mastery of language means it can now hack and manipulate the operating system of civilization"

@M.A.R. Why not both?

(hopefully that link bypasses paywalls)

Robusto

@Mitch I can't get there from here.

Mitch

> In the beginning was the word. Language is the operating system of human culture. From language emerges myth and law, gods and money, art and science, friendships and nations and computer code. A.I.’s new mastery of language means it can now hack and manipulate the operating system of civilization. By gaining mastery of language, A.I. is seizing the master key to civilization, from bank vaults to holy sepulchers.

M.A.R.

@Mitch the fecking exam! AAARGH

Robusto

@Mitch OK, did you read the entire article?

Mitch

Yes, last month when it came out.

Robusto

14:51

That is certainly the provocative lead-in.

@Mitch How would you summarize his argument?

Mitch

I found it hard to distance this article from the rest of the uninformed AI hyperbole, both the doomsday revelations and the science fiction nirvana.

Robusto

So you're summarizing your reaction.

Mitch

@Robusto Yes

Robusto

I'm not arguing one way or the other. I'm just curious what he argues.

Mitch

rereading

(it's an opinion article so any argumentation is by association, not necessarily syllogism.)

> For thousands of years, we humans have lived inside the dreams of other humans. We have worshiped gods, pursued ideals of beauty and dedicated our lives to causes that originated in the imagination of some prophet, poet or politician. Soon we will also find ourselves living inside the hallucinations of nonhuman intelligence.

Robusto

14:57

He got a lot of flak for his Sapiens from a lot of people who thought he was writing a historical treatise when he was actually writing a book of philosophy based on premises from accepted historical information.

Mitch

It's pretty ... airy.

Robusto

@Mitch Well, it seems like it may be an interesting point.

Mitch

It makes me feel like ... is this a perfume commercial?

Robusto

Perhaps you are bringing more to it than is there?

Mitch

@Robusto uh... isn't how a lot of this works?

Robusto

14:59

Define "this"?

Mitch

Reading

Robusto

Well, you always bring something to reading, but ideally you would be open to taking something from it as well.

Mitch

The writers are assuming quite a lot.

Robusto

That's the problem with the AI discussion: people jump to conclusions before they examine the evidence. I've been guilty of that, but now I'm trying to keep an open mind, uninformed by fear or my own ignorance.

Mitch

@Robusto What I take from the opinion piece is that he (and the other two authors) don't have a good grasp of the technology. There are a number of (non-factual) things they say that I agree with (eg we should be concerned with what people do with generative models).

Robusto

15:05

OK, that sounds like what I was asking you to explain.

Thank you.

Mitch

@Robusto Just so you are aware (if you ever get a chance to read that article in full) there is no evidence in it to examine. They're just asserting things.

Robusto

OK. So we could call it a "thought experiment"?

Mitch

(I mean -everybody- is only ever asserting things when they write)

M.A.R.

@Robusto personally, I'm always skeptical of "whole new thing" perspectives, because real, practical stuff is almost always gradual. I don't wanna know about how AI is this whole new thing, I want to know how it is already helping companies scam me and steal my stuff, enhancing methods that were already there.

Mitch

@Robusto or polemic

Robusto

15:07

Perhaps it is that. Or perhaps it partakes of both.

M.A.R.

"whole new thing" is once every 400 years. And we've already discovered gravity.

Mitch

@M.A.R. I'm pretty sure people were aware of gravity earlier

I mean they weren't just floating around and then newton said 'aha' and we all fell to the ground.

haha

hah

Robusto

@M.A.R. And relativity. And antibiotics. And many, many things in your own field, including wholly new vaccine models.

M.A.R.

It has an air of ignorance about it, to speak of breakthroughs as spontaneous

@Mitch see?

Mitch

By your calculation we're up for another earthquake right about now.

why not AI?

why not the Matrix?

M.A.R.

15:09

@Robusto but those new vaccines improved on older vaccines, which improved on even older vaccines, which improved on whatever they injected people with, which was pus and things like that

Robusto

@Mitch That is a disingenuous statement masquerading as an argument. His idea of gravity pertained to planetary motion, not apples falling on your head.

Mitch

> Soon we will also find ourselves living inside the hallucinations of nonhuman intelligence.

M.A.R.

@Robusto didn't he also admit later that he was trying to be poetic by the whole apple fall story?

Mitch

@Robusto Yes, I am aware of both my disingenuosity and what Newton did.

Robusto

@M.A.R. Sounds like we are at odds in our definition of new.

Cerberus

15:12

@CowperKettle Did he work at Deep Mind when Google bought the company?

Mitch

> Soon we will also find ourselves living inside the hallucinations of nonhuman intelligence.

THat's just horseshit talking.

Robusto

@M.A.R. I don't know that, but what if he did? What bearing does that have on all "this"?

Cerberus

Because that is mainly where Google's 'AI' stuff comes from.

Mitch

I do that kind of talk all the time.

CowperKettle

Amtolmetin guacil

Amtolmetin guacil is a non-steroidal anti-inflammatory drug (NSAID). It is a prodrug of tolmetin sodium. == Background == Tolmetin sodium is an approved NSAID, marketed for the treatment of rheumatoid arthritis, osteoarthritis and juvenile rheumatoid arthritis. In humans, tolmetin sodium is absorbed rapidly with peak plasma levels observed 30 min after administration. It is eliminated rapidly with a mean plasma elimination t½ of approximately 1 hr. The preparation of slow release formulations or chemical modification of NSAIDs to form prodrugs has been suggested as a method to reduce the ...

Robusto

15:13

@Mitch OK, I see you've made up your mind on this. I would very much like to read the article and see if I agree or what I might imagine is his argument.

CowperKettle

A gastroprotective NSAID O_O

M.A.R.

@Robusto I should, if I'm curious, find out what exactly we had before these chatbots, and what we will have after, and only then would I know where these bots stand. Not by daydreaming about some VR dystopia or whatever.

Mitch

"hallucinations of nonhuman intelligence." - that's just arty talk for made up bs ... we already listen to other peoples hallucinations.

Cerberus

@CowperKettle Ohh where can I can I get this?

Robusto

@M.A.R. OK, but "daydreaming" can lead to discovery as much as, or perhaps more than, rigid adherence to accepted wisdom.

CowperKettle

15:15

@M.A.R. Maybe those two patients had some rare mutations/metabolic derangements manifesting as schizophrenia, and by chance got lucky with the diet?

Mitch

Anyway talking about random collections of words statistically inclined to be like what other people say can only -very- loosely be called hallucination.

And also implies some kind of alien intelligence that just isn't the case.

If there's any meaning in these random words it is what people are giving to them.

CowperKettle

@M.A.R. I also don't like Harari - I would liked him if he kept some scientific advisor to weed out the most blatant mistakes from his books. That would have made his books so much better.

M.A.R.

@CowperKettle there is no such thing

CowperKettle

He has a sweeping style, but he needs someone to review and fix his writing.

@M.A.R. Ah

M.A.R.

Fewer ADR yes, but gastroprotective? Get out of town

CowperKettle

15:18

> Amtolmetin guacil stimulates capsaicin receptors present on gastrointestinal walls, because of presence of vanillic moiety and also releases NO which is gastro protective.

Mitch

@M.A.R. To be fair to him, he is already out of town.

M.A.R.

NSAID-induced peptic ulcer has two mechanisms: One is local, mitigated by better formulation or better drugs. The other is systemic, and related to drug action: Inhibiting cyclooxygenase will mean less gastric mucus.

Cerberus

@M.A.R. What is ADR?

Mitch

@M.A.R. wait... really? aren't there some NSAIDs that don't 'do' things until they reach the small intestines? (good for colitis and ... IBS and such)?

M.A.R.

You can't do anything much about systemic mechanisms of NSAID GI effects. The most selective agents are still not safe. So it would be a comparison of how much capsaicin receptors would alleviate lesions

@Cerberus oh, sorry, I'm suffering from post-exam acronymosis. Adverse Drug Reactions

Cerberus

15:22

Ah, merci.

How did it go?

M.A.R.

@Mitch oh those aren't NSAIDs. They're just antiinflammatory.

You might ask why if it's antiinflammatory and not a steroid, we don't call it a non-steroidal antiinflammatory drug. To that I say, I have no idea.

Mitch

@M.A.R. So they're steroidal?

M.A.R.

@Cerberus well, it was really difficult. I did my best but will probably get around 34 out of 46

Cerberus

Will that be enough?

M.A.R.

Sure. I've never failed a course in uni

Cerberus

15:25

Great.

Mitch

@M.A.R. mesalazine?

Cerberus

I failed only one. I didn't study hard enough.

M.A.R.

@Mitch the point is we just call some select drugs NSAIDs. There's no criteria, and certainly not every antiinflammatory drug outside steroids is called an NSAID.

@Mitch yep

For example, we don't call every antioxidant an NSAID, though theoretically most should help with inflammatory conditions.

Mitch

@M.A.R. Pardon my literalness, but that is an anti-inflammatory and it is not a steroid, so I would be inclined to categorize it as a non-steroidal anti-inflammatory.

M.A.R.

Tomatoes aren't NSAIF

Mitch

15:28

@M.A.R. That's because they go in salad and aren't sweet.

M.A.R.

Adalimumab is a fricking monoclonal antibody. Calling it an NSAID is just misleading

(Used in IBD and Rheumatoid Arthritis)

Mitch

@M.A.R. OK but mesalazine is like a weird cousin of aspirin, so it would make total sense to categorize it as an NSAID.

@M.A.R. Yeah totally understood

M.A.R.

So NSAID is just mostly code for "common analgesics"

Mitch

oh

but

but not mesalazine

M.A.R.

@Mitch maybe because it's always formulated to have local effects. So someone with some tension headache can't use mesalazine

We're more inclined to list acetaminophen with NSAIDs for this reason (although we don't since it's not antiinflammatory)

Mitch

15:32

@M.A.R. got it. Scientific language isn't always one-to-one onto

but it's annoying when it's not.

M.A.R.

@Mitch I've reached the conclusion that in pharmacotherapy, the pharmaco part often makes sense and the therapy part is a hot mess.

Robusto

@Mitch Perhaps you're right. But I think if the words were truly random they wouldn't be useful, would they? People wouldn't be able to use them for school assignments or to assist in programming tasks.

Mitch

@Robusto If they were truly random it would more likely look like gibberish.

Robusto

My point exactly.

Mitch

The less random it is the more likely it will look like what has already been written.

M.A.R.

15:36

"statistically arranged words"?

Mitch

thousands of monkey's typing eventually will recreate Shakespeare (but also many near copies of Shakespeare and many bad reformulations of Shakepeare and etc and etc). These LLMs just mean it'll take fewer monkeys and a shorter amount of time.

But any 'new' things, any new literature? Of course it'll create those too, (and lots of poor copies of them with variations).

How would we know that a particular production (of these more constrained monkeys) is actually Shakespeare worthy?

M.A.R.

@CowperKettle doesn't seem like a new drug, but seems it has never been introduced to most markets? No entries in UpToDate. Maybe it turns people inside out or explodes their heart or something. When they tried to improve celecoxib to get more selective drugs (e.g. rofecoxib), they caused heart attacks and were quickly withdrawn. Now nobody would touch them with a ten-foot pole

CowperKettle

@Mitch Maybe if someone comes up with a structure that will unite several specifically-shaped neural networks, it will work "a bit like a human mind"?

Robusto

@Mitch I think if you look at the odds it would take trillions of monkeys trillions of years to accidentally create even a single one of Shakespeare's plays.

Mitch

@CowperKettle 'a bit like' is doing a lot of work here.

CowperKettle

15:43

I mean, we also have these networks in our brains, but there's some structure that is correcting things.

Mitch

There's a loose analogy between biological neurons and the matrix ops/thresholding that is done in neural networks but it doesn't get much better than 'loose analogy'

CowperKettle

So one needs to recreate actual dendrites, dendritic spines, and so on?

Mitch

You could say the same thing about a new NN architecture - they might be inspired by some biological knowledge. Or even psychological knowledge.

M.A.R.

Oh, BTW, they're impeaching the minister of industry here. Real exciting stuff. The replacement would surely propel economic gro . . . Dollar prices.

Mitch

@CowperKettle Oh...no.. I'm not saying that.

CowperKettle

15:45

A new artificial intelligence system called a semantic decoder can translate a person's brain activity—while listening to a story or silently imagining telling a story—into a continuous stream of text.

> The work relies in part on a transformer model, similar to the ones that power Open AI's ChatGPT and Google's Bard.

Mitch

There's a lot of philosophical controversy about what is possible. I was only talking about what is happening today.

M.A.R.

@CowperKettle this stuff is also puzzling for me. Is it EEG-based?

Because to my knowledge EEG has mostly remained a dormant technology

CowperKettle

> For example, in experiments, a participant listening to a speaker say, "I don't have my driver's license yet" had their thoughts translated as, "She has not even started to learn to drive yet." Listening to the words, "I didn't know whether to scream, cry or run away. Instead, I said, 'Leave me alone!'" was decoded as, "Started to scream and cry, and then she just said, 'I told you to leave me alone.'"

@M.A.R. No, MRI-based, and is quite not very precise yet, and requires "fine-tuning" for each specific person

> The paper describes how decoding worked only with cooperative participants who had participated willingly in training the decoder.

M.A.R.

. . . Which might as well be two steps removed from fabricating data huh

Mitch

As to what's possible, sure, like @Cerberus has intimated, one could theoretically simulate the action of molecules in nerve cells, and the collection of nerve cells arranged in an architecture exactly like someone's brain. And this would simulate human intelligence.

(that is philosophically questionable, but I am on the side of simulatability at any level of brain/psychology organization)

The difficulty is that we just don't know enough about -any- of the levels of brain/psychology organization to simulate.

Cerberus

15:51

@Mitch And how about creating a different, un-human-like intelligence? Do you agree that true intelligence can be created as an emergent feature of some extremely complex combination of data and algorisms?

Emphasis on emergent.

M.A.R.

What threshold would have to be crossed for intelligence?

Mitch

1) that's amazing tech if they can do that
2) I find it highly unlikely that the headline captures what the capabilities are (all similar headlines I've seen recently mislead to the point of lying)
3) people don't think in a 'streams of language' (we are fooled into thinking so because we use a stream of language to tell others what our thoughts are)
4) There are a lot of people we'd all rather not know what they're actually thinking.

M.A.R.

We are (astronomically) far from both

Robusto

@M.A.R. One that is all too infrequently crossed even by humans.

Cerberus

@M.A.R. That I do not know.

@Mitch Point 3 is certainly true. But does that mean no intelligence can be created based on these streams of language?

Robusto

15:56

Maybe we should define intelligence at this point, since we're all talking about it.

Mitch

But it is plausible to me that transformers (the NN architecture behind ChatGPT, etc) create a good enough language model that one could use it to convert brain patterns to text (and that text corresponds to some brain internal stream of consciousness meaning like the recitation (internally/imagined) of some text.

@M.A.R. without looking at the paper, it is plausible. technologically they have to take these first tiny steps that seem like nothing.

M.A.R.

@Robusto Getting to bananas out of reach? Using a stone to crack walnuts? Proving 1 + 1 = 2?

Mitch

@Cerberus Totally I believe that. That it is possible. I totally believe that these LLMs are not even slightly that at all.

But 'emergent' is doing a lot of work here.

Robusto

@M.A.R. Those are products of intelligence, but can we define what that motivating principle really is?

Mitch

@Robusto yes. That joke is often taken too seriously by people trying to sell AI

M.A.R.

16:01

Cracking walnuts with bananas. Self-interest?

Robusto

So it is the motivating practice of an organism to better its condition?

M.A.R.

But then is self-harm unintelligent per se?

Cerberus

@Mitch OK, sure. Not these. But how long until one could appear?

Mitch

@Cerberus I'm not sure if I've parsed/understood what you mean by that as intended but there is a big controversy over what intelligence is in language alone (ie can you get an intelligent agent by training solely on language and not using any other senses). This is called the 'grounding' problem... do you need to be grounded in a full sensory environment (sometimes also called 'embodied cognition) to be fully intelligent.

Some people say there is no intelligence in just the strings of text - that you need to have the language connected with an environment to have any intelligence in it.

Some people say that text is enough.

M.A.R.

If someone knows smoking is bad but continues to do so, it's probably not unintelligent. But bots don't have desires. So maybe it is if chatGPT smokes.

Cerberus

16:06

@Mitch I understand that argument.

Mitch

And as usual as with all these philosophical arguments is that there is something between the two extremes. (or rather -I- think this way)

Robusto

@M.A.R. I suppose it depends on other factors? The guy who cut off his arm to escape death may have been acting intelligently, right?

M.A.R.

All good and dandy until someone comes and says it's a moderation fallacy

Rinse and repeat

Mitch

There is a lot of info (logic, inference etc) -in- the literally blind text, but its meaning would be much closer to what we think of meaning if it had been in an environment of seeing and living with other ... bodies.

CowperKettle

@M.A.R. Human desires arise out the necessity to copy the DNA to the next generation. What if a bot's desires arise out of the necessity to have its computer code copied on other computers and used for as many CPU cycles as possible?

Mitch

16:09

But there is definitely a trend to add this kind of extra sense data (or at least meaning data)

Robusto

@CowperKettle That's a good point. Perhaps we need to step back from discussion of individuals to the real motivating principle: life recreating itself in whatever form it needs to keep doing so.

Mitch

The LLM intelligence stuff is very misleading. It's mostly people filling in the gaps and giving the output meaning, ascribing intention to what is literally statistically likely text.

Cerberus

@Mitch But consider this. We have no idea what our environment is like. Our brains merely get impulses from what we think might be our visual organs. Those impulses are merely bits of data that our brains interpret. And the interpretation is never tested by God; its only test is whether it is *useful* to us (related to pragmatism in philosophy).
So is there really any fundamental difference between receiving some electrical impulses from a source basically unknown to our brain on the one hand, and huge numbers of linguistic impulses from some other source?

I mean, yes, the computer will not understand the outside world the way we do. And it may forever continue to make mistakes that we won't, because of this. But it will understand the world in a different way. Perhaps one that is functional enough?

Mitch

@M.A.R. The mirror test? Getting out of a sheet wrapped around you?

M.A.R.

I'm unintelligent on some winter nights

Mitch

16:13

haha that last one is actually part of a dog intelligence test... some dogs pass it (obviously border collies and other shepherd dogs) some don't (beagles? bulldogs?)

or those dogs on leashes that go around a pole and can't figure out that they need to go back around.

or maybe they're just being willful

Robusto

@Mitch Some humans are like that. A lot of them.

Mitch

@CowperKettle That's the 'paperclip factory' idea.

CowperKettle

@Mitch Maybe the oxygenation event on Earth was the result of the runaway paperclip factories

I mean the primitive organisms that produced a lot of oxygen as their waste product, and dropped the temp on the whole planet

Robusto

@Mitch That sounds rather condescending, to sum it up in a ~~flip~~ ~~clip~~ glib metaphor.

CowperKettle

Great Oxidation Event

The Great Oxidation Event (GOE), also called the Great Oxygenation Event, the Oxygen Catastrophe, the Oxygen Revolution, the Oxygen Crisis, or the Oxygen Holocaust, was a time interval during the Paleoproterozoic era when the Earth's atmosphere and the shallow ocean first experienced a rise in the amount of oxygen. This began approximately 2.460–2.426 Ga (billion years) ago, during the Siderian period, and ended approximately 2.060 Ga, during the Rhyacian. Geological, isotopic, and chemical evidence suggests that biologically-produced molecular oxygen (dioxygen, O2) started to accumulate in Earth...

Mitch

16:20

@Cerberus I think there is some 'signal' in a text if seen as a string of bits (that is I don't think there is none at all) but it's not much. I can see all sorts of sentences like "The dog bit the man" and tell that it's usually this 'dog' that has some sort of 'biting' relation to the other thing, the 'man'. But as of this moment I'm starting to doubt if one could even extract the concept of noun or verb from plain text (ie without a human tagging lots of text with parts of speech first).

Cerberus

@Mitch Maybe you don't need that concept for intelligence. Maybe you can distil some other concept that is also useful enough.

Mitch

@Cerberus Sure... maybe to act on just those bits

Cerberus

So my point is, we are basically not 'grounded' in reality either, or so it could be argued.

Mitch

@Robusto Cats use that as an intelligence test for their staff.

Robusto

Hahaha, that they do.

CowperKettle

16:22

@Cerberus Yes, because we falsely see and feel solid stuff all around, and in reality it's 99.9999999% empty space

Mitch

@CowperKettle Which I think resulted in the 'Snowball Earth'?

CowperKettle

And we don't even see radiowaves. We are so removed from reality.

Robusto

You don't?

CowperKettle

)))

Cerberus

@CowperKettle Yes, for instance!

Mitch

16:27

@Robusto I'm just reporting to you in very telegraphed way a well established metaphor that everyone in AI is aware of (and has lots of commentary behind it). It is still controversial. It's Nick Bostrum's fable.

@Cerberus Yes, that is certainly a possibility and a difficulty with assessing (this so far undefined concept of) intelligence.

Robusto

I think at some point much of what we are arguing about is unknowable. Things like "what is intelligence?" or "what [even] is reality?" or any of a number of abstractions. To some extent, intelligence may be defined as one's ability to create abstractions.

Mitch

@Cerberus I'm not sure I see exactly what you're getting at but I -think- you may be alluding to the 'brain in a vat' idea (what's the difference in intelligence between us living our lives and a brain in a vat that is getting its sensory input from stimulations on it's optical and cochlear nerves.

@Cerberus Oh... if you're saying we don't sense -all- of reality, right we don't have access to all that our sense could (the full EM range, the full sound range, all sorts of other possible sensory data that we can't even imagine.

Robusto

BTW, it's the first of May! Workers of the world, unite!

Mitch

But what we do have (that small sliver of light bandwidth) has quite a lot more meaningful data in it than petabytes of ASCII

@Robusto Yes, it is difficult to define well.

Usually when people come up with a definition, it's always got problems and doesn't match our intuition well.

@Robusto It's still not warm enough

Robusto

> I refute it thus!

Mitch

16:38

Is there a 'jealousy' emoji?

Of course there is.

@Robusto Very apropos

CowperKettle

Sorrel soup

Sorrel soup is made from water or broth, sorrel leaves, and salt. Varieties of the same soup include spinach, garden orache, chard, nettle, and occasionally dandelion, goutweed or ramsons, together with or instead of sorrel. It is known in Ashkenazi Jewish, Belarusian, Estonian, Hungarian, Latvian, Lithuanian, Romanian, Armenian, Polish, Russian and Ukrainian cuisines. Its other English names, spelled variously schav, shchav, shav, or shtshav, are borrowed from the Yiddish language, which in turn derives from Slavic languages, like for example Belarusian шчаўе, Russian and Ukrainian щавель, shchavel...

Mitch

(to ideas of intelligence and consciousness)

CowperKettle

Sorrel soup, a soup that may cause death by hypocalcemia.

> .. was admitted to hospital because of vomiting, diarrhoea, and progressive impairment of consciousness soon after the ingestion of a vegetable soup containing about 500 g of sorrel.

Mitch

@CowperKettle Or severe trauma from having a pallet of sorrel soup cans falling on you.

@Robusto I also find the 'paper clip maximizer' to be problematic, but even so it is a way to make some ideas about AI more concrete. It seems to be driving a lot of the terror mongering.

CowperKettle

Maybe the AI will not kill humans, but misshape them by greatly affecting societies.

Like humans did not kill all wolves, but fashioned them into wierd shapes.

Mitch

17:01

@CowperKettle Good point. It doesn't have to be 'total destruction' or nothing.

Robusto

Perhaps AI in itself isn't the danger—that it would act destructively out of its own volition—but that it could be used by bad actors to spread disinformation and inflame hostilities and bring about more conflict.

Mitch

But also, it doesn't have to be some science fiction AI. Just plain old automation with a simple decision rule can be a problem. (denying loan applications automatically based on zip code, predicting recidivism based on race)

@Robusto Yes, because these LLMs make it so easy to produce text of a certain kind (that was very difficult for bad actors to do before)

Robusto

The difference is of degree, not kind. We already have bad actors doing bad things, but now they have discovered AI steroids.

So we will have convincing video of Nancy Pelosi and George Soros molesting children, or something like that. Deepfake, anyone?

Mitch

Maybe some one can use a text to speech API to mimic a grandchild to call the grandparent and convince them to wire money to a bank account

Robusto

The avenues of perfidy are numberless. As usual.

Current NYT headline: "‘The Godfather of A.I.’ Leaves Google and Warns of Danger Ahead"

https://www.nytimes.com/2023/05/01/technology/ai-google-chatbot-engineer-quits-hinton.html?unlocked_article_code=fQfpMFSfvnt3X0Yse73WYutieOVOPK3c7c7Lrba94MYNVMg35AASVBbrWXjEHfFXgIJFgYnSFXKtMVklgsltch3Vw6Umw1SYQpjDyw9KECBubV2UbTOPZKn-S-12sHh2Nx2CSldTXWSkBGWQGgw2vQKL-iSn36wf-mGLm6mGMxLW6pukTZohPNL1UAwKNXy5ZDNYM9ZNXyKtDHCyl39ByVdKzinIJP1aT7Zs9jWoXiFcnNnMwC7ikKNpjlpxBT6rv4QYtdsMommF73XwtvZ1q5AfX_R8zz8QFMfW6vw6A9VdXOE0llD2_5UM7KAIohmAhIxAyBwXjMGTh-q_G86rdxvtPaweKcSMPNx6H7H1OdXCbzsz&giftCopy=0_NoCopy&smid=url-share

...

> But gnawing at many industry insiders is a fear that they are releasing something dangerous into the wild. Generative A.I. can already be a tool for misinformation. Soon, it could be a risk to jobs. Somewhere down the line, tech’s biggest worriers say, it could be a risk to humanity.

“It is hard to see how you can prevent the bad actors from using it for bad things,” Dr. Hinton said.

Which we were just talking about.

Mitch

17:21

@Robusto @CowperKettle linked to the article just above his reference to Harari.

I'm beginning to think he may have access the the NYT

I'm gonna go out on a limb here and state that what he really should be reading there is their Cooking section.

Easy to follow recipes.

Health conscious

Ingredients not too exotic (unless you want that)

Robusto

Yeah, I don't really want to know more about pharmaceuticals. But good food recipes? Nice. There, I said it.

Mitch

@Robusto Everything I know about pharmaceuticals I get from ads for medications that start with 'Ask your doctor about...'

Robusto

@Mitch Those are so helpful! Where you see people smiling and dancing and having fun while the small print says "May cause convulsions, blindness, paralysis, ugly sores all over your body, anal leakage, and death."

Mitch

They all look so happy

I want to be one of them

Robusto

Not the ones with anal leakage, though.

Mitch

17:35

Oh

yeah

good catch

unchecks anal leakage

Robusto

I remember when Nexium used to advertise the "purple pill" and showed people flying overhead with no effort. I so wanted to be one of them, but when I tried Nexium I never even got off the ground.

Mitch

Now that you mention the flying thing, maybe I'll mention Nexium to my doctor.

Also, did you say 'purple'?

That's gotta be good.

checks wikipedia

I don't see anything about 'exploding hearts', so we're good.

Robusto

Anal leakage?

Transcript for

English Language & Usage: Multi-Layer…