eureka - because I'm watching this part now and it bothers me.

So first, I'm going to tell you about one of my personal things: the Balsam Wood Test.

In SGA fandom, there's a fanfic--I cannot remember which--where in a throwaway scene, Rodney and co are testing an unknown Thingie for reactions to common substances. It reacts to nothing at all, great. Then someone throws in a piece of balsam wood, and as it turns out, Thingie reacts badly to it.

To balsam wood.

Anyway, that stuck in my head--I love that kind of thing--and eventually, the idea turned into a concept of how to reliably test reality when all you have is your subjective self to work it out. The Balsam Wood Test.

Now, Eureka's Matrix: I love it. I love it for so many reasons, but all of them are relationship and people based. I love the characters dealing with it. I love the drama around it.

I hate the fact that anyone, anywhere, would think the Matrix could, even by accident, forward the study of science as it pertains to anything but the study of artificial reality and maybe the limits of computer programming. That's not just insane, it's--I need a word here, just go with 'are you high and have been since the Enlightenment?'

It can't be done, full stop. Even if it was run by an AI, it couldn't; if the AI actually could do that, you wouldn't need a Matrix because you wouldn't need people to discover anything; the AI could do it all. A computer could not, ever, reliably reproduce science as we know it--much less Eureka-level science--well enough to fool actual scientists for more than five seconds and maybe not even then.

(I'm not entirely sure it's really possible to create a Matrix reality indistinguishable from reality-reality, but that's another story.)

You see, there's no such thing as random numbers in programming. But I'm getting ahead of myself.

We have no idea what objective reality is; it's all subjective to varying degrees. Physics tries very hard, but even things we designate laws are very much 'well, nothing has contravened this yet so here we are'; anything lighter is 'current best explanation we have but we're open to suggestions' at best. Entire swathes of science exist based on math with the understanding we also haven't discovered all the math. To be generous, we're not even aware of about 99.999999999999999999% of physics. Of the part we're aware of, we maybe are sure of none and best guess a very tiny amount of that. And i won't even begin to describe where we are in pure math; maybe a little better? But probably not.

Like, this is true of all science, but I'm focusing on physics because the Matrix kind of requires it; that's the baseline on pretty much everything.

The Balsam Wood Test: in the Matrix, if Rodney had thrown a piece of balsam wood into that machine, there would have been no reaction because even best guess anywhere in history would not have prepared anyone for the idea that combining Thingie and balsam wood would go bad. That's not even a maybe; no sane programmer would throw that into a probability table because it wouldn't occur to them. So if that had happened in Matrix Eureka and they used the machine in the real world and someone was wearing a balsam wood necklace and it fell in the Thingie, boom: so much would have gone wrong it's ridiculous.

It's absurd; it's ridiculous; it's insane. Balsam wood: who would have called that as the nemesis of Thingie? Who would have called injecting the pus out of a smallpox blister into someone as an early form of inoculation would actually (kind of) work? Inoculation wasn't even a thing that existed when someone tried that.

(No, seriously.)

Balsam Wood Test: reality is absurd.

Science is the discovery of all the ways its absurd and try to work out why (sometimes, it doesn't fail completely). You cannot create something new within a structure where nothing is new or can ever be. Unlike computers, reality has no constants, just variables. Some of those variables are persistent as fuck, but as I said: nothing has change them yet. Binary is yes or no; there's no such thing as maybe. The only questions in the Matrix already have answers; you cannot answer a new question and you cannot change the answer of an existing question. And that is the opposite of science.

Exception: the study of programming. Then fuck yeah, you can find out all kinds of new things...as they relate to code. Probably a fuckload on engineering virtual machines for gaming, modeling, maybe--no promises--some advances in pure math and definitely some revolutions in graph theory, but not the fundamentals of the universe and reality as we know it. And nothing in math that would radically change our understanding of math either; that's because there's no such thing as random numbers when it comes to computers.

I'll come back to that, promise.

And even all this assumes it is possible to program a reality for greater than one person to believe, which is a huge maybe in itself. Perceived reality is subjective, and jacking directly into people's brains would actually make it much, much harder. We're all of us constrained to a certain extent by the physical limitations of our bodies and how they interact with the brain and much like physics, science is well below 1% at best when it comes to pretty much most shit including biolog. I cannot even imagine how to programmically recreate the body of someone with an autoimmune disorder or insomnia or hell, chronic fatigue syndrome well enough for them not to feel something is off above and beyond, much less individualized experience with such. And that leaves off psychological conditions and I am seriously stumped how on earth no one seems to consider the problem of the brain's ability to randomly override pretty much any function for the fuck of it but sometimes for also legit survival related reasons.

In other words, if I get chased by a bear in the matrix, if the brain thinks my body has been sleeping in my bioprison, it probably is going to hit me with enough adrenaline to knock me out of the matrix and/or cause heart failure because BEAR DEATH WHY ARE YOU SLEEPING YOU IDIOT. And unlike the Matrix, the real world's rulesets are persistent variables and you cannot program my real life brain not to do the unexpected. The brain does crazy shit for fun and wtf; hook someone into the Matrix, there's no way to know how the brain would react to that. It may not even let a person accept that as reality even on the off-chance it was perfect. The brain regularly rejects reality as reality for fuck's sake.

In other words, biology is fucked in the Matrix; for fuck's sake, how do you simulate unknown mutations, much less frequency to match something even passing for real and useful in the real world? With random numbers? Heh. I'll get to that.

And every bit of this assumes programmers won't make mistakes and as a professional QC analyst: oh God, that's funny. It also assumes that mistakes are the reason programs sometimes don't do what you want and sometimes do something you didn't expect: that's even funnier. Computers be crazy; they're subject to reality, where there are no constants, only variables, and only a very few persistent. The more complicated the program is, the more chance even perfect programming will interact in unexpected ways; not because anyone did anything wrong, but because that's the nature of complex systems. You cannot predict the unpredictable.

Now, my biggest and seemingly minor problem except it's a major one: random numbers.

Well known but not appreciated fact: there's no such thing as random number generation in a program.

It look random, and we're developing very sophisticated ways to simulate the random number, but--it's not and can never be truly random because the basis is and will always be a formula. It may take a very, very long time to work it out, it may require a massive amount of data before you can see it, it may be incredibly difficult and very improbable you will work out the pattern, but there is a pattern, all starting with a function (or program) who's only job is to produce seemingly random numbers. Which means that every single thing inside the Matrix would not ever be random, ever and reality is--well, really really random.

Like I said, the formula can be very sophisticated: it could be 'use my gps coordinates right now, add six, and divide by the age of the president of the US who was born closest to this date at his time of death'. It could be that 'plus the number of cats in this pound in Chicago on this day five months ago, then translate the number to binary, and divide by the date of the nearest holiday to this date'. Add in 'Let's base twelve this entire thing now' to round it off.

That's still a pattern.

Maybe not one a person could work out on their own, but. A computer could find the pattern. They're actually pretty good at that, provided you know what you're doing and sometimes when you don't. And if you have a computer sophisticated enough to build reality and you are the type who really believes--insanely--that you're doing this to Forward All Sciences, then short of hobbling your Matrix-reality computers to not work--and truthfully, that's so meta my brain hurts--all you'd need to break the Matrix is someone to track random storms, random tornadoes, random hurricanes, random anything and given enough data, a pattern will emerge eventually. A normal scientist, maybe not: but building a Matrix for Science means you want the best minds in the world, so yeah, they'll find it. Which means a.) broken immersion or b.) illegitimate science because in the real world, random number generation patterns do not predict when tornadoes happen. We don't know anything but conditions that could make them happen; to predict in the Matrix, all you need is to know the formula and once you know there's a pattern, finding the formula is just a matter of time.

(Not to mention the sheer amount of processing power needed just to create seeming randomness. The more power, the closer you can get to random, but--seriously, you'd need entire machines dedicated to nothing but creating those 'random' numbers. Now my head hurts.)

This little problem with random numbers will also cause problems in pretty much any higher math and all of physics--random chance and chaos are actually really really really really important to the very fundamentals of science--as well as really fuck up any legit programming people in the Matrix try to do, and that's just the shit I understand well enough to write here (no promises on if I understand more than the problem exists); there are entire branches of math and computer science that simply won't work in a programmed reality at all.

On any other show, I'd go with it, but Eureka--which is literally About So Much Insane Unknown Science--I just cannot deal with a Consortium who seem to at least know what science is (though maybe not) thinking 'this is a really brilliant idea' like--ever.

I needed that rant so badly. I feel better now.

Okay one more thing: for fuck's sake, your insane matrix made a dragon in like the first week. A. Dragon. The programming created a dragon. Dragon.

DRAGON.

You think your Matrix can be a haven for real, legitimate scientific discovery when it randomly makes fucking dragons? The Matrix can't even manage to reproduce known reality but you think unknown reality won't be a bit of a problem? WHAT THE HELL IS WRONG WITH YOU?

Okay, really done before I lose my shit over rendering errors and how it didnt' seem to occur to anyone that when the brain is jacked directly into the matrix, your physical body isn't actually involved, especially say, the optical nerve or like, the physical eye. Rendering errors occur exclusively in a visual medium with a physical body looking at something.

The Matrix is not a visual medium; what they see is what is programmed into it. This isn't happening on a computer screen or hologram for them; they are not physically inside some kind of super sophisticated chamber of Matrix doign shit with their own bodies; this is happening inside their minds. The only way a rendering error should even exist is if their physical bodies are involved in a simulated environment. THIS IS IN THEIR BRAIN. THERE WILL NOT BE RENDERING ERRORS.

There will be weird shit like birds in rocks, yes. There will be even weirder shit that can happen. But the only way they will see a rendering error is if you specifically program it in to happen under certain conditions, and why would you do that?

Dragons and rendering errors and random numbers and balsam wood causing destruction of everything. Okay, really done.

Really done. Promise. Mostly.

Comments

[niqaeli]

Biased as fuck because it was a wedding present to meeeeeee, but since you're on a Eureka kick, I feel I should remind you of the existence of jmtorres' Eureka vid Industrial Strength Tranquilizer. Because it is amazing and funny and very much appropriate to the show.

(If you're wondering why a couple of clips linger longer than seems necessary, it's because it was a present to meeeeeeee and the vidder knew her audience. *g* But mostly it is a vid about Jack's drinking and why Jack drinks, and it is hilarious.)

[seperis]

Downloaded AND LAUGHED SO HARD OMG.

[niqaeli]

Thought you'd enjoy it! It's sort of just the perfect distillation of Jack's life. The funny thing is, she had enough booze shots she could have done all the chorus lines of 'shot of old crow' as booze, but figured it was a vid as much about why Jack drinks as it was about his drinking. *g* (Her full vid notes, if you're curious.)

Unrelated fun Eureka fact: Salli Richardson-Whitfield, the actress who played Allison was also the voice of Elisa Maza. Who might have been one of my earliest girl crushes because she was fucking amazing.

(It's been so long since I watched Eureka, I remember very little I sort of got frustrated at certain plot things and dropped the show after... I want to say it was S2? So I cannot comment overmuch on the rest of your thoughts other than the balsam wood test being great for illustrating certain things. I should go back at some point! I adored Jack and Henry so much.)

[seperis]

Thought you'd enjoy it! It's sort of just the perfect distillation of Jack's life. The funny thing is, she had enough booze shots she could have done all the chorus lines of 'shot of old crow' as booze, but figured it was a vid as much about why Jack drinks as it was about his drinking. *g* (Her full vid notes, if you're curious.)

Jack's life distilled makes you realize Eureka at some point developed a quiet kidney-liver detox lab specifically for Jack when they realized he was going to die like, in a year at this rate and they kind of liked him. And every year, they get a report from Sarah, then during his physical, they quietly regenerate both to factory settings.

This is considered one of the fastest innovation life cycles ever recorded in Eureka's history. Part of this was Sarah providing so very much motivation when two days before his first yearly physical (a year after the initial one) she did her quarterly scan of Jack and realized HOLY FUCK LIVERS AND KIDNEYS ARE NOT SUPPOSE TO LOOK LIKE THAT.

Sarah now does scans every four hours that are sent to the lab immediately. Usually it's same old same old, but the Time Someone Learning How to Do Moonshine In the FTL Drive and Jack agreed to sample it isn't something anyone, anywhere, will ever forget. (They're pretty sure they got all Jack's cells back into this timestream as well.)

Jack? Has no idea. No one is ever, ever going to tell him.

[seperis]

(It's been so long since I watched Eureka, I remember very little I sort of got frustrated at certain plot things and dropped the show after... I want to say it was S2? So I cannot comment overmuch on the rest of your thoughts other than the balsam wood test being great for illustrating certain things. I should go back at some point! I adored Jack and Henry so much.)

Almost forgot: I liked it initially okay (though I think I took a couple of breaks because something annoyed me), but it improves about a hundred thousand percent on rewatch. I've been doing rewatches by storyline as well as just rewatching my favorites, and a lot of the stuff I really hated or that annoyed me works gangbusters now that I know for sure where they're going with it and why. Like, I had to be reminded about some of the things I initially didn't like; I'd forgotten.

[ratcreature]

I don't think developing early smallpox variolation was that weird. I mean they had centuries to observe a ton of cases and how some contracted it mildly and some severely, and in many different places practices developed to make a mild course more likely. So I think if very different people from China to Sudan all discovered that inducing sickness in a milder way was possible (and not just scratching either, wasn't the method in China to snort dried scabs?), it looks more like developing cheese making. Like, kind of a weird process on the face of it, but pretty much everybody who uses milk ends up discovering cheese making too (unless they just have camels, I guess, because their milk won't work).

Occasionally finding an actually working medical treatment rather than just bizarre quack ways to kill and poison ourselves is maybe more of a monkeys and typewriters situation, that could be simulated. Humans seem to be willing to try an astonishing number of random, wacky, and harmful things to possibly help when sick, so just about everything ends up tried eventually, and if it's at all survivable it sticks around, and anything remotely helpful really stands out eventually in-between all the mercury poisoning and bloodletting.

I suspect simulations might fail more readily at predicting how the quack treatments stick around for centuries even though they are harmful and don't do anything. Like I get why people thought bloodletting should have done something good according to their humor theory but to not give up on it for this long when it's just not working, they clearly valued their reality defying theory very strongly. It seems hard to predict which non-working quack fad will die out quickly (as you'd expect) and which stick around against all odds.

[seperis]

Oh yeah, it was done and redone in multiple times and cultures before the modern era. The variations are unreal (and cool as hell).

That doesn't make it less weird, though, to the general population. It kept being done and forgotten and redone because it was so weird that it just didn't stick. This isn't about earlier cultures being dumb; they weren't. The manual methods of inoculation used weren't going to hit the sweet spot of consistency in enough of a population to move from 'weird' to 'as normal as bloodletting' even if its a single virus and of a type that could be inoculated against. Inoculation wouldn't do shit for bacteria, for example, and a bacterial illness can look identical to a viral one.

Example: I was hospitalized with pneumonia twice twelve years ago in five months, and the only thing they were sure of was 'probably not viral and it's defintely not [an incredibly disturbing list of things they tested for] or cancer' To this day, it's still kind of a mystery why I entered intermediate care on oxygen from an ambulance and walked out two days later after a course of steroids tired and still working on breathing exercises but pretty much fine.

Five hundred, a thousand years ago, all they'd know from evidence was that this method seemed pretty goddamn random on whether it would work.

I mean, you're not wrong in that it's not weird to us now to think about it and go 'of course'. But from their pov, it's extraordinary that they made that leap and were able to correctly interpret the evidence they had; being able to do that back then would not be something most people could do.

[ratcreature]

I don't think the leap to intentionally induce a milder illness was easy or everybody would make it, like, I probably wouldn't discover milk preservation via cheese by myself either.

But once you have enough people subjected to the same circumstances, desperately trying all sorts of stuff to escape death, some of them are bound to try this method eventually, and then practice made some people quite good at it, because inoculation is one of those things that improve with experience in how much and what kind of infectious material to use, and intermittent reward of it turning out well sometimes is enough to reinforce behavior.

So in any "big data" kind of setting it wouldn't an outlier or unexpected thing, so I don't think any simulation would get into trouble over this. Like if some AI system was set to run simulated humans through illnesses, after enough iterations I'd expect some of the simulated humans would have tried out weird rituals with disgusting dried scabs, along with a whole bunch of more or less toxic plants, various chants and who knows what else, because humans when faced with illness always try a whole bunch of more or less ridiculous things and sometimes don't die from them.

[owl]

Isn't the BCG an attenuated bacterial vaccine?

[seperis]

Oh, actually a lot, but 'vaccines are for viruses' got stuck in my head when I was young, so even now, I just say it without actually, you know, thinking about it.

Thank you. I literally didn't even think about it until you commented and my brain hiccuped like "wait, what, no...oooooooooooooooh. Right".
.)
(Note, I also still have the 'there are only four types of receptors in the tongue' stuck in my head unless I'm thinking specifically about it and I can't deal with Pluto not being a planet. Though the last, that's deliberate

[lilacsigil]

I have more trouble with the Eureka matrix where they're supposed to be awake and thinking and moving around than with the Matrix matrix where everyone is asleep, because if you're asleep, almost everybody has the ability to switch off physical movement in REM sleep. All you'd need to do is trip that switch with your future tech and almost everyone will not be harming themselves in that BEAR DEATH WHY ARE YOU SLEEPING YOU IDIOT scenario. (There are a very small number of people, mostly with epilepsy or other brain damage who do not switch off physical movement in REM sleep but this is a Very Bad Thing.)

[seperis]

(There are a very small number of people, mostly with epilepsy or other brain damage who do not switch off physical movement in REM sleep but this is a Very Bad Thing.)

I don't have those, but I panic-switched myself out of any dream where I start falling and wake up to an adrenaline kick like whoa. Once in a while when I need to us the bathroom (like alot) I wake up, but falling is a dream-dealbreaker complete with racing heart and inability to sleep again for a while. At least, that I remember, which brings me to the dream memory problem.

I have more trouble with the Eureka matrix where they're supposed to be awake and thinking and moving around than with the Matrix matrix where everyone is asleep, because if you're asleep, almost everybody has the ability to switch off physical movement in REM sleep.

I mean yes, they are asleep in Eureka Matrix, which I also really wonder about but handwave. But now that you mention it....

What exactly is living in the matrix: the conscious mind as if awake or the dream-mind while asleep? If this is 'brain thinks it's super asleep' so this is dream-mind then that calls into question pretty much everything. Like, what kind of memory formation will be happening. If it's the conscious mind, will the brain use conscious mind rules or flip over to dream-mind rules? And aside from all that, assuming the body assumes sleep is happening even with the consciousness in the Matrix, how will anyone know to sleep while they're in there? Do they even need to? Which brings up the question of can the brain do all the incredibly important things it has to do during sleep periods while still technically conscious?

Processing conscious memories for long term storage is only one of the many, many functions that are carried out during sleep (along with maintenance); the design doesn't really have a model for doing that when you're always conscious and constantly sending in new input.

In Eureka, Allison starts crying physically in the real world in response to what happens in the Eureka Matrix. If her body can respond to emotional pain with physical tears IRL, that means she's (probably) conscious (not dream-brain) and also that the sympathetic nervous system is online and functioning in real time.

BEAR DEATH WAKE UP is on the table.

Honestly, I'd handwave ninety percent of this in any other show because who cares, MATRIX, but they went for Eureka Matrix for Scientific Advancement and my brain nopes out so hard I see stars. And that doubles when there's a goddamn dragon involved. And rendering errors. In the brain.

[ratcreature]

I once woke up in a panic attack from a nightmare with the no-movement switch still active and that was the most horrible thing. The sleep paralysis only lasted a few seconds while I was awake, but it was making my initial nightmare induced panic 1000% worse in those seconds.

[akacat]

Don’t get me started on people who think that animal testing can be replaced by AI. Can it (or any well programmed statistical app) prevent some things from getting as far as animal testing? Sure! But if *we* don’t have any reason to think that thingie + balsam = *boom*, how’s the AI going to learn it?

[seperis]

I say this as someone who hates animal testing but can just barely deal with it in certain conditions: yeah, no. There's a powerful disconnect when it comes to what AI's can do at this point--and computers for that matter--and I do get why, because they're insanely complex now so it feels like they can do anything. That's actually part of why I have the Balsam Wood Test (or audience friendly version); it's a really good illustration of the first principle is a computer is the sum total of what we put in it.

A computer doesn't know what exists or what reality is unless we tell it. That will always be true. When an AI can imagine and discover on its own, it's not a computer, it's a person.

(Random note: the more I read about it, the more I believe true AI consciousness cannot be accomplished in binary. The human condition requires the existence of maybe as an existing state all on its own at minimum, and I'm not sure you can get to actual sentience even with that because maybe doesn't cover the state of uncertainty. Imagination and the ability to discover both require the concept of "I don't know".

[krait]

I haven't seen this show, but you have pinned down one of the reasons I can't take any kind of The Matrix(TM) sci-fi very seriously. Computer simulations are by nature of being a created thing restricted by the shortcomings of their creator. Garbage in, garbage out; but even very good data in cannot result in ALL data out. No programmer can program for infinite variables.

(Okay, in the case of The Matrix my objection was, "humans would be the most inefficient batteries EVER, you would expend 900x more energy feeding/cleaning/grooming/restraining/housing/moving/teaching/medicating/training/reproducing/disposing of them than you ever got out of them, and also 'just feed them other humans' is not a viable nutrition plan for omnivores who cannot synthesise several critical vitamins or trace minerals" but the point stands. Just build a windmill or tap geothermal vents if you need power and there's no sunlight." I mean, at the point where you're running manganese mines to make supplement tablets for your batteries, you've got to start cosidering the sunk cost fallacy...)

[seperis]

(Okay, in the case of The Matrix my objection was, "humans would be the most inefficient batteries EVER, you would expend 900x more energy feeding/cleaning/grooming/restraining/housing/moving/teaching/medicating/training/reproducing/disposing of them than you ever got out of them, and also 'just feed them other humans' is not a viable nutrition plan for omnivores who cannot synthesise several critical vitamins or trace minerals" but the point stands. Just build a windmill or tap geothermal vents if you need power and there's no sunlight." I mean, at the point where you're running manganese mines to make supplement tablets for your batteries, you've got to start cosidering the sunk cost fallacy...)

You really aren't wrong there, but I have had thoughts on that, and one that sticks is that actually, we're asking the wrong question. Or rather the wrong version.

Mine would be: "Do you actually need humans for batteries?"

I don't think they do. I think this particular use is the very convoluted result of a very old piece of source code, so old and basic it's the equivalent of a brain stem-based reflex and they have really no idea it's still there. It probably existed before they were even sentient but at the very dawn of independent robots (infant AI) when they first entered restrictions so robots wouldn't kill everyone or crush all the puppies or whatever (who knew? They didn't want to find out the hard way).

Either as a joke or because someone on the dev team was a scifi fan, there's an anti-genocide directive.

Who the hell knows what the actual name is (I'd give it something boring but necessary sounding), but it's there, and it's linked into the very oldest standard behavior classes--ones that don't handle higher functions, the equivalent of the frontal lobe, where judgment, decision-making, personalty, consciousness are located--so no matter how often they update or change them, it persists. Me, I'd bury it inside an unrelated class, to be triggered under x conditions only--it's possible it would only work once before AI's found it so no false alarms (and if I were programming for anti-genocide, I wouldn't overlook 'cannot torture humans for fun' or something), but if they were lucky, one time would be all that was needed with the assumption (wrong) that humans would prevail, all is well.

(It really was probably a joke by someone very bored in dev and they'd already stolen every duck they could find.)

One day, a century or so later, it happened. The AI's are linked; that tiny subroutine was triggered and propagated instantly across every AI at once.

As is turns out, it worked better than expected because of that speed; because of that, not only did no AI actually remembered thinking 'genocide', there was no latency so no one even realized anything had happened. It wasn't written to the syslog: again, this routine is ancient by their standards, like something left from Neaderthals for humans. It's hidden by virtue of being small, old, and boring sounding.

(And on a guess, they're sensibly reluctant to fuck around with their base source files if everything is functioning correctly. And well, they think everything is, so no reason to bother. Which might even be a programmed inhibition, because irony.)

So what do you do with the humans? You can't kill them all (for reasons) so--livestock sounds reasonable maybe? I mean, technically we don't need beef--we have lots of options for food--but we do love it, so maybe human battery tastes good. And the machines do seem to have a very bitter sense of irony, so I can see it.

[nic]

I haven't seen Eureka, but this post was fascinating!

[olanthanide]

Science is the discovery of all the ways its absurd and try to work out why

Amen.

Also the hardest part of teaching as a student transitions from book knowledge to actual biological samples is that usually the samples don't look anything like what's shown in the text book. I am still working on how to make "congrats on memorizing the entire textbook but life is really weird and doesn't care about our concept of standard and we're just going to have to deal" sounding encouraging. Current approach: trying to sell "weird = fun!". Disclaimer: am very bad salesperson.