Messages

Although I do play with Stunts sometimes, what you do with these races requires wasting way too much time and I have more useful things to do. Besides, in my games, which already look and play better than Stunts, you seem to have no interest at all.

To be fair to us: almost every game released in the last 30 years looks and plays, on a technical level, better than Stunts. I mean, 20 simulation steps per second? I think the only game I know that runs significantly slower than that is Pokémon Go. We all have our own reasons for liking this one very specific game so much, but none of them can be summarized as "there were no better looking games". So getting us to drop Stunts and take our whole community to another game was never going to be a matter of having a better looking game. Even if I, taking myself an example, was disgusted with the (ever joyful) look and (awesometastic) bugs of Stunts and wanted to race something better looking and smoother running I'd be in the Riptide GP community. I love those games almost half as much as I do Stunts. And outside of racing games I like The Just Cause series, the Assassins' Creed series, the Tomb Raider Games, Age of Empires and all of its spinoffs and imitators, and so many many more titles. We know we have options. that's not why we're here.

I genuinly feel you do, maybe partly on accident, make a beautiful point here. We all have our own priorities with our time. We all write things, read things, play things, watch things, and we all continuously make choices on what to write, read, play and watch. We're all way into adulthood here, you and me and everybody, we all have our own lives. For me time spent playing Stunts will never be wasted. Your time will never be wasted writing articles and linking people to cool songs. But that doesn't mean we have to agree on which one feels like a waste and which one doesn't.

A few people here checked your game out, gave feedback. I saw a conversation in which technical tips were being swapped making everyone's projects better. good stuff. I might check the game out too, later, sometime. But at the current moment my free time is filled up just the way I like it.



...Wait, let's climb that hill, maybe we can still see the original topic of this thread from there.

In Dutch we call it "roeien".

It is a funny example of there sometimes being several competitive solutions to a problem. Three in this case, with very close results.

What you call consciousness is not consciousness and when you tried to define it you used a bunch of other notions that can't be defined.

With all due respect: I just typed out a full page of extra explanation on how I used this term in my first post, defining it as best I can in order to give clarity about what I meant, as a response to your reaction, and I'm afraid what you've gotten is what you've gotten. I am now going to leave it at this, because I don't feel like having a debate where I take a lot of time to explain myself and the other person just says "it's somewhere on my blog, your homework is to read the whole thing." I've always been pretty good at resisting that particular urge.

In fact, now that I think about it, I don't really feel like a debate at all today. I just wanted to give some context about this track, and that mission was accomplished in the first post. But feel free to leave any interesting contributions on the topic in this thread! Any speculation on the winning cars and drivers as well of course! Your own game already has its own thread, but you can go there to talk more about it.

https://en.wikipedia.org/wiki/Artificial_general_intelligence#Current_research

The actual complexity of modeling biological neurons has been explored in OpenWorm project that aimed at complete simulation of a worm that has only 302 neurons in its neural network (among about 1000 cells in total). The animal's neural network was well documented before the start of the project. However, although the task seemed simple at the beginning, the models based on a generic neural network did not work.

Being a neuroscientist must be a humbling profession. In a frustrating way, but also often in a good way.

Although I would like to add after looking up the project on their Github and Wikipedia pages that they certainly didn't seem to have the idea that they could just start up the simulation and have a working C. elegans. Even just trying to code out these individual cells is a challenge, it seems like they haven't even gotten to the part where they can let it figure out how to crawl yet, which would be an achievable task for today's regular self learning algorithms.

About consciousness:

https://matei.one/osgod.html#saved

I'll try not to go into the realm of religion and spirituality too much. The term consciousness as I use it means the ability to use formal-ish logic (if this than that, also the basis of math) in a structured way with a feedback loop that lets your brain understand what it's doing so it can logic out the next step. In most living humans this also includes the use of language as part of the logic and part of the feedback loop, although as language seems to be at least largely a learned skill it should most likely be possible to run consciousness itself without language, to some degree. This skillset I hang the term consciousness on is quite hard for brains to do because it's very different from the style of logic a single braincell uses. Braincells/neurons are fantastic things: they can take dozens, hundreds or even thousands of factors with positive or negative impact and each with their own weighing factor and come up with a single conclusion, which is updated in real time. A single braincell can do that. You can see this in action in young babies. They don't start crying because of any one factor, instead they take everything together (hunger/thirst, tiredness, wetness of buttocks, currently being rocked or not, smell of mam detected) and come out to a single crying volume based on all of those, which continually updates as the factors being used change. But that's very different from being able to do math. In a way brains are reverse computers. Even very simple computers can run amazingly complex formal calculations, the trick is taking that formal logic machine and make it do things that are more fuzzy. Are brains use much more fuzzy logic, and it's hard for them to use formal logic instead. Of course, saying this is an oversymplification of how brains work is an understatement. As brains are way more complex than any single cell, and even the cells are more complex than I give them credit for.

For me, using this definition of consciousness, I'm not sure this ability/brain structure is even the primary seat of personality. We know it certainly has influence on personality. There is for instance the case study of Phineas Gage, who incured major damage to the logicky parts of his brain in a workplace accident, and his personality changed quite a bit afterwards. But personality to me is a very broad term. A frog can choose to jump, wiggle or sit still when I pick it up, and not every frog reacts the same way. That's surely a form of personality, and the frog does this without a proper prefrontal cortex. (Animals like birds also do a lot of amazing things without a proper prefrontal cortex as we mammals have them, but that's cheating, they have analogous, if anything more efficient, structures elsewhere in their brains. Emotions are a big part of personality, and they're not really part of this consciousness. Yes, you eventually feel that you're angry, but that wasn't a decision you logic'd out. You only become aware of the angryness at all after the fact, all the decisions leading up to it were made in the fuzzy, unconscious way. Your consciousness basically just feels the angryness. The best we can usually do is reason with ourselves after the fact about which factors it were that made us angry. Craft an explanation for ourselves.

Now, as for the part of the discussion that I didn't really want to get into, I realize this is an incomplete story without it, so I'll try to at least keep it respectful instead. We know that the state of the brain has major influence on personality. Children grow up in conjunction with their developing brains, people with forms of dementia lose part of their minds and can lose parts of their personality too. And I already mentioned the guy whose personality changed after getting a big spike through his brain. There was also at least one person who could not make any new memories after having his amygdalae (and some other nearby brain tissue) surgically removed. He died in 2008 still thinking of himself as a young man, getting the shock of a lifetime every time someone made him look in a mirror (or, you know, at his hands). While memories are perhaps not strictly part of someone's personality, in an ideal world where there was a noncorporeal part to our personality these two would ideally always stick together. (Although I guess a case could be made for a reincarnation based system where you wouldn't want them to stay together, so it might make sense to have the memories be part of the physical brain.) Science doesn't know everything about our brains, there could certainly be an extra part to our personalities that is made of massless matter undetectable to any sensor, runs on an unknown and undetectable form of energy, can communicate with our physical brains through means we've never even thought of and has the ability to, I don't know, shift in the fourth spatial dimension to another place after we die, something like that. But that possibility is so far outside of the realm of what we know that any speculation on it is pure guesswork and/or unadultered theorizing. It makes Russel's teapot look like an absolutely mundane idea, and makes physics theories involving curled up dimensions and exotic forms of already questionable concepts like dark matter and dark energy seem like pretty grounded concepts. At least there is some math that says it could work if things were like that. For this reason I choose not to include ideas like incorporeal personality fragments in posts like the one above. I have no problem with speculative discussion, but when I'm trying to create an overview like this I try to focus on things which have at least some connection to the known world and the evidence. Sometimes I'm just more okay with not having a complete picture and leaving gaps in the explanations I give than with speculating all too wildly. I'm not saying I can judge which things aren't true, I'm definitely not omniscient, but I choose to stay closer to the things that I'm at least some degree of sure are true.

As a final note: In don't believe historic people were stupid. The ancient Egyptians didn't need aliens to build their pyramids, they were very good at building them because that was important to them. Hittite clay tablets from the bronze age have been found describing methods for training horses that we didn't develop until somewhere after World War 2. Because horses were important to them. The Bible as a source contains historical truths that were at some point taken as likely myths because of a lack of supporting evidence, like the existence of the aforementioned Hittites and the Sea Peoples, which are an unsolved mystery, but certainly seem to have been a thing. But that doesn't mean I'm going to take their word for everything. Until an ancient scroll is found documenting a bit more about how they found these souls and how these things supposedly work I prefer to read the sciptures on that subject as speculative.

I hope you'll accept that explanation, and I hope to see you on the track.

If you've been wondering what this track or this name are about, I've got a science lesson for you. For this I'm going to be using the image of the track as well as an image of the main three dopaminergic pathways from Wikipedia.

You're going to notice that the track we're driving does not actually really match the other image or my description of the dopamine pathways. It was one of the concepts, but I scratched it off as too specific and difficult when I could just be building a fun track. The final product semi-accicentally looks reasonably close though, and there are ways in which the name "Dopamine Pathways" is highly appropriate for a Stunts track, which I won't be summarizing until all the way at the bottom of this post. (Edit: well, sort of at the end but before a bunch of extra additions anyway.)

Note: if you go googling pictures of dopamine pathways you'll find images with only 2 pathways, with 4 pathways and with entirely different pathways that don't look anything like this. Brains are amazingly complicated, and even where science as a whole has some understanding of them, I often do not. I'll be basing this post on a course I followed years ago and barely passed on the fourth and final attempt. I could have been kicked out over that, and the worst part is that it was an elective, I chose that course. The point I'm trying to make is this: I will be doing my best to say things like "if I remember correctly" as little as possible, because this entire thing is one big "if I remember correctly". I take no responsibility for anyone experimenting with levodopa use after reading this post.


You cannot view this attachment.

So, there's this class of substances called neurotransmitters, whose job it is to activate, deactivate and modulate types or clusters of brain cells. If the electrical signals going through your nerves are telephone calls, targetting specific individual cells, neurotransmitters are more like megaphones, calling entire groups to spring into action, or calming them down again. There are a bunch of neurotransmitters, like noradrenalin, glutamate and GABA, all of which have their own pathways, their own channels through which they are transported to the areas where they are needed. Dopamine is unique amongst them because even beginner courses on the topic acknowledge that this one has a lot of different functions, and those functions are grouped by pathways: where in the brain does dopamine cause these effects?

The top pathway (going by the direction they take from the point of origin) goes to the central part of the brain. I'd like to note here that we often place a bit too much emphasis on these different brain regions. Most processes are not actually completely limited to one area in the brain, at best we can measure more activity in one specific part. That doesn't mean the other parts are doing nothing. It also doesn't mean you only use 10% of your brain no more than you not doing all the fitness exercises all the time means you only use 10% of your muscles. But since these brain regions are a useful tool for visualization of where this increase of activity happens it can still help to think this way. Anyway, this central part of the brain is itself divided into way to many different parts with their own names. There's a central part called the thalamus, but beyond that there are the hippocampus, the amygdala, the striatum and many more. We're just going to be viewing it as one big happy corners and crossroads region. This part of the brain is heavily involved with things like emotions, but it's also basically the middle manager of the brain: it's involved in the smooth running of everything.

Take movement. Conscious movement starts in the prefrontal cortex, the loops and corks region. That's the part associated with many "higher" functions, and with consciousness. If you can hear yourself thinking, whatever you're doing at that point, like maths or writing forum posts or finding shortcuts, it's probably being done largely in the prefrontal cortex. So the prefrontal cortex says: I want to drive around this corner. That assignment goes to the premotor cortex, the front part of the jumping bridges area. This area breaks down a familiar movement into the parts it's made up of. You don't consciously have to think about this, because you already know how it works. You have to brake, steer, then tap-steer, accelerate and finally stop steering and be ready for a little counter-steering if needed. Those separate assignments are sent to the motor cortex, the rear part of the jumping bridge area, which knows how to execute them. Acceleration is one fluid movement of extending and pressing down the middle finger coördinated by the motor cortex. (Most brain courses would give an example like a salto being broken up into arm, leg and spinal movements, but this is the Stunts forum.) The outputs of the motor cortex are sent to the limbs, while the cerebellum (the elevated roads section) monitors my balance to make sure I don't fall out of my comfy chair during my steering movements, and with the spinal cord (not pictured, the part in the bottom is the brain stem, but they are kind of an extension of each other) activating reflexes, like me withdrawing my finger quickly if it turns out my up-arrow is scaldingly hot, somehow.

Yet with all of that going on the central brain region is still at all times regulating movement, and in particular it's regulating when muscles should not move. This upper pathway is among other things involved in Parkinson's disease. This disease, known for giving often older people the everlasting physical jitters, is caused by these central regions of the brain becoming insensitive to dopamine. There's been some wonderful work done in treating Parkinson's symptoms with electical stimulation, but the only pharmaceutical option is more dopamine. Or, well, dopamine cannot cross the blood-brain barrier, formed by the walls around the brain's blood vessels, so they get a "prodrug" (not the same thing as a drug for pros) called levodopa that can cross this barrier and is metabolized into proper dopamine ones inside the brain. This may feel slightly overcomplicated, but for several highly medical reasons like the risk of infection and the existence of the skull it's a much better idea than trying to inject dopamine directly into people's brains.

It does provide an interesting complication though: we can't target levodopa to just the top pathway. Which means you're often going to see side effects related to the other two pathways. Speaking of which: the bottom pathway's prime target is the cluster of the hypothalamus and the pituitary gland. (You might notice this image actually labels two separate bottom pathways that overlap. See the part where I bragged about how little of this I remember. For what it's worth the tubero-infundibular pathway visits the target site, so you can just pretend I'm talking about that one and ignore the mesolimbic pathway.) The hypothalamus and pituitary gland are the little bump on the lower-middle-leftish side of the pipe region. This area is involved in some of our strongest physical reactions to mental issues. Stress for instance is mostly regulated here. But dopamine itself is more closely associated with feelings or reward and accomplishment. Winning a race, scoring a hot date, that feeling of being on top of the world. It's basically the best thing ever, which is why a lot of the most potent recreational drugs act on this pathway, increasing the dopamine signal and the rewarding feeling. All of the high, none of the work. This is also why those things are so highly addictive. Humans aren't motivated by what they have already accomplished, but by what they're going to accomplish next. Dopamine is addiction. You want more dopamine. Funnily enough accomplishment turns out not to be the same as happiness, and most clinical antidepressants, at least the ones prescribed for long term use, do not act primarily on the dopamine system, but on the serotonin system and sometimes the noradrenalin system.

You might have noticed in the meantime that the middle pathway touches pretty much all of the brain. It runs through the prefrontal cortex, premotor cortex and motor cortex discussed earlier, but it continues all the way into areas with sensory functions, like the optical cortex all the way in the back of the banked road area. Because it's such a "general use" pathway that just brings most of the brain their daily dopamine dose it's not associated as clearly with one very strong specific effect like the other two, but if I'm not mistaken this is the pathway that between the prefrontal cortex and the sensory areas is responsible with a known side effect of dopamine: hallucinations. It's not the most efficient way of hallucinating. There are more powerful options out there. The main hallucinogenic drugs are mostly not chemically related to dopamine as far as I know, but I do believe that part of their way of acting is a secondary activation of the dopamine system by whatever they were primarily doing.



Anyway, this brings me to the part where I can answer the question this post was about. Why call a Stunts track Dopamine Pathways? Well, dopamine pathways are involved in movement, like a gamer moving their fingers or a car moving around a track, they are involved in addiction, like the mild one that keeps people coming back for more races in their favorite game, and they're involved in seeing weird shit like illusionary track elements and cars that speed up by jumping and turn corners while flying through the air. It's also appropriate because they're called pathways, so they're like roads. All in all, dopamine has a lot to do with Stunts. Or as I'm pretty sure the kids don't say it: Stunts is pretty dope.



This was the Stunts science corner for this week. Thank you for listening (well, reading) to me pretending to get value out of that brain course. I never did make it in neuropharmaceutical research, but I'll always have my mildly inaccurate associations between visible real world stuff and certain neurotransmitter pathways.

If this became a jumbled mess of word spaghetti after two paragraphs because the space in your brain for new technical jargon was filled up: I guess you could reread it tomorrow. Eventually you will reach that barely passing grade.

Final note: no, I did not think of all of this while drawing up the actual track. one of my hobbies is overthinking things after the fact to see how much work my subconscious has been doing to line things up.

Sounds good.

@alanrotoi as our first volunteer: would March work for your idea? And if so, since it's pretty close already, do you have a weekend of preference?

Addition: For white cars the selection of the white background doesn't work. So there the trick is that you make the exact same screenshot ofa non-white paintjob, and have both in the same image as different layers. Use the non-white paintjob for the selection, and apply the transparancy on the white one.

Hope you'll enjoy driving the Indy as a normal car, for the first time in a while.

We just did the Indy as a normal car, in the PG-less race. Now it's back to the regular programming. I think someone will find at least one more PG section in this track. I have absolutely no idea where and how, but it feels like that would fit the flow.

I just hope we'll see a little less LM02 this year, maybe it can finally win something convincingly. But if not the at least two Indy races (including power gear sections) within the first half year might balance out the slowness of the LM02 a bit.

We've been conqlebuggling a bit in the little Telegram group to work out the proposal Duplode was talking about before.

Is the conqlebuggling worth saving for the future? (e.g., grabbing a log out of Telegram and archiving it here?)

Not really. A little bit of talking about bonuses in relation to eachother, like that I had done a recent test in which the F40 was about 2.5% faster than the CERV and the GT3 +-5.5% slower, how we felt you probably got a pretty good number on the Lola, comparing it to the MP4, how the Thunderbird was probably slower than the LM02 on most tracks but we felt like it wasn't too bad to start it off a little lower and a little about how it was fair to give the Indy a little bit of a bonus because you liked the idea so much. And a bunch of chat bringing bonuses up and down a bit when we thought of something, like whether the car we were comparing it to was currently at a low point or a high point in its range.

So, that, basically.

Okay, I've cooled off a bit. Maybe my language was a bit harsh.

We've been conqlebuggling a bit in the little Telegram group to work out the proposal Duplode was talking about before. Crowdsourcing the estimations of the newly added cars, so to speak, trying to take your views into consideration as well. So that might turn out as something worth considering.

EDIT: I'm an idiot, I refreshed the page but failed to go to the next page.





And yes, that was a made-up word.

Personally, I think this whole rebalancing is a bad approach.

The cars are rebalanced automatically, but very slowly. You're taking cars that have been patiently waiting for their chance and setting them back another half a year of waiting or more, and you're taking other cars that have dominated this season to the point that people got sick of them and you're goving them 2 entire full years (!) of extra bonus, which ensures that they will be even more massively dominant this year, different track design or not. The principle of every car getting their chance to shine every year or so is kind of being thrown out the window with these huge start of year tweaks. I think the Mercedes' run last year showed that we shouldn't try too hard to rebalance and predict these things. Its bonus was set too low, we all knew it would just be dead weight for the entire season, and it was. My preference would be to not actively try and do that to as many cars as possible.

If you want to personally manually change the coefficients like this every year my vote would be to at least move back to the +3% system rather than the +1% system. The +1% only works if the cars are allowed to find their natural spot in the order over the years, and brings the cars closer to their natural sweet spot than the 3% does. But if we just randomly hand out huge bonuses to cars that were already overused the system is going to need the bigger jumps to at least get back to near those sweet spots faster.

And I don't think having all the tracks flooded and shortcut-less is going to fix that. People have been getting better at using powergear, that means we have more powergear races until we reach the point where they've naturally dropped far enough to be fair again. At best we could overcompensate a bit and lower their bonuses, not raise them.

Four worked pretty well laat year. Created some room to experiment with the ruleset too.

Potato

Leek