Gran Turismo Sophy: Sony AI x Polyphony Digital

[Yard_Sale]

If you could provide me some reference material for ACC's in-game AI maybe I could say something but at this point I know nothing except that it is likely AI similar to many other games which adapts to various inputs using things like decision trees but most of it is pre-baked. It might have some code stating what the ideal line is, some code to make it change its behavior if its about to lose traction, other code that governs its behavior for overtakes. For Ex: if the upcoming straight is long enough, try to pass with slingshot if in range. I may be very wrong here though.

So as I have not seen anyone else tackle this, I with my very very limited knowledge of ML but a graduate's level of understanding of computational methods and optimization algorithms will take a crack at it.


People here have a shaky grasp of what is happening with ML specifically reinforcement learning used here and what words like "AI" or "model" even mean in LM literature. I have not worked with ML or deep learning BUT I have worked with evolutionary algorithms (CMA-ES) https://en.wikipedia.org/wiki/CMA-ES
which have also been previously used for similar optimizations as the abstract of the paper states. Fortunately, there are some similarities in these approaches even though the implementation differs. Let me try to explain a bit here:

Evolutionary algorithm: Usually you are looking to optimize some cost function (let us assume here its lap time or average speed or some function that incorporates multiple of these desirable factors). Now as the name suggests this is done using "evolution" and "selection" of candidates that provide the best result for optimizing the cost function. You start first with some random guess of a good candidate, then the algorithm essentially "remixes" the initial generation to make a new generation according to some criteria which then are evaluated and the best ones are then crossed together with some pre-determined randomness always to make the next generation and so on. The CMA-ES algorithm can also "move in the right direction" so to speak, meaning that the crossing is done in such a way that the new generation is closer to where the algorithm thinks the true optimum lies. You can think of 3-D graph that looks like a mountain, the algorithm essentially tries to drive the cost function to the top of the mountain by looking at which area has a slope rather than randomly going in all directions looking for the peak (Very basic strategy and not a very robust one). This is all done with some numerical weights to the functions that spit out the new generation and the evaluation of each generation. These weights are quite similar to the "rewards" in deep reinforcement learning. Essentially, some weights make certain good behaviors like no penalty make the evaluation higher while bad ones like contact might make it lower. You can tune these weights for making the algorithm faster or behave differently. You can also make these weights themselves be automatically tuned by looking at the slope of the mountain I mentioned earlier, and this is what CMA does.

Deep reinforcement ML: Similar, but here after our initial agent, lets call him, instead of using evolution by mixing candidates, the current agent's parameters are tuned using data that it can "learn" from as we use the data to feed the algorithm. Here, Sophy can be fed the data for the track and and its opponents and we keep rewarding Sophy positively for good things like course progress and negatively for behaviors we want to discourage like contact, going off-track, etc. These rewards the the ones cited in Sony's paper along with other ones. Sophy can basically run the track many times to try to find the optimum with the help of the algorithm.

In fact, I the weights used for Sophy are listed in table 1 and are as follows: View attachment 1112915

With all this said, you might ask what exactly is the "thing" that gets saved or what do you get with the optimum solution. Well a matrix, really. In Sophy's case they use a modified QR Q-function (disclaimer, I am not very familiar with this algorithm or function myself) https://en.wikipedia.org/wiki/Q-learning

I will point out that the rub here is that they have modified this to accept continuous actions as inputs instead of numbers defining a state at each step.

Still reading? Well the answer to your questions is basically here. This Q-function is what a trained agent will have and obviously will need some work done to be implemented as an AI in the game as it will have to be consistent with whatever the game's implementation of AI model currently is.

So, while I cannot for certain obviously (who can in any scientific matter), once the model has been trained, running it on local hardware as pre-baked AI should not be any more taxing than running the current AI cars.

Very well explained. It took 2 reads through, but I eventually got it 👍🏼

 

[05XR8]

What’s really insane about all this, everything they programmed for sophy and the questions about is sophy doing something good or bad, these could have been implemented in the game for players.
Seriously. Rewarding players over a lap, letting a player know if a turn was good or bad, braking early or too late. The coaching. It should be in the game since day one.

In ACC, there are some basic objectives. A bar at the top of the screen lets the driver know if they are getting smoother over several laps. A target of trying to complete four laps smoothly, not perfectly, but smoothly, is realistic. GT players are having to beat the clock for the first time ever playing these games. That’s unfair for new players.
The basics should be about encouraging smoothness. Getting the player accustomed to the physics(even if GT physics/controls are easy for novices) and speed.

At the beginning, the clock should only be used for progress. Not the reward. Not the target. Reward the player for keeping the car on the track and for smoothness and consistent clean laps. Good coaching goes a long way for driver confidence. Something Sophy seems to have done and what some other games do.

 

[05XR8]

It's slightly more obvious than that.

The suffix "-sophy" - as in "philosophy" - literally means "knowledge". It comes from the exact same Latin (and Greek, oddly) root as the name Sophie/Sophia (generally held to mean "wise one").

What better name for an artificial intelligence than "knowledge"?

This is cool because, “Phil” means “to love”. Might have to change B-Spec Bob’s name to Phil.

 

[JD_18]

First of all, hats off to Sony & PD for this collab. Even though we've had the 2 previous studies showing AI driving faster than humans in a time trial setting, this is much larger in scope. I didn't expect to see AI machine learning outracing the best GT Sport drivers so soon. If they can implement this in the game in a realistic time frame, it would probably be GT's biggest contribution to the racing game genre since the first one's bringing simulation to the masses.

Now, a couple of my rambling thoughts (some might already have been covered in the thread, I didn't have time to read every post in detail):

- In terms of pace, the AI is stupendously fast. But we already know it's capable of doing this from the previous study at Tokyo Expressway. The AI can be perfect in its input all the time, and it can react in literally frame-by-frame of the physics engine. That's why it can pull off the 4 wheel drift kerb entry at DTS' hairpin. A human will never be able to do this with any sort of consistency, because our reaction time and input device is limited. In a twisted way, I find it very satisfying to see the top GT players feel the desperation us plebs feel when watching their replays. I've always said that the difference between a good driver and an alien is just accuracy. Well, the difference between AI and an alien, is just superhuman accuracy

- Taking the above then, an easy way to make the AI more "human" would be to limit their input frequency (e.g. if they are now running at 60 Hz, possibly tone it down to 5 Hz - human fastest reaction time is around 0.2 secs I believe). So a difficulty slider in-game would basically just be a frequency slider. Then they can add extra parameters like mistake probability, overtaking/defence aggressiveness, etc.

- Moving on from that, the weird way the AI drives shows that GT physics still has many unrealistic loopholes. For example the DTS kerb drift and using the grass on Maggiore last corner entry and Sarthe Dunlop curves. So in a way, having the AI push the limits will also help PD plug those loopholes in the future (hopefully). Back in GT5/6, drifting the corner entry used to be the fastest way to go round the track (just look up the GT Academy replays from 2012-2014). I thought this was finally gone, but it seems it's still there, just more difficult for humans to do.

- Same with abusing track limits (Sarthe pit entry lol). It's comical that PD thought training the AI on the track with the worst track limit exploits would make for good advertisement. Fix your penalty boundaries PD!!!

- All 3 races are done with no tyre/fuel consumption. It would be interesting to see if bringing these 2 factors would make the AI take more realistic racing lines, because 4 wheel drifting every corner entry would definitely kill the tyres quicker than the conventional line.

- For the racecraft, it is genuinely impressive to see the AI improve from the July race to the October one. However I think they pushed it a step too far in aggressiveness to compensate. The much higher number of penalties in the second round shows that. Because the AI is so much faster, it can pressure the human player into a mistake, and then just crowds/divebombs them at the nearest opportunity. Once it gets ahead, game over. It feels very forced and not very nuanced. It's not a classic one-on-one battle like with a similarly skilled human player, where you can be trading blows for multiple corners and not know who will come out ahead. The battle with Yamanaka at La Sarthe in the first round makes you think they can, but Yamanaka is just hanging back here for the final overtake into the Porsche Curves (common strategy for Sarthe in slipstream dependent cars). It's not the AI's brilliance that kept it in the battle for so long.

- In terms of commercial implementation in the game, this is where the biggest barrier lie IMO. The time (and computational effort) to run the simulations to get the AI to this point would be enormous taking into account all the cars/tuning/tracks/weather/tyre/fuel possibilities in the game. In the article it says it takes 300,000 km (10-12 days) for an AI to get to alien level. Just taking GT Sport, we have 338 cars and 82 raceable layouts. That gives 27,716 combinations. Multiply by 10 days = 759.3 YEARS of simulations I don't know about you, but I think none of us would be alive by then Of course you can speed up the process by having more computational power, but how much money are Sony/PD willing to spend to get this done? If it takes 1000 virtual PS4 machines to run one combo, to get everything done in 5 years you'd need roughly 152,000 virtual PS4s. And would you accept if shifting their budget means we have less car/track licenses for example? And whether that's good use of computing power and energy instead of solving other world problems/scientific research?


Future thoughts/possibilities:

- Could the AI be used for Tuning cars? No need for us humans to test manually, just let the AI run every permutation of every suspension/LSD/downforce/gearing setting to find the perfect tune. But as we've seen, the AI doesn't drive like humans so it could result in a very fast, but also very unstable car.

- B-spec! Let Big Brain Bob complete the races that are too difficult for you (and also grind money while you're AFK ). But if all the AI are super, then Bob is just back to being its average self again...

- Strategy. Similar to tuning, no need to run your own practice race to figure out the best strategy for FIA. Just let the AI pound round and figure out the best tyre/fuel management.

- More of a curiosity, I'd love to see the AI take the Red Bull X1 and Tomahawk X around Nordschleife to see what a perfect lap would look like. Human record is just under 3 minutes currently (in fact there's a TT in GTS right now).

- I'd love to see Sony/PD team enter Roborace. Look it up if you don't already know

- And lastly, I hope this doesn't turn into Skynet/Ultron

👏👏👏👏👏👏

 

[happycorey]

I want to see Sophy on fuel and tire saving mode.

These are complex variables that I think could be easily implemented to watch. Furthermore, strategy could be nearly perfect every time too. But these are variables I believe they will leave randomized so that there is a human element to Sophy.

In fact, outside of demonstration or teaching, I don't believe we'll ever see the true alien Sophy in GT. Only toned down versions.

 

[Pete_J_C]

picture the scene: you load up the game, buy your first car, and enter the sunday cup race. you overtake the field in the first lap and are gaining on p1 by half a second through every corner. you overtake p1, and they transform into... igor sophy

 

[MroczMR]

Incredible. I am very hyped about this. But I doubt it will come as "normal AI" in custom races or something. I think that future implementation for GT7 will be some kind of separate game mode like Sophy Challenges or something similar to Louis Hamilton challenges.

 

[05XR8]

Incredible. I am very hyped about this. But I doubt it will come as "normal AI" in custom races or something. I think that future implementation for GT7 will be some kind of separate game mode like Sophy Challenges or something similar to Louis Hamilton challenges.

I can see that.

 

[snowgt]

Imagine if Sophy was injected into ASIMO(or whatever robot) and sitting behind the wheel at a world tour event? How awkward would that be?


Edit: @Jordan
This is my question to ask Kaz: Is there potential for us to ever see Sophy sitting next to human drivers, at a future World Tour event?

Balance of Perfomansophy. Seems logical.

BoP(Toyota) = BoP + 0.5s

This is my question to ask Kaz: Is there potential for us to ever see Sophy sitting next to human drivers, at a future World Tour event?

Well, did you ever see a chess computer taking part in the World Chess Championships?

 

[MroczMR]

Also...

i love that Porsche livery:

 

[Verschmant]

I find the mostly underwhelmed reactions here quite surprising.

I watched the first race of the July event and the first race of the October event this morning and thought, that this AI is nothing less than a revolution for future gaming. Not only limited to racing games, but imagine football games or even action-adventures with deep-learning AI. I'm really looking forward to all the possibilites in other genres, too.

But coming back to GT: The maneuvers of Sophy are insane if you compare it to classic AI drivers from other games. I really enjoyed the AI in the F1 games in the past but even this is appearing so simple compared to Sophy. Look at how two Sophies are fighting with each other. They even pulled of the old Switch-a-roo in deliberate manner. Not because it was programmed to do that every lap, but because it was the best way to overtake in this specific situation. Also how it stays behind a real player in the second race and goes to the left in the last moment to do a suprirse overtake. This is what I've never seen in racing games before. But the best part is, that Sophy is not slowing down if you push her to the track limits. She knows she does not have to lift off the gas because the player leaves her enough space on the track. In every single racing game so far the AI lifts and you stay in your position.

Ofcourse there are flaws like the super-human speed. But Sophy is still in its research phase. I am 100% certain you can set different difficulty levels for this AI. Maximum corner speed for different lap times shouldn't be too hard to be implemented. And the last update is from October. The improvement from July to October was already spectacular, can't wait to see what she can do now.

All in all, this was the biggest improvement in racing games I have seen in the last 10 or 15 years. Can't wait for the update for GT7.

 

[TorqueHappenS08]

Video showing updated GT7 AI. Looks promising.

Including Ai making mistakes

 

[GTalaune]

Very good AI. I'm impressed . Just need polyphony to simulate the lost of grip when 2 tires are dirty , just as they do with 4 dirt tires.

I think this a byproduct of a setting that affects grip when going off road. When set to realistic, even when dipping 2 wheels in the dirt, you have to slow down or you are ****ed.

 

[Dave A]

Kaz talked about how the PP system calculations were based on an AI running a fictional track - is Sophy being used for that already ? If not, would there be any plans to update the PP system to use the Sophy AI within GT7 ?

Also, in the races they ran against real players, I noticed stewards gave warnings to Sophy and that the AI cars got track limit penalties as well - is it something the AI is able to learn from and adapt to currently ?

I doubt it, he said it's hopefully coming to GT7 and implied in the future not at launch.

Also to have a machine learning AI do something like that wouldn't be relaible, cars tested as it progressively learns will have infalted PP scores as the AI is driving better. It would need to be a fixed AI, likely one that doesn't make any mistakes and doesn't learn. There's no need for it to be learning, in fact it learning in that application would cause unrelaible results.

 

[YRS2703]

Everyone was saying that they were bored of catch the rabbit races in previous GTs.
They'll still be in GT7, except this time you're the rabbit and Sophy is the one hunting you down.

 

[Greatbee]

Awesome!

 

[Dave A]

Everyone was saying that they were bored of catch the rabbit races in previous GTs.
They'll still be in GT7, except this time you're the rabbit and Sophy is the one hunting you down.

Certainly at the level they were driving in the demonstrations it wouldn't take long to hunt me down.

It's very promising stuff, but it needs refining and obviousely it will need to be scaleable if it's every going to have mass appeal and use in races. Some of what it was doing wasn't very intuitive though, which is likely due to falws/limitations of the physics engine being exploited by the AI.

 

[LeGeNd-1]

Agend was running at 10 Hz, read #391 comment above yours.


I think that Sony doesn't need to train agend at every track for every car and condition. I think that situation is similar to DLSS: as I know DLSS was trained originally at some games and can be implemented to any game just by adding some DLL's to game folder and enable them.

---------

GT's career design is not for Sophy - races are only about 3 laps, you start at last position, so you have only 5-8 minutes to push forward to 1st. Sophy must press throttle slower when exiting turns, must brake far before turns. In other way your laps can not be faster 5 seconds than Sophy's laps

Thanks for that, I really wish they made the Nature paper available freely so I could read the whole thing before commenting.

Surprised that it's "only" running at 10 Hz, but if you look at it as 200% as accurate as the world's best, it's probably good enough. Or to put it another way, if you could slow down gameplay by 50%, it will make you seem to have superhuman reactions when it's run at real time. There's a cheating scandal at Trackmania a while back where players were doing this exact thing (i.e. slowing down gameplay to have super accurate inputs and thus better times).

Re: DLSS, I'm not a tech geek by any measure but I imagine it's a lot simpler upscaling a resolution than driving a car on the limit and battling with others trying to do the same. Possibly the AI could get faster at learning with each car/track combo, but they would still need to run a few simulations for each combo just to rule out any weird behaviour, and all that takes time.

If this AI makes it into GT, then I expect the old chase the rabbit format to be gone. They need to bring back qualifying, and have grid starts or closer rolling starts and just let the now competent AI have a proper race with the players.

 

[snowgt]

It says in the article that it needs 10-12 days for the AI to reach the top level. So, does that mean they would have to run the AI for every car and every track for 10-12 days? What about different weather conditions? How do differently tuned cars affect Sophy's performance...?

With 420 cars and 97 course variations, that's already about 40,000 such cycles. Different tunes and conditions could well make that number grow exponentially.

I'm sure they can run these simulations in parallel to some extent, but it might delay any addition of cars and tracks once it's introduced in GT7, if they need to run these before. Depends, of course, on how Sophy will be integrated in the game.

 

[Azerach]

I will have the chance to speak with Kazunori Yamauchi and the Sony AI team about Gran Turismo Sophy in the coming days — let me know if you guys have any specific questions about Sophy that you'd like answered. 👍

Will the sophy AI on gt7 run on the console or on Sony ai/pd servers? - if on console; how many instances of the ai a single ps4/5 can run? ie. Can you fill a grid with Sophie's for offline races

 

[RickyGT]

First of all, hats off to Sony & PD for this collab. Even though we've had the 2 previous studies showing AI driving faster than humans in a time trial setting, this is much larger in scope. I didn't expect to see AI machine learning outracing the best GT Sport drivers so soon. If they can implement this in the game in a realistic time frame, it would probably be GT's biggest contribution to the racing game genre since the first one's bringing simulation to the masses.

Now, a couple of my rambling thoughts (some might already have been covered in the thread, I didn't have time to read every post in detail):

- In terms of pace, the AI is stupendously fast. But we already know it's capable of doing this from the previous study at Tokyo Expressway. The AI can be perfect in its input all the time, and it can react in literally frame-by-frame of the physics engine. That's why it can pull off the 4 wheel drift kerb entry at DTS' hairpin. A human will never be able to do this with any sort of consistency, because our reaction time and input device is limited. In a twisted way, I find it very satisfying to see the top GT players feel the desperation us plebs feel when watching their replays. I've always said that the difference between a good driver and an alien is just accuracy. Well, the difference between AI and an alien, is just superhuman accuracy

- Taking the above then, an easy way to make the AI more "human" would be to limit their input frequency (e.g. if they are now running at 60 Hz, possibly tone it down to 5 Hz - human fastest reaction time is around 0.2 secs I believe). So a difficulty slider in-game would basically just be a frequency slider. Then they can add extra parameters like mistake probability, overtaking/defence aggressiveness, etc.

- Moving on from that, the weird way the AI drives shows that GT physics still has many unrealistic loopholes. For example the DTS kerb drift and using the grass on Maggiore last corner entry and Sarthe Dunlop curves. So in a way, having the AI push the limits will also help PD plug those loopholes in the future (hopefully). Back in GT5/6, drifting the corner entry used to be the fastest way to go round the track (just look up the GT Academy replays from 2012-2014). I thought this was finally gone, but it seems it's still there, just more difficult for humans to do.

- Same with abusing track limits (Sarthe pit entry lol). It's comical that PD thought training the AI on the track with the worst track limit exploits would make for good advertisement. Fix your penalty boundaries PD!!!

- All 3 races are done with no tyre/fuel consumption. It would be interesting to see if bringing these 2 factors would make the AI take more realistic racing lines, because 4 wheel drifting every corner entry would definitely kill the tyres quicker than the conventional line.

- For the racecraft, it is genuinely impressive to see the AI improve from the July race to the October one. However I think they pushed it a step too far in aggressiveness to compensate. The much higher number of penalties in the second round shows that. Because the AI is so much faster, it can pressure the human player into a mistake, and then just crowds/divebombs them at the nearest opportunity. Once it gets ahead, game over. It feels very forced and not very nuanced. It's not a classic one-on-one battle like with a similarly skilled human player, where you can be trading blows for multiple corners and not know who will come out ahead. The battle with Yamanaka at La Sarthe in the first round makes you think they can, but Yamanaka is just hanging back here for the final overtake into the Porsche Curves (common strategy for Sarthe in slipstream dependent cars). It's not the AI's brilliance that kept it in the battle for so long.

- In terms of commercial implementation in the game, this is where the biggest barrier lie IMO. The time (and computational effort) to run the simulations to get the AI to this point would be enormous taking into account all the cars/tuning/tracks/weather/tyre/fuel possibilities in the game. In the article it says it takes 300,000 km (10-12 days) for an AI to get to alien level. Just taking GT Sport, we have 338 cars and 82 raceable layouts. That gives 27,716 combinations. Multiply by 10 days = 759.3 YEARS of simulations I don't know about you, but I think none of us would be alive by then Of course you can speed up the process by having more computational power, but how much money are Sony/PD willing to spend to get this done? If it takes 1000 virtual PS4 machines to run one combo, to get everything done in 5 years you'd need roughly 152,000 virtual PS4s. And would you accept if shifting their budget means we have less car/track licenses for example? And whether that's good use of computing power and energy instead of solving other world problems/scientific research?


Future thoughts/possibilities:

- Could the AI be used for Tuning cars? No need for us humans to test manually, just let the AI run every permutation of every suspension/LSD/downforce/gearing setting to find the perfect tune. But as we've seen, the AI doesn't drive like humans so it could result in a very fast, but also very unstable car.

- B-spec! Let Big Brain Bob complete the races that are too difficult for you (and also grind money while you're AFK ). But if all the AI are super, then Bob is just back to being its average self again...

- Strategy. Similar to tuning, no need to run your own practice race to figure out the best strategy for FIA. Just let the AI pound round and figure out the best tyre/fuel management.

- More of a curiosity, I'd love to see the AI take the Red Bull X1 and Tomahawk X around Nordschleife to see what a perfect lap would look like. Human record is just under 3 minutes currently (in fact there's a TT in GTS right now).

- I'd love to see Sony/PD team enter Roborace. Look it up if you don't already know

- And lastly, I hope this doesn't turn into Skynet/Ultron

I think we will be the guinea pigs, and possibly this could be implemented since day one without they telling us is there. then they would have the biggest array of skills possible to train their algorithms with (and the most realistic one), so the time to perform all the different simulations would be drastically reduced.

It would be interesting to know if the car characteristics are "slider" variable that can simply be modified for each run or they require a complete training from the ground up for each car. with the tracks of course it's the latter, but with the car I don't know honestly

 

[Kalder]

All I'm hoping for is that Sophy doesn't turn out like Sophia who wants to destroy humans to make the world a better place...


Anyone remember this?:

 

[RickyGT]

I will have the chance to speak with Kazunori Yamauchi and the Sony AI team about Gran Turismo Sophy in the coming days — let me know if you guys have any specific questions about Sophy that you'd like answered. 👍

Yes: how do cars factor in? Is the car a variable which requires training or the algorithm can understand car parameters and immediately adapts itself to the car that is driving?

 

[Dave A]

Will the sophy AI on gt7 run on the console or on Sony ai/pd servers? - if on console; how many instances of the ai a single ps4/5 can run? ie. Can you fill a grid with Sophie's for offline races

The AI should easilly run on a base PS4 with a full grid. It's not the complexity of the AI running that demands special hardware, it's the development of it and training it as afst as they are doing (i.e. getting ti run thousands of scenarios simultaniously) which wouldn't be possible on a base PS4.

That doesn't mean it'se definitely coming to PS4, if it's not ready for a couple of years they might not be supporting the PS4 version of GT7 then. They might, but we've not been guarenteed anything in that regard.

 

[Nuschel01]

After looking at the hot lap videos to me it just looks like the AI is doing a TAS, as those inputs it does are impossible to do for a human with a controller/wheel.

 

[Kazuya720]

So basically, GT gets its own Skynet :/ .. great.

"When Sophy became self-aware, humans panic and "try to pull the plug," so "Sophy fights back" no fate I guess.

anyway.. the sun is shinning outside, I should go for a nice walk.

 

[lancia delta hf]

Proofs in the pudding, sounds great, but if it still feels "cheaty" in places and doesn't follow the same laws of physics as the rest of us humans when compared to wet weather, general cornering physics, speed power and momentum as well as tyre wear and self awareness it will all be for nought.

This. It's all typical PDI marketing until you see it first hand. Also it's not in GT7 at launch which means it could be 4 years before it makes it's way to GT7, what was wrong with just adding good AI at launch the traditional way?

 

[Kazuya720]

Kaz: Usually, when you go into a curve, the load is only on the two front tires, but Sophy has the load on three tires: two in the front and one in the rear as well. It allows the car to break as it is turning and is not something human beings would be able to do, conventionally.

I gon't get this? I mean, why I couldn't do the same as a human? As long as Sophy is only allowed to use the regular possibilies of a steering wheel, ****ft paddles and pedals...the outcome must be the same or still reproducable for a human. Its not a "Deep Blue" who just runs 100 millions of calculations per seconde to win in chess. Racing is a bit different

Howevery, I would be curious how Sophy reacts to greyish unfair, dirty driving which is not worth to get a penalty.

 

[Famine]

I gon't get this? I mean, why I couldn't do the same as a human? As long as Sophy is only allowed to use the regular possibilies of a steering wheel, ****ft paddles and pedals...the outcome must be the same or still reproducable for a human. Its not a "Deep Blue" who just runs 100 millions of calculations per seconde to win in chess. Racing is a bit different

In the rest of the quote, Yamauchi says that this behaviour is similar to top-tier real-world drivers like Verstappen and Hamilton, and somehow she's learned these same techniques.

So it is reproducible for a human.

 

[Scaff]

Kaz: Usually, when you go into a curve, the load is only on the two front tires, but Sophy has the load on three tires: two in the front and one in the rear as well. It allows the car to break as it is turning and is not something human beings would be able to do, conventionally.

I gon't get this? I mean, why I couldn't do the same as a human? As long as Sophy is only allowed to use the regular possibilies of a steering wheel, ****ft paddles and pedals...the outcome must be the same or still reproducable for a human. Its not a "Deep Blue" who just runs 100 millions of calculations per seconde to win in chess. Racing is a bit different

Howevery, I would be curious how Sophy reacts to greyish unfair, dirty driving which is not worth to get a penalty.

Neither do I, as it's complete nonsense and I hope something has been lost in translation.

The only way for you to only have load only on two or three tyres would be for the other one or two to have fallen off or you're lifting a wheel (as in tripoding an FWD or in older RWD cars managing to lift a front under acceleration).

As for allowing a car to brake and turn not being possible for human beings, umm trail braking?

I really do hope this is just a 'lost in translation' moment.