And...

that's yet another specific characteristic of the X-Plane Ecosystem, but it's particularly rewarding to know it starts at the very core of the development behind this Great Flight Simulator, at LR...

For years I've been able to directly address Austin regarding the most varied subjects regarding X-Plane, specially in the flight dynamics area.

Austin is a brilliant mind, dedicated to his causes, and when a decent and as well documented as possible is presented to him he does his best to find a time slot and dedicate time to it.

Sometime ago, more than a year ago, I started posting about a strange effect experienced with some prop aircraft in Xp12, specially multi-prop. With their props rotating, say, CW, there was a notorious tendency under a considerable range of power, speed, pitch regimes, for them to roll to right, which is counterintuitive.

After showing Austin this strange phenomenon, he promised to look at it as soon as it would be possible given the many tasks under his hands. Sooner than later he started looking at it, and made adjustments, but they only made their way into internal betas this month, and Austin reviewed the solution and has been at it to fine tune the FM for one of the next public beta releases.

Now, it turned out to be, not really a "bug" but rather the outcome of a detailed feature which is implemented natively in X-Plane 12 for the simulation of propwash effects over the various surfaces that lay on it's way, like the wings, the fuselage, the tail surfaces, etc... All of that is taken into consideration to a level of detail that is truly AMAZING, but the fine grain of the simulation required for a more precise outcome in terms of moments "created" by that swirl of air leaving the prop(s) wasn't there yet, and needs to be added when the necessary "FM cycle slots" are implemented, optimizing memory and CPU / GPU use.

In the mean time Austin made the necessary adjustments to the present model, already one of the most sophisticated I experienced in any flight simulation I ever used, creating a compromise solution that gives us a kind of effect which is real, and yet not making it impact the simulation negatively, while he tries to continue to improve it and make room for even more fine grained calculations in future updates, or for the next version of X-Plane.

Being able to experience how passionate Austin and the whole LR Team are, and even more feeling that we can directly address them and receive their attention, provided we find acceptable forms of defining problems, limitations, features we would like to see implemented, etc... makes X-Plane truly **unique** in the World of general purpose desktop flight simulation !!!

Edited March 9Mar 9 by jcomm

Though I am not an "X-Planer", I agree with your statements about Mr. Austin! He is brilliant.

I've only recently joined the AVSIM clubhouse, I've been snooping around here for years, and I enjoy the family community of flight simming.

Direct support from a developer to the customer is an extremely rare thing.

- Don

3 hours ago, jcomm said:

Austin is a brilliant mind

That man is brilliant. A couple of weeks ago I watched a video he did about the stock market and actually adjusted my thinking on investing. He is one of those souls similar to Elon Musk who has a gift, talent and discipline to mentally perceive in the abstract then but their perceptions into practical application. 
I sincerely hope he never tires developing XPlane since I’m sure our user base is quite small compared to MS sim.

We don’t forget and do appreciate his talented team as well.

It concerns me as to what impact AI may have in, or against, this excellent team.

16 minutes ago, philmurfin said:

It concerns me as to what impact AI may have in, or against, this excellent team.

That’s interesting. Why would it concern you? Any ideas how AI would negatively affect Laminar Research?

1 hour ago, Patco Lch said:

That’s interesting. Why would it concern you? Any ideas how AI would negatively affect Laminar Research?

Seeing how AI hasn't done very well with a bunch of Nvidia drivers, I can see the concern if proper human oversite isn't introduced going forward with AI.

AI isn't as much a threat as a lot of people might think it is.  AI is to coding what software, like Substance Painter is to texturing.  It's a tool to get things done quicker.  It's only as good as the person using it.  

17 hours ago, Patco Lch said:

That man is brilliant. A couple of weeks ago I watched a video he did about the stock market and actually adjusted my thinking on investing. He is one of those souls similar to Elon Musk who has a gift, talent and discipline to mentally perceive in the abstract then but their perceptions into practical application. 
I sincerely hope he never tires developing XPlane since I’m sure our user base is quite small compared to MS sim.

We don’t forget and do appreciate his talented team as well.

Yes, but the sound quality of that video is very bad.

AI should really help Laminar to speed up development.

7 hours ago, philmurfin said:

It concerns me as to what impact AI may have in, or against, this excellent team.

Gödel is our friend... Don't worry that much 🙂

This being said, also as a joke OFC, you rise a very valid point, and for sure something that most of the technological players of this World might have to face, specially those with smaller IT businesses, and not only, businesses may have to learn new ways to deal with...

Thankfully, if that problem turns out to get to a threatening dimension, we could use AI to solve it as well, or..., couldn't we ?

Anyway, LR is still made up, most of it, I guess 100%, of human agents only 🤣

Edited March 10Mar 10 by jcomm

21 hours ago, jcomm said:

And...

that's yet another specific characteristic of the X-Plane Ecosystem, but it's particularly rewarding to know it starts at the very core of the development behind this Great Flight Simulator, at LR...

For years I've been able to directly address Austin regarding the most varied subjects regarding X-Plane, specially in the flight dynamics area.

Austin is a brilliant mind, dedicated to his causes, and when a decent and as well documented as possible is presented to him he does his best to find a time slot and dedicate time to it.

Sometime ago, more than a year ago, I started posting about a strange effect experienced with some prop aircraft in Xp12, specially multi-prop. With their props rotating, say, CW, there was a notorious tendency under a considerable range of power, speed, pitch regimes, for them to roll to right, which is counterintuitive.

After showing Austin this strange phenomenon, he promised to look at it as soon as it would be possible given the many tasks under his hands. Sooner than later he started looking at it, and made adjustments, but they only made their way into internal betas this month, and Austin reviewed the solution and has been at it to fine tune the FM for one of the next public beta releases.

Now, it turned out to be, not really a "bug" but rather the outcome of a detailed feature which is implemented natively in X-Plane 12 for the simulation of propwash effects over the various surfaces that lay on it's way, like the wings, the fuselage, the tail surfaces, etc... All of that is taken into consideration to a level of detail that is truly AMAZING, but the fine grain of the simulation required for a more precise outcome in terms of moments "created" by that swirl of air leaving the prop(s) wasn't there yet, and needs to be added when the necessary "FM cycle slots" are implemented, optimizing memory and CPU / GPU use.

In the mean time Austin made the necessary adjustments to the present model, already one of the most sophisticated I experienced in any flight simulation I ever used, creating a compromise solution that gives us a kind of effect which is real, and yet not making it impact the simulation negatively, while he tries to continue to improve it and make room for even more fine grained calculations in future updates, or for the next version of X-Plane.

Being able to experience how passionate Austin and the whole LR Team are, and even more feeling that we can directly address them and receive their attention, provided we find acceptable forms of defining problems, limitations, features we would like to see implemented, etc... makes X-Plane truly **unique** in the World of general purpose desktop flight simulation !!!

Should post that on other forum, like to see janovs reaction, hes being saying its fine all along

4 hours ago, mjrhealth said:

Should post that on other forum, like to see janovs reaction, hes being saying its fine all along

He is aware, and I believe he understands the problem, but interprets it in a different way, and..., honestly, I can understand his pov too, even if I sometimes feel a bit irritated when someone tells me something I find evident to experience / see / test, isn't there.

Case is, I believe, Janov tried to emphasize that the effect is actually resulting from an advanced simulation feature that hasn't yet been fully modelled and thus can bring some unwanted side effects. This is probably why he fought the designation of the effect as a "bug"...

It's not really, from a software development and physics modelling pov a "bug", but rather an incomplete model of the associated effect, which is actually realistic, and modelled in X-Plane most probably much more in detail than in many other general purpose flight simulation platforms.

The effect is very well identified in many text books, and in particular in various NACA reports, and Austin, as usual, always tries to bring into X-Plane as much realism as possible, within the limits imposed by the "need for speed" given the fact that complex calculations can contribute to a slower / more stuttery sim, and many users look at the graphics in the very first place, and want cars on roads, rocks, fauna (no pun intended...) and the like more than into the realism in terms of FLIGHT simulation ...

So Austin was aware of it early 2025 when I first pointed it to him, changed the code around March 2025, but due to the many development fronts under LR's hands, it took a while before it finally got into the beta release line... where it is now internally.

I had the chance to try one of the latest versions and find it satisfactorily tamed down, although Austin himself will get an even better solution when he finds the time to optimize the calculations of this sort of rather complex stuff without impacting the sim performance.

I believe Janov was well aware of this, and so he didn't want to see it debated as "a bug", although it's true that in a few occasions he did try to make me think I was dreaming 🙂

(he even hinted at me going War Thunder again... and he wasn't far from the truth since, it's not WT, but now AoT that is taking away 90% of my simming desktop time... I always try to find "shelter" in another type of flight simulation genre whenever I feel tired or less motivated by my mainstream sims, and I fight really hard not to hit that tempting UNINSTALL button 🤣)

Edited March 10Mar 10 by jcomm

4 hours ago, GoranM said:

AI isn't as much a threat as a lot of people might think it is.  AI is to coding what software, like Substance Painter is to texturing.  It's a tool to get things done quicker.  It's only as good as the person using it.  

And the problem is that most people using AI are not good enough and the result of that usually finds its way into databases or the open, where it is used for training by other AI, which in turn produce lower quality content that gets used for training by other AI...it's basically an endless spiral of diminishing quality.

In a closed environment, AI is a helpful tool though. I wish every company would run one that was trained on all of its data and documents and could save oneself from having to dig through heaps of PDFs to find something or generate reports.

2 hours ago, Bjoern said:

And the problem is that most people using AI are not good enough and the result of that usually finds its way into databases or the open, where it is used for training by other AI, which in turn produce lower quality content that gets used for training by other AI...it's basically an endless spiral of diminishing quality.

In a closed environment, AI is a helpful tool though. I wish every company would run one that was trained on all of its data and documents and could save oneself from having to dig through heaps of PDFs to find something or generate reports.

They trained them on ebooks they pirated from tpb.

Meta is currently in court argueing how that is cool and normal:

https://arstechnica.com/tech-policy/2025/02/meta-torrented-over-81-7tb-of-pirated-books-to-train-ai-authors-say/

https://torrentfreak.com/uploading-pirated-books-via-bittorrent-qualifies-as-fair-use-meta/

 

18 hours ago, philmurfin said:

It concerns me as to what impact AI may have in, or against, this excellent team.

So far, it's only extended to making memes about each other. 

5 hours ago, mSparks said:

They trained them on ebooks they pirated from tpb.

Meta is currently in court argueing how that is cool and normal:

https://arstechnica.com/tech-policy/2025/02/meta-torrented-over-81-7tb-of-pirated-books-to-train-ai-authors-say/

https://torrentfreak.com/uploading-pirated-books-via-bittorrent-qualifies-as-fair-use-meta/

 

“Hey, man… I just found this bag full of money with First National Bank printed on the outside. Right after an armored car with police chasing it went by. 

How was I supposed to know it was stolen…?”🤣

The problem is that the books were pirated in the first place and put out where anyone could get them. 

So, everyone is gonna (“has already”)grab stuff like that for AI training. 

But no way is it Fair Use. And now that the cows have left the barn, Zuckerberg should peel off a few bills to license their use from the authors/publishers. 

I hear he just paid a record $170,000,000 for a new (additional) home in Miami.

 

Edited March 10Mar 10 by UrgentSiesta