Mainly Live, but I use Reaper sometimes.
Yeah, Reaper UI is a little old skool winampish, but it does do what is says on the tin.
1 Like
I asked it how to create an accurate flight model.
To simulate an accurate flight model for Microsoft Flight Simulator (MSFS), you typically follow these steps:
Gathering data: Collect real-world data on the aircraft you want to simulate. This includes information such as aerodynamic coefficients, control surface deflection ranges, engine performance, and other relevant parameters.
Aerodynamics modeling: Use the gathered data to create an aerodynamics model. This involves representing the forces and moments acting on the aircraft during different flight conditions, including lift, drag, thrust, and weight. Complex mathematical equations and computational techniques are used to model these aerodynamic forces accurately.
Flight controls modeling: Model the behavior of flight controls, including the response of control surfaces such as ailerons, elevators, and rudders. This is done based on the aircraft's control system design and the physics behind how these surfaces affect the aircraft's motion.
Systems modeling: Simulate the various aircraft systems, including engines, fuel systems, electrical systems, hydraulics, and avionics. These models ensure that the simulated aircraft behaves realistically in terms of its systems' operations and their interactions with flight dynamics.
Integration and validation: Integrate the created models into the MSFS simulation environment. Test and validate the simulated flight model against real-world flight data and pilot feedback. Iterative refinements may be necessary to improve the accuracy and realism of the simulation.
Fine-tuning and adjustments: Adjust the flight model parameters and characteristics to match the specific aircraft being simulated. This can involve tweaking values to reflect specific aircraft performance data or incorporating custom modifications or avionics systems.
User feedback and updates: Continuously gather user feedback and monitor real-world changes in aircraft performance or systems. Use this feedback to make necessary updates and improvements to the flight model through regular software updates or patches.
It’s worth noting that creating an accurate flight model for MSFS requires expertise in aerodynamics, aircraft systems, programming, and access to reliable aircraft performance data. It can be a complex and time-consuming process that often involves collaboration between developers, pilots, and subject matter experts to ensure the highest level of accuracy and realism in the simulation.
As usual, ClippyGPT sounds supremely confident but fails to deliver the full goods, leaving out key information. The ultimate “idea guy (or girl)”, leaving it up to someone else to do the work only to discover that a couple of chapters were skipped in the instruction manual.
Just go direct to the source, skip the stochastic parrot: The Flight Model
1 Like
The response seems like an excellent way to organize the team at Asobo or any other flight sim.
Even if it was 100% correct (which it’s not btw) that’s using the “even a broken clock is right twice a day” to assume the clock is good for 24 hours a day time reading. That’s extremely risky to do for something that requires both trust and accuracy in its results.
However, if you use it to write a creative backstory for your flight plan of 3 people, that’s a great use case where it works fine - assuming it doesn’t take liberties with the flight plan details. You’d still need to check them.
Please read Just Calm Down About GPT-4 Already - IEEE Spectrum for some insight into the issues from an AI & Robotics veteran. eg: What the large language models are good at is saying what an answer should sound like , which is different from what an answer should be .
It’s an important issue, to understand exactly what these new tools are and not just believe the vested interests boosting it along on the AI Hype train. There’s a bridge out just around the next bend, don’t be on that train when it crosses it, because you know the vested interests will have got off it long before and moved onto the next thing they can promote😂
Also, to be clear, LLM’s are/will be a useful addition to the AI toolset for certain applications (once constrained), just like computer vision is. But they are nowhere near AGI (Artificial General Intelligence) and are just one tool among many to advance AI R&D and systems development.
Fwiw here’s a good article that explains equally well how good AI can be when implemented properly, and also how bad it can be when not. Six Questions Every DOD AI and Autonomy Program Manager Needs to Be Prepared to Answer
2 Likes
You love articles don’t you. Why don’t you let it rest bud? Everyone can choose to believe or predict or be on a hype train or not be on a hype train.
Time will tell if it’s only a “tool” or will it evolve to much more.
Let’s wait it out.
True believer huh. Good for you!
2 Likes
Just waiting it out and happy to be living to see it 
I didn’t say anyhting different here:
You don’t know me personally and that’s totally fine and normal. But everyone who does in real life, know that i am a realist who lives by the objective code. Heck i’ve been told i make a great judge.
My point is, i always like to wait things out and watch them unfold while staying objective all along.
Now opinion and intuition are a totally different thing. My INTUITION tells me that this thing will evolve more.
Anyway…
Everything evolves, that’s a given.
That doesn’t mean you ignore the risks and the pros/cons of a technology in the now, the very near future, or even the far future.
These are already being deployed into production systems, without proper oversight or even disclosure in some cases that you are talking to an AI or reading its output. That’s a problem, given the trust and accuracy issues this tech has currently. But that’s not stopping people deploying it into production, they don’t care.
The current move to impose some regulation on AI is only due to smart committed people making the effort to raise these issues, persisting in the face of Vested Interests who were telling them to shut up with “it’s not a problem, you’re being hysterical”, then a few months later those vested interests having to do a 180 in the face of overwhelming evidence it is a problem.
Sure, wait things out and watch them, but be please be aware too.
1 Like
I agree. Raising issues is important and i’m glad it’s being raised. This is one of the (very few) positives i always give to our current era: Awareness, smarts and caution.
But how about thinking about it this way: Maybe the warnings and cautions are a good thing to make this thing better in the future, by staying “obedient” and under control? That’s a good thing.
For example, we all know the dangers of Instagram and how much it affects mental health in teenagers. But those dangers weren’t being “aware of” if it wasn’t for the warnings and cautions being raised.
In other words, i really like the non-passiveness in our current generations and in my opinion it’s the main thing that’s preventing a WW3 (drifting away here sorry).
Someone finally explained to Sam Altman, CEO of OpenAI, exactly how ChatGPT really works.
The latest chapter in OpenAI having to walk back it’s hype train about ClippyGPT, LLM’s, superintelligence, AGI etc and acknowledge the issues, risks, threats, and actual opportunities that they were refusing to engage with.
Those “more sophisticated” critics were in fact AI researchers with deep experience, who were doing very pragmatic critical analysis (ie: they were not “AI critics”, but AI analysts and researchers asking the right research questions) but were denigrated and put down by Altman et al until the evidence became undeniable.
As I said:
Anyway, onwards and updwards…with awareness.
OCT 2025. I just had ChaptGPT optimize my UserCfg,Opt file. In VR it set my visual options to a higher level, but my fps performance increased. A lot! So now I have better performance AND better visuals. It took 20 minutes because I never have used ChatGPT to optimize .cfg files before.
The exact question to ask ChatGPT.com I found in a thread on Avsim. In that thread it gave the exact question to ask the AI in order for the AI to optimize 3 files of MSFS 2020 or 2024. But I only asked AI about one file: UserCfg.Opt. I did not optimize the other two (because it was late and time for bed).
3 minutes later I started a flight in my normal test flight area that was the best VR experience in any flight ever. Way more fps, better visuals, and better tracking in the headset (Pimax Crystal Light).
I have no clue as to what was messed up in the UserCfg.Opt file that I have been using for months now. But ChatGPT fixed whatever it was when it was aksed to optimize that file.
I don’t think we are allowed to link directly to that message thread on Avsim. So I won’t try. MSFS gives us all Settings Options in MSFS. All ChatGPT did was to make better choices that Asobo has given for us to make in the sim.
Closed.
As per the Code of Conduct:
Improve the Discussion
We recognize that our forums are an official place for feedback on Microsoft Flight Simulator, whether it is positive, neutral, or negative. Help us make this a great place for discussion by always adding something constructive to the discussion, whether big or small. We do not remove posts for being negative; however, we do remove posts that are toxic or non-constructive (click here for specific examples on how to improve the discussion). Non-constructive posts do not add to the community and include but are not limited to:
…
- Using AI to generate spam-like posts
…