47 minutes ago, RALF9636 said:

Which was a workaround for a dysfunctional AI traffic engine from some 20 years ago. Now it is 2023 and I would expect ATC / AI to do better.

Then keep asking Asobo to fix it!
The complaints in this thread that two guys aren't doing what a team of 100-250 people at Asobo should be, is side tracking what looks to be a good app.

Edit: If you don't like the 'zapper' function, something similar to the excellent 'callsign blocker' as used in PSXT. Just enter a specific callsign and that AI is no longer displayed. The program also suggests a callsign to block.
 

Edited July 6, 20232 yr by F737MAX

24 minutes ago, F737MAX said:

Then keep asking Asobo to fix it!
The complaints in this thread that two guys aren't doing what a team of 100-250 people at Asobo should be, is side tracking what looks to be a good app.

The same would be true for ATC.

IMHO you cannot reasonably look at ATC and AI traffic separately. Both must go hand in hand. Otherwise you can only get halfbaked workarounds for persisting issues.

 

Edited July 6, 20232 yr by RALF9636

Just now, RALF9636 said:

The same would be true for ATC.

Big difference between delivering correct phraseology and navigating one aircraft compared to dynamically controlling 50+ AI aircraft, at different airfields and everything else that entails.
As I said, controlling AI has been tried before by small third party devs. The end results and subsequent performance loss meant we still haven't reached the holy grail.

2 minutes ago, F737MAX said:

Big difference between delivering correct phraseology and navigating one aircraft compared to dynamically controlling 50+ AI aircraft, at different airfields and everything else that entails.
As I said, controlling AI has been tried before by small third party devs. The end results and subsequent performance loss meant we still haven't reached the holy grail.

Let's see what VoxATC and FSHud will bring to the table when they get out of Beta.

Personally I am not interested in another chat-robot simulating ATC in an ambience of independent AI traffic.

 

7 hours ago, sd_flyer said:

And if there is no SID?

Then we give the altitude restriction

VFR - C-ABC, VFR to the East, fly runway heading, not above 2500ft, depart runway 08, squawk 1201.

IFR - ACA123 cleared to Sudbury airport via the flight planned route maintain 5000ft. Depart runway 30, climb runway heading for vectors, squawk 0506

On 7/4/2023 at 11:55 PM, Abriael said:

I like it.

I also would be surprised if Microsoft doesn't use ChatGPT for ATC in FS2024, given that they own it. 

And how would a conversational bot help with ATC?

23 minutes ago, RALF9636 said:

Let's see what VoxATC and FSHud will bring to the table when they get out of Beta.

Personally I am not interested in another chat-robot simulating ATC in an ambience of independent AI traffic.

 

So why don't you just buy FShud, instead of being so negative on this thread, which is not about FShud. 

4 minutes ago, Bobsk8 said:

So why don't you just buy FShud, instead of being so negative on this thread, which is not about FShud. 

I already did and enjoy it.

Nevertheless it's always good to have competition and I'd like to see B-ATC be a complete ATC addon as well, because it looks promising.

I see many people seem to be content with just another ATC chatbot, which is fine. It will still be better than default for sure. But if B-ATC stops there it will just not be for me (and some others), which is fine as well.

 

 

Edited July 6, 20232 yr by RALF9636

3 hours ago, Farlis said:

And how would a conversational bot help with ATC?

They would use AI/machine learning by exposing it to all recorded ATC instruction ever recorded.  From there AI would learn how most ATC communication operates, including subtle voice inflections and other nuances including content.  This would translate, potentially, to a Turing Test level quality in ATC communications...or something along those lines.  And when one flies into a specific region communications would mimic local variance in ATC speech etc.

3 hours ago, MarkW said:

Then we give the altitude restriction

VFR - C-ABC, VFR to the East, fly runway heading, not above 2500ft, depart runway 08, squawk 1201.

IFR - ACA123 cleared to Sudbury airport via the flight planned route maintain 5000ft. Depart runway 30, climb runway heading for vectors, squawk 0506

In US we only use clearance delivery only for IFR filing. We are given clearance limit. It's essential part in case of lost coms.So what you suppose to do in Canada when you lost coms (other than squawking 7600)

6 minutes ago, sd_flyer said:

In US we only use clearance delivery only for IFR filing. We are given clearance limit. It's essential part in case of lost coms.So what you suppose to do in Canada when you lost coms (other than squawking 7600)

You always have a clearance limit either in the SID or provided in the clearance.  Here is the verbage from one of the YYZ SID charts.

 

Communication Failure On recognition of failure 20 minutes or less after take-off and in IFR weather conditions proceed as follows: 1. Select transponder code 7600; 2. Beyond 10 NM from CYYZ proceed directly on course; 3. Do not climb above last assigned altitude for 5 minutes after recognition of failure, then; 4. Climb to flight planned altitude.

I think those guys have gold in their hands right now. It could possibly become an absolutely "must have" add-ons for all msfs simmers who don't use Ivao or Vatsim

1 minute ago, sdirand said:

I think those guys have gold in their hands right now. It could possibly become an absolutely "must have" add-ons for all msfs simmers who don't use Ivao or Vatsim

My guess is the vast majority of flight simmers do not use IVAO or Vatsim. I fly every day of the week, when I am not at work, I start flying after breakfast, and finish in the afternoon. I fly in the US normally, and the number of controllers anywhere in the US, during those times is almost none. Then there is the question, of where I want to fly. South America, Japan, New Guinea, etc..no controllers there either. Then if I was lucky to snag a controller , when I used to fly Vatsim, years ago, either at departure or arrival airport, having them sign off mid flight was pretty common. After a few years of this, I gave up, and went to ATC programs, starting with Radar Contact, then Pro ATC X, and finishing with PF3. I use ATC on every IFR Flight I do, normally around 12+ a week, from ATIS-clearance- ground-tower-departure-centers-approach-tower-ground. 

From what I have heard and seen with Beyond ATC, I probably will get it on day one. The voices are realistic, with plenty of variety, and not sounding like robots. The vectoring to final approaches seems good,and hearing other AI communications is a plus. As far as AI control, I have been using FSLTL for a couple of months now, and it beats hands down, AI programs that I purchased when I was flying P3D. The number of problems I have had on approaches, because of AI traffic, have been minimal. In FSLTL you can always press a button that stops traffic in your area, which , I have never had to use. I don't think eliminating AI traffic  problems, by restricting the number of AI aircraft, is a good solution. 

1 hour ago, Noel said:

They would use AI/machine learning by exposing it to all recorded ATC instruction ever recorded.  From there AI would learn how most ATC communication operates, including subtle voice inflections and other nuances including content.  This would translate, potentially, to a Turing Test level quality in ATC communications...or something along those lines.  And when one flies into a specific region communications would mimic local variance in ATC speech etc.

That's my biggest hope: That we see a turing level of AI use in ATC, AI traffic, the co-pilot, FSPax addons, and FBO/Dispatch/Company. Essentially a "Virtual Twin" of real world communication capabilities.

Feed AI 10 million recordings from Live ATC, cabin announcements, cockpit videos, and let it do its thing.

I would take a duplicate of Radar Contact at this point from a procedure level. I have asked for years why TC cant be duplicated. Does it need improvement, YES. But again I would take it without griping for now. The AI thing is an issue but no way a deal killer for me or a subscription base pricing.