Because after 8 years the idea of eating meat has no remaining appeal and is switching more to mild revulsion. Why would I order a substitute that is a close copy of that?
I'll still get them if there's literally no other vegetarian option on the menu, but that's rare.
I've been vegetarian for about 8 years and won't buy them and try to avoid them in restaurants because they're too meat-like. Unfortunately they've made good non-fake meat vegetarian burgers (black bean, wild rice, etc) harder to find.
It's a situation of "You know that thing you don't eat, don't like, and don't have cravings for anymore? We made something that tastes exactly like it. You're going to love it!"
I'm glad they existed when I first went vegetarian as they made the transition easier, but its a tough market when people will go off them in a couple years.
I still kinda wonder if they saw the success of the invite system for gmail (I remember a lot of late nights begging for an invite on various forums) and thought that it would work again.
The critical difference is gmail still worked just fine with hotmail, yahoo mail, aol, etc. Wave was useless if both sides didn't have it.
It might be part of the feedback loop, but from my experience it always starts at the company level.
I think I've been under a pay freeze for 4 of the last 6 years, and a capped 2% raise one of the others. No matter how much effort I put in, my wages would have stagnated.
I don't see any reason its not theoretically possible but I doubt it would be that beneficial.
You'd have to map the results back onto the traditional model which has overhead; and using shaky results as a precondition is going to negate a lot of the benefits, especially if its (incorrectly) predicting the part is already in the non-linear stress range which I've seen before. Force balances are all over the place as well (if they even bother to predict them at all, which its not always clear) so it could even be starting from a very unstable point.
Its relatively trivial to just use the native solution from a linear solution as the starting point instead, which is basically what is done anyway with auto time stepping.
I've been "lucky" enough to get to trial some AI FEM-like structural solvers.
At best, they're sortof ok for linear, small deformation problems. The kind of models where we could get an exact solution in ~5 minutes vs a fairly sloppy solution in ~30 seconds. Start throwing anything non-linear in and they just fall apart.
Maybe enough to do some very high-level concept selection but even that isn't great. I'm reasonably convinced some of them are just "curvature detectors" - make anything straight blue, anything with high curvature red, and interpolate everything else.
Ended up being fairly easy to look for - I compared the David Woodard list to the United States list and found instances where it claimed there was an article for the former but not the latter. Most David Woodard articles have a link to where he was born (United States), so an easy crosscheck.
I like Babbel a lot for reading/writing/listening but their speaking is a little weak. It's there but I find it pretty flaky - either so permissive it'll accept just about any sound you make, or so buggy it won't accept a single thing.
I haven't done a lot with it, but Pimsleur (https://www.pimsleur.com/) seems quite good for conversational. I've done a couple trials of it and plan to dive in when I finish my Babbel courses.
For conversational though you might be better off just finding an online tutor. 1 hour a week with a native speaker is probably more effective than any of the apps.
Duolingo is little better than a gacha game these days. It wasn't bad early on but its nearly unusable these days.
I like Babbel a lot for early levels - its gotten me back to a rough-B2 German level. I was probably approaching C1 in college, slipped to an A2 from lack of use, but I'm building it back up; and spent a while learning Norwegian from scratch a couple years ago on it.
Pimsleur is probably next on my list once I top out on Babbel to build up speaking.
Interesting idea on the ChatGPT sentence prompts. I'm not sure I fully trust it for that, but its worth a try.
Edit: Hm. Interesting idea. It's definitely a bit better at German than me, but its still making a handful of mistakes (as compared against other sources).