In order for an AI to take over my old job every detail and assembly would come from a catalog to start.  The trouble is the variation that can't be anticipated to catalog.

That applies to airliners, too, especially the wings.

However the products from Altair Corporation will be able to produce sensible analysis from a blank sheet soon enough.  A human will set the boundary conditions and at minimum define the load cases.  In the past Hyperworks produced optimized solutions but they were never producible by machining and never employed complete load sets.  But damn, it was still impressive.

I used Hyperworks for some solutions on MOP.  Afterwards I started coaching my young guys to learn how to use the software.  The interface at that time was not polished, but it worked well and was easy to learn.  Maybe some of them cared enough to pay attention.

There are airplanes in development for that mission.  They'll be ready in 10 or 15 years.

They will probably not look like the current "Loyal Wingman" from Australia or MQ-25 derivatives.  They will be similar size, speed, and range of current fighters, and they won't be "able to pull 20g's".

Every measure of airplane performance is weight sensitive.  Excess weight for vanity steals performance; sparkling performance beats sluggish, muddy, slouching every time.

On materials -

Structures are measured in four ways (not including producibility, corrosion resistance, maintainability, and others) - strength, stiffness, fatigue life (which is mostly dependent on the details), and the effective specific strength and stiffness (the strength or stiffness divided by the airplanes weight).

The strength of particular materials mostly comes in to play as dependency on the volume available.  If there is plenty of volume, the the least dense materials distributed in a manner consistent with buckling is appropriate.  As the volume, or depth of the space reduces, a stiffer and stronger, but denser material comes into play.

This leads to titanium alloy spars in the inner torque box of the F-15 wing, with very wide caps at the most inboard part.  The cost is a material about 150% the density of aluminum, with essentially identical specific stiffness (Young's Modulus divided by material density).

Then we are forced to carry landing gear, a necessary but icky feature that robs performance.  We don't want to allocate any more volume than absolutely necessary to squeeze them in, so high strength steels are selected to make the parts as compact as possible.  At a specific density cost of 3 times aluminum with the same specific density.

The program that blows off weight is doomed to fail.  As fatigue life requirements increase, weight avoidance becomes more important.  And, real life fatigue life analysis methods are not much better than in 1985.  We still impose scatter factor of 2, and sometimes more, to structure that will undergo fatigue testing, and scatter factor 4 on structure that will not be tested.  Crack initiation and crack growth lives required have both been increased.  Load spectrums are still what amounts to a guess, because the life is dependent on the order the loads are applied, and the magnitude of each load.

8.  BlueScreen