Yes you can put multiple active devices inside a single tube, but it gets hard. "Multisection" tubes like the
Loewe 3NF exist. I've never dealt with them myself, but I suspect it's hard to keep each device separate because they all share the same vacuum. You can try and put physical dividers in, but unless you completely seal each one off into its own vacuum tube they will interact with each other. The
one circuit I've found for the 3NF shows the tube being used in a way where positive feedback from each stage is not a problem, because they all amplify the same signal.
Analog computers still use transistors. They're needed for amplifiers, multipliers and inverters (subtractors).
Without transistors you can only do memoryless computations, which in the analog computing domain is something akin to adding numbers. If all you can do is add numbers then you are not
turing complete, so you would not be able to make anything even resembling even video-game AI, let alone real intelligence.
I'll also jump into the definitions debate and be broad with how I use the word 'transistor'. To me anything that can amplify might be called a transistor, I don't think it should be a term limited to silicon devices. Especially since many transistors made today don't use silicon
Did they found a way to put more than just a single transistor in that vacuum tube or just found a weird work around
I'd vote both. Perhaps making the vacuum tubes flat -- like vacuum displays used on the front of DVD players and VCRs to show the time:
From there you might be able to place 'blockers' between elements, or weave them in clever ways to make fabrics of transistors.
The other thing they'd have to get around: self-repair. Tubes have glowy bits that degrade over time. A tube with many filaments will have to be self-repairing or have redundancy to survive.
Whilst I'm here, some other cool stuff with pretty pictures.
William's Tubes are a form of memory using CRTs and glowing phosphor:
Core memory is also lots of sex:
Selectron tubes baby!
