Prototype conscious AI?

[Morgoth99]

Okay, here's my question. In the High Guard book we have a TL16 conscious ship's system.

In the Robots handbook, Conscious brains are TL17.

Now it makes sense because even a small ship probably just has lots more processors than a robot brain.

BUT, the Central Supply Catalog book allows a prototype to be up to two TL's higher.

BUT... robot brains don't really use mass and the central catalog has mass for computers/electronics incrased by 10 at one TL and 100 at two TLs above SOTA.

So how would you work a prototype CI computer brain at both TL15 (two tech levels below 17) and TL16 (1TL below TL17).

 

[Terry Mixon]

Okay, here's my question. In the High Guard book we have a TL16 conscious ship's system.

In the Robots handbook, Conscious brains are TL17.

Now it makes sense because even a small ship probably just has lots more processors than a robot brain.

BUT, the Central Supply Catalog book allows a prototype to be up to two TL's higher.

BUT... robot brains don't really use mass and the central catalog has mass for computers/electronics incrased by 10 at one TL and 100 at two TLs above SOTA.

So how would you work a prototype CI computer brain at both TL15 (two tech levels below 17) and TL16 (1TL below TL17).

Personally, I ignore the mass for robotic brains that are prototypes. @Geir could provide a better answer, I’m sure.

 

[Condottiere]

If you make it large enough, and feed it lots of energy.

 

[Sigtrygg]

What could possibly go wrong with putting a prototype TL15 CI on a 1,200 warship...

 

[ross]

Because of the way computers are built in to ships today I’d say the mass is 10x0=0. However it’s completely arbitrary and the GM is capable of saying anything they want.

 

[Limpin Legin]

A conscious brain computer would need to model an estimated 228 Trillion synaptic operations per second in order to mimic a human brain. Given the brain's synaptic operations don't work like today's digital computers, not even the fastest digital supercomputer, consuming MegaWatts of power, can achieve this target.

Eg the Fugaku supercomputer, equipped with 158,976 nodes (48 cores per node), could mimic 26 billion synapses. This is roughly the size of half of a full mouse brain, but a full sized mouse brain is the size of an almond and the Fugaku supercomputer is the size of a large room.

 

[Sigtrygg]

And the energy requirements are several orders of magnitude different...

 

[Geir]

Personally, I ignore the mass for robotic brains that are prototypes. @Geir could provide a better answer, I’m sure.

In High Guard, computers have no displacement, not even the Core/xxx models, so that would be fair. Especially since Conscious Intelligence is just a program eating Bandwidth on a standard Core in that case. (whoever thought bandwidth was the right word... well, fine, but I also get upset when people mix mass and weight, or propellant and fuel, so I'm just a pedantic old pendant.)

 

[collins355]

The whole Prototype, TLs before TLs thing, was IMHO a Bad Idea (TM) and should be abandoned before further confusion and chaos results.

 

[Vormaerin]

I think it is often forgotten that TLs are quite broad. The Imperium has been TL15 for a while, but that doesn't mean that nothing has been significantly improved and iterated on. The iPhone 1 and the iPhone 17 are the same TL.

Bringing things entire TLs early can be interesting in moderation and to explain specific points of interest, but applied as a general rule can quickly become problematic.

 

[Limpin Legin]

The whole Prototype, TLs before TLs thing, was IMHO a Bad Idea (TM)

Well ... it is a good idea as science and tech do kind of evolve out of prototypes, etc. The thing 'wrong' with it is that it is too promiscuous like any maker can choose to prototype features just because they have wad of cash or spare space for carrying the extra weight. The Prototech theme really needs further criteria for what is going to make it happen in a more precious and rare kind of way (aside from quirks.)

 

[Morgoth99]

The whole Prototype, TLs before TLs thing, was IMHO a Bad Idea (TM) and should be abandoned before further confusion and chaos results.

I think one problem is that there sees to be a little confusion on just what a TL encompasses. In Singularity a TL 18 AI is fluffed as bieng barely understandable by lower TLs. Which implies that the jump between tech levels represents a change in kind, not degree, and yet in other places, like the prototype rules, tech levels seem to be far "narrower" in scope with TL 16 mostly being incremental improvements, and then from 17+ things start shooting up again.

For CI robot brains, it seemst hat your sitting in the "incremental" stage, since it's mentioned that CI's have existed (albeit as large programs in large installations) are extant at TL15/16 and robot conscious brains sit at 17, which TBH, isn't a big shift from the initial CI's in terms of impact. (Yes, a terminator is worrisome--but a self-aware robot can be that, and what is more scary, a CI in a body or in a Tigress class DN?) But 18 is specifically fluffed at "beyond Imperial understanding).

It's a bit late, and I doubt it'd be worth the page count, but I wonder if it might be better to divide TLs into two categories--conceptual TL's, for those changes that are very much a paradigm shift, and then within that have more granular "sub levels" that are more focused on "okay, we just developed gravitics, when does it go from super expensive to common).

On my question, I'm thinking of possibly just making them much more expensive, and having a high "failure rate" in constructoin which explains why everyone isn't rushing to a TL16 robot brain prototype when TBH, in 99 percent of cases, a self-aware AI does just fine.

 

[Vormaerin]

I think TLs are far more granular than they have a purpose in being and that it was done that way just to match the hexadecimal (ish) system used by other UWP stats.

Personally, I just worry about Primitive, Pre Industrial, Industrial, Digital, Early Space, Standard Space, Advanced Space, and "Nope, don't have that yet".

I don't have a mental vision of how a TL 13 society differs from a TL14 one, for example. So that difference doesn't do me (or my players) much good. Especially since a TL 12 world could have imported TL14 rowboats to deal with their needs, causing those distinctions to be very muddied anyway.

 

[Condottiere]

Consider what would be the minimum requirements for sustaining a given technological level.

 

[Morgoth99]

Of course one issue is there are still aspects of the TL system dating from when Traveller was written--and in that time, things like Nanotech, fabrictors, and such really didn't exist in hard sci-fi settings at least. They7 were more often moved to the realms of pulp and sci-fantasy. One of my old teachers, way back in teh 1980s mentioned and It stuck with me that "Science fiction is always the future of the era it was written." And that applies to traveller as much as anything.

 

[ross]

Of course one issue is there are still aspects of the TL system dating from when Traveller was written--and in that time, things like Nanotech, fabrictors, and such really didn't exist in hard sci-fi settings at least. They7 were more often moved to the realms of pulp and sci-fantasy. One of my old teachers, way back in teh 1980s mentioned and It stuck with me that "Science fiction is always the future of the era it was written." And that applies to traveller as much as anything.

Yeah I wrote a column or an article about Cyberpunk 2020 after a similar conversation called "an obsolete future".