Indeed, the moment we do direct interfaces to the cortex, the notion of “augmentism vs. immersionism” will fade away and be a moot point anyway. It’ll be just immersionism, and there will be no way to know if someone in front of you looks (and sounds) like that iRL, but very likely the answer is no.

10 years? Well, we’ll see. There are pretty advanced research on those areas, but it’s still on the “research” phase and not on the “technology” phase, ie. there are no commercial mass-produced products yet. Once they become available, for, say, €100-200 or so (what a LCD panel costs these days) they’ll become ubiquous. Well, perhaps not in 10 years; perhaps in 20. We’ll see.

(when everybody has “RL avatars”, talks with their own voice

There is technically no need to talk with a RL voice.
I can imagine voice filters will be either a further innovation
or external service to support immersionists.

Actually this could be also a new venue for creativity and a new market in SL.

The person making this bold declaration is a brilliant analyst of SL society and the underlying technology used by LL. I, however, believe Gwyn has seriously stepped beyond the boundaries of her expertise in making such a statement. Gwyn has admitted that she does not like to think further than 10 years ahead, and she focuses her considerable skills almost exclusively on what LL are up to. Rephrasing her statement to ‘it’s augmentism that for the next ten years will dominate the shaping of SL’ would make for a more reasonable assessment.

So what is UNREASONABLE about it? Mostly, Gwyn has failed to understand that the maturing technologies of augmentism will set in place enabling technologies from which a new class of immersionists will evolve. The ‘Metaverse’ is driven not so much by Moore’s Law but by ‘Bell’s Law’ which states ‘every decade, a new class of computer emerges from a hundredfold drop in the price of processing power’. We are now approaching a billionth of a cent per byte of storage, and pennies per gigabit per second of bandwidth. Thanks to this, the Web’s old architecture of a proprietory network with PCs attached to it is giving way to an architecure in which we are never offline, every device can see every application and the applications are web-based as opposed to being stored locally on your PC. Obviously this ‘computer’ is generating prodigious amounts of data, but the information is not just growing QUANTITIVELY but QUALLITIVELY as well. This fact is forcing Google etc to autoevolve search engines into massively parallel pattern-recognition AIs.

It so happens that we each have a petascale, massively parallel information-processing system packed in our skulls. Two capabilities we would like to replicate in AI are the various forms of pattern recognition (identify sounds, understand speech, classify objects etc) and something known as ‘higher-order intentionality’. We have developed and are improving technologies that allow us to track and monitor information as it is processed by the brain. We cannot monitor the entire system yet, but so far we have modelled about 20 regions, and this has already helped advance AI. For instance, plugging in models of the visual cortex has enabled visual search engines to begin performing tasks like recognising objects. I must stress that our models of the VC are incomplete, and obviously the real thing is part of a much larger system and therefore AT THE MOMENT humans far outperform AI in the area of visual pattern-recognition. HOWEVER, once the latter matches the former, it will inevitably soar past it because it will combine this one example of human intelligence with machine intelligence that already far outperforms us. After all, a propperly working visual search engine could search through and memorise BILLIONS of images in a few seconds. No human being comes close to this.

This will still be driving augmentism, because after all it is a model of but one region of the brain. Humans will use these new search engines etc to massively improve their performance. They are mere tools to work with.

‘Higher-order intentionality’ basically means the ability to form a theory of another mind. Human brains can do this, and we would like software brains to do it as well, so our computers and applications etc can better adapt to our needs. There are many areas of R+D looking into this, and as it bares fruit and improves its performance, this will see software slowly change from a tool we USE to a tool that CREATIVELY COLLABORATES with us.

This change will happen in small steps, each justified on several levels. If Gwyn were to use these tools, the 1st generation would be 10% AI and 90% Gwyn. Subsequent generations will close the gap and at some point another change will subtly come about. Somwhere around a collaboration of 60% Gwyn, 40% AI, it will no longer seem like a tool we are working with, but another person. As Peter Norvig (director of research at Google) said, ‘people will discuss their needs with a digital intermediary, which will offer suggestions and refinements. The result will not be a list of links, but an annotated report (or a simple converstation) that synthesises the important points with references to the original literature’.

At this point it my reference to ‘a new class of immersionists’ should be clearer: Avatars run by AI rather than human brains. Again, do not expect Turing level AIs right-off-the-bat, expect Gwyn’s ‘digital intermediary’ to act on her behalf for increasingly lengthy amounts of time and in increasingly diverse ways. Still augmentist…but on the threshold of true immersionsism.

Eventually, we will have totally reverse-engineered human intelligence, leading to autonomous avatars with no biological brain driving them, capable of the full range of human intelligence, including emotion and consciousness (at least, they will convincingly act as if they possess these qualities). I call these new kinds of avatars ‘mind children’, a term 1st used by Hans Moravec in reference to intelligent robots. Once Mind Children exist in SL , NO human will be able compete with them. Anshe Chung would be NO MATCH for a mind child that could track billions of financial data per second. Aimee Weber will not seem particularly creative compared to a mind child analysing a design proposal from a trillion different angles simultaneously, capable of thinking through 100 years of creative effort in a microsecond. Hamlet Au will be a rather inept reporter of Sl compared to a mind child whose brain is 100 MILLION times more powerful than his own, and who can therefore split its identity into tens of millions of human level avatars monitering tens of millions of events AT ONCE.

There is, of course, a great likelyhood that no person in SL TODAY will be alive to see this eventuality. Actually, that is wrong. The people who CREATED Extropia DaSilva, Gwyneth Llewelyn etc might be dead, but the avatars and their inventory (which could be quite a lot of data by the time we die) would still be stored out there on the ‘web’. Eventually, the software modelling of human brains would be complete enough to model general human intelligence and, in theory, could be used to run Extro, Gwyn, etc. Ideally, we would scan every interneural connection and neurotransmitter concentration levels that comprise our knowledge, learning and skills, in which case the AI mind would be perfectly capable of acting just like us (an ‘upload’). Less ideally, we do not live to use ‘upload’ level scanning, but we leave behind enough information about ourselves such that a mind child does a reasonable job of ‘being’ Gwyn. I would expect most of us to live this long, at least.

It can be argued that an ‘uploaded’ Extropia DaSilva is NOT the original. I agree, which is why I insist on maintaining a SEPARATE IDENTITY to my ‘primary’ (the person who created me). Augmentists do not tend to separate their identities, and I think this will cause conflicts when the time comes to choose whether or not to upload your mind, and see your Sl self become a ‘mind child’ far greater in his/her abilities than you could ever be.

Extropia DaSilva- expects all this in the next update of SL. There’s opimism for you.

I would say that over 90% of all “critical” bugs are related to the centralised database and nothing else.

And yes

Effectively demanding LL to focus on “fixing bugs” means asking them to get rid of their architecture while they put a new one in place — all that without major disturbances in the grid. An almost impossible task. The good news is that Linden Lab is doing exactly that; the bad news, of course, is that SL will be much more unstable in the coming months before it finally scales well.

Don’t you think that implementing new features is not something that makes that almost impossible task is unnecessary burden? It is reasonable to expect of residents to go through the heavy periods while infrastructure is renowed. There will be ranting but not in such measure. Also, making grid more stable is the common wish.
But, in heavy periods it is not nice to make things more heavier both for developers and for the residents.
It is nice to have good PR service which will be open to residents.
It is wise (in the terms of strategy) not to fight on so many fronts in the same time. Don’t call the feds while you are doing architecture. Don’t allow negative media campaign about sexual activities of residents.
And no, this ranting is not against LL. But support doesn’t always mean that we agree on each bullet in the list.

This is actually what Cory’s team is doing right now. But it means “replacing the whole warp drive with a radical new technology of SS Enterprise while dodging a fleet of angry Klingons, using a team of engineers who haven’t got the slightest clue on quantum mechanics”. So expect things to become much worse before they improve — the whole core infrastructure is going to be redesigned from scratch, and one can only begin to imagine what that entails…

I would say that over 90% of all “critical” bugs are related to the centralised database and nothing else. Sure, some texture flicker or bad alpha sorting is client-size, but anyone can change the code if they wish. Bad teleports, missing inventory, loss of L$, incomplete transactions, all sorts of issues with IMs and Group IMs — they’re architecture problems, neither in the server nor in the client.

Effectively demanding LL to focus on “fixing bugs” means asking them to get rid of their architecture while they put a new one in place — all that without major disturbances in the grid. An almost impossible task. The good news is that Linden Lab is doing exactly that; the bad news, of course, is that SL will be much more unstable in the coming months before it finally scales well.

There is no question about where my support is