I'm with Ben, on this. To amplify a few points: Universities were invented 800 years ago as a social structure to allow old smart people to have the freedom to do research without worrying where their next meal is coming from, to recruit and train young geniuses to carry on, and to create walls to keep out the liars, cheats, morons and other destructive elements. It is the only social structure that I can think of to have survived for so long. The only other thing I can think of is the legal theory of evidence, which was invented around the same time.
That said, there are problems. Being an academic requires you to take a vow of poverty. If you are lucky, you can pay your bills, but just barely. Modern capitalistic thinking has helped damage the university; assistant and associate professors are abused. Tenure and publish-or-perish has created the crisis of replication, with no cure in sight. As to being a jack-of-all-trades, that takes time and patience. Grad school is designed to be a forced march to the top of the mountain-peak; there is no time to stop and smell the roses. Yet reading a bit of everything takes a very long time - a decade, or two or more. It is essential to obtain a strong foundation. Without that foundation, you become one of those people on facebook (or wikipedia) who ... I dunno... post smart-sounding drivel and nonsense about QM or general relativity or whatever. And then get into endless silly arguments about it. These people are jacks-of-no-trades, and anti-masters of all. Re: cog-sci -- do not confuse it with software engineering. They are very different things. Cog-sci is theoretical, mathematical. Software engineering teaches you how to build things in a safe, functional, dependable fashion. Re: AGI -- it requires research, not engineering. You can't assemble a team of engineers and say "build me an AGI". That said, let me contrast to "big science physics", and to "tabletop biochemistry". So in "big-science physics", e.g. telescopes, colliders, you have 10-100 million dollar budgets, teams of 20-500 people working for a decade to construct a scientific instrument. An army of grad students function as engineers, building the thing, with professional engineers providing guidance. In "table-top biochemistry", you mail-order some reagents and some bacteria, and a week later, you are crispr-cas-ing some genes in your kitchen. AGI research is mostly in the middle between these two. It's hard to do anything in AGI without "lab equipment", it's hard (impossible) to procure that "lab equipment", so you have to build it yourself. And it's almost impossible to convince someone else to build it for you (e.g. an "engineer") because they tend to mis-understand the problem, and build the thing they know how to build, instead of building what needs to be built. Perhaps Microsoft or maybe google has good "lab equipment" lying around for you to use, but its ... proprietary, and it might take a decade before they let you lay your fingers on it. Or something. The goings-on in those companies are opaque to me. --linas On Thu, Jun 25, 2020 at 8:48 PM Ben Goertzel <[email protected]> wrote: > > - Can an AGI be made without there existing a single person knowing the > relevant parts from Neuroscience, Psychology, Machine Learning, Philosophy > - and perhaps some more relevant parts from Computer Science? I'd guess > this question is impossible to answer, since we don't have an AGI yet; but > from the perspective of how teams work - does it become necessary for at > least one person to know the relevant parts from the various fields, so as > to be able to coordinate the team's efforts? I myself don't have much (any > perhaps) experience with leading teams; and hence, I wanted to seek > experienced opinions. In essence, is the "broad yet deep" background too > much to aim for? > > I think having one or preferably more people w/ that sort of > integrative knowledge is highly valuable for any AGI project > > > > > - Are there any opinions about whether a Masters in Cognitive Science is > worthwhile, or would I be better off pursuing the Masters in something more > specific? > > MS in Cog Sci is a great idea if you want to work seriously on AGI > > > - In case I'm better off pursuing the Masters in something else, is it > feasible to just do it from online courses? I've a strong bias towards > online self-directed learning - and I want to learn things without being > much involved in the research itself. For instance, I am learning machine > learning, but I do not want to invest myself in ML research. I'm also not > very convinced by the way academia exists today in the age of internet, and > think it can be improved. This goes off on a tangent though. For > self-learning AGI itself, there exist a ton of resources at > agi-society.org (the links seem broken in recent days though; > internet-archive helps); but I'd be very dubious if studying that would > help me pay my bills. > > > Universities suck badly in many ways, yet they are the most reliably > OK institution humanity has yet found for systematically fostering > research and education. Online learning is fantastic, but does not > quite substitute for the complex implicit learning that comes from > being part of a social group focused on learning and advancing a > particular area of knowledge (such as one gets from good old F2F grad > school, as least in non-shitty cases...) > > -- Ben > > -- > You received this message because you are subscribed to the Google Groups > "opencog" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/opencog/CACYTDBdnxU%2BSBDdY-5daWiKHdQuZgr752cGRjL6N7yGPv1C4tw%40mail.gmail.com > . > -- Verbogeny is one of the pleasurettes of a creatific thinkerizer. --Peter da Silva -- You received this message because you are subscribed to the Google Groups "opencog" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/CAHrUA35pwC%2Be%3D8Vcc-fgwe851MDHn795paaERs3y_0vM4AfwpA%40mail.gmail.com.
