Comment by sytelus | Hacker News Reader

Comment by sytelus | Hacker News Reader

Parent Full thread

Yes, wild ideas except for the "2100" part. The scientific and engineering breakthroughs have been slowing down dramatically in past 50 years or so. Science fiction from 60s and 70s is embarrassingly filled with predictions such as we would have colonies in space, computers indistinguishable from humans, flying cars on street, humankind without poverty and end of cancer by 2020. How close are we? If we go back in time to meet all these writers from 60s & 70s, the major breakthroughs we would have to show for are likes of Instagram and Angry Birds. Yes, I'm downplaying all these increase in computing power and better algorithms in search etc but guess what? Moors law is dead as we know it and search is still keywords in a textbox. To me it looks like we probably had 10 fold increase in human productivity during past 50 years but major portion of it goes towards "enhancing shareholder value" or wars rather than advancing science and technology. So unless this changes somehow, all the magic that can happen is still pretty far in distant, probably much beyond 2100.

You're also downplaying the creation of a global communications infrastructure that is nearly ubiquitous and unprecedented in its capacity.

Other than that, no progress at all. /s

But how much did the internet actually improve people's lives or change things? The major changes of our time are nowhere near as major as the ones of the last two centuries. Industrialization and mass production and trains and electricity, and massive population growth, and crazy amounts of scientific progress.

Almost all of the major improvements happened early on in the century, not so much in recent times. All the low hanging fruit has been picked. There is never again going to be a single technology that improves people's lives as much as mass adoption of electricity did (short of a singularity at least.) More efficient or even self driving cars compared to today's cars, will be nothing compared to what cars were for horses and trains. No single medical technology is going to have as big of an impact as antibiotics did. Cell phones are not nearly as big of an improvement as the first phones were.

Did the world really change and improve as much between 1970 and now as it did between 1930 and 1970? Did it change as much then as between 1890 and 1930?

Mass use of shipping container in 1970s alone has changed more lives than all those inventions from 1930 to 1970 combined.

Seriously? The way I see it, containers made transport of goods more efficient. Without containers, it was still possible to have all the goods, may be at higher prices, but possible.

> may be at higher prices, but possible.

Because lower prices don't affect anyone's lives?

The scientific institution has stopped really paying that much attention to space (or computation, in relative terms), and has started paying attention to biotech and materials science instead. Not a one of those science-fiction authors thought to suggest "get your DNA sequenced for $100" in their wildest flights of fancy.

Also,

> search is still keywords in a textbox

what else would you suggest? The "hard problem" part of search now doesn't seem to be anything to do with indexing information usefully, but rather getting people to be able to clearly communicate (and before even that, clearly decide) what they're looking for.

Current search would be pretty embarrassing to future generation as much as it is awesome to current generation. Right now you can find a page only if that page contains all keywords in it. In some cases, there are band-aids to include popular synonyms and correlated keywords or remove some keywords or suggest alternative keywords (spell check). If you think of it, this is very primitive, pretty much the same stage as baby tryieng to sit up at 6 months.

Let's say you read the book Black List by Brad Thor and you had an awesome memory to remember everything. Your friend won't ask you about this novel by keywords. They will ask you questions like "What is the best way to steal a car?" or "How would you get rid of strangers chasing you?" or "what breed of dogs are very vicious when you signal but very friendly otherwise?". And you would be able to answer those questions, not only by quoting related text but also summarizing and enriching the information. This what the ideal search would look like... pretty much how you can converse with computer on Star Trek. That's why I said current search is just "keywords in a textbox", still a very far thing from making science fiction a reality.

Oh, I see where I was confused; you're not actually talking about "search" here. "Search" is the ML term for, basically, "finding something you previously stored, by using parts of it." Google is a search company. Search problems ultimately boil down to the simulation of an ideal http://en.wikipedia.org/wiki/Perfect_hash_function generated from all the relevant data ever input. Search is an objective problem, best solved with lots and lots of Big Metal.

This, on the other hand, is recommendation--finding something you don't know you want, by clustering things together and modelling your desires and personality. Amazon and Netflix are recommendation companies. Recommendation problems ultimately "boil down to" a full Bayesian network modelling your own brain, so they can have the right context to answer you. Recommendation is a subjective problem, best solved, probably, with a neural implant or something which can literally "read your mind" to know what you would like the most even if you don't currently "want" anything in particular.

SF from the sixties and seventies is also filled with predictions which were not nearly ambitious enough.

(Two examples that come immediately to mind: I recall one of Heinlein's characters describing something that seemed a lot like a microSD card. It had a capacity of something like a million words. I got the impression that it was quite expensive. And in Neuromancer, "three megabytes of hot RAM" was contraband.)

(I think actually those examples are both from the eighties, but I don't think that's particularly important in this case.)

Please go read Kurzweil. Scientific advancement is dead on track on an exponential curve. Moore's law is still accurate.

Will it keep this rate as Kurzweil postulates? I have no idea. Until now, though, he's correct.

Comparing tech today with sci-fi of the past is wrong. Predictions fail both ways: there's lots of tech we do have that is missing in sci-fi.

Kurzweil has been getting it right because his near-term predictions are pretty safe bets, many of them true in some form even in 2000. For examples,

* Research has been initiated on reverse engineering the brain through both destructive and non-invasive scans.

* Autonomous nanoengineered machines have been demonstrated and include their own computational controls.

* People can talk to their computer to give commands.

He has crafted his long term predictions such a way that 8 out of 10 are pretty trivial and 2 most likely to go wrong. No wonder he often claims 80% success rate. For example, for 2020 he is predicting trivial stuff like this which has been in works already since he started predicting:

* Computers are embedded everywhere in the environment (inside of furniture, jewelry, walls, clothing, etc.).

* People experience 3-D virtual reality through glasses and contact lenses that beam images directly to their retinas (retinal display). Coupled with an auditory source (headphones), users can remotely communicate with other people and access the Internet.

* People are able to wirelessly access the Internet at all times from almost anywhere

* Nanotechnology is more capable and is in use for specialized applications, yet it has not yet made it into the mainstream. "Nanoengineered machines" begin to be used in manufacturing.

However his non-trivial predictions are almost guerenteed to be false:

* By 2020, there will be a new World government.

* The basic needs of the underclass are met.

* Most human workers spend the majority of their time acquiring new skills and knowledge.

* Phone calls entail three-dimensional holographic images of both people.

The point is predicting long term future is not just hard, it's impossible. But there is an expectation from long term future that sets our sense of progress and achievements.

I didn't mean the futuristic almanac style predictions. Kurzweil's main point is that mankind's evolution and abilities are exponential, or that Moore's law is not a special case but rather the norm in scientific evolution. His charts on stuff not related to transistor count do hold up for the near future.

More important than that, in this context, Kurzweil shows that scientific evolution has not stagnated or decelerated, in the last few decades, as the parent comment was proposing. Much to the contrary, the current pace is faster than in the past.

Reply on news.ycombinator.com