4 ms·

Would you rather watch cabletv or YouTube? Would you rather fish out a map and look for directions or simply ask your car for navigation directions? Would a c

Would you rather watch cabletv or YouTube? Would you rather fish out a map and look for directions or simply ask your car for navigation directions? Would a cancer researcher use his library card and read literature or simply Google it? Would you like to solve the protein folding problem? All of these have been paid for by advertisements.

echelon 5y ago
Outside of protein folding, each of these has been solved for 15+ years. And protein folding isn't exactly solved.
- random314 5y ago
  YouTube was solved 15 years ago?
  - echelon 5y ago
    > February 14, 2005; 16 years ago https://en.wikipedia.org/wiki/YouTube https://en.wikipedia.org/wiki/YouTube It had a lot of content back then. Work since has gone into scaling and transcoding, but much more work has gone into ads, platform stickiness, music rights management, abuse detection, etc. Not really solving problems at the forefront of humanity.
goodpoint 5y ago
The following technologies have been researched between 1900 and 1990 almost entirely with *tax money*: Semiconductors, computing theory, GPS, GSM, satellites, airplanes, X-ray, optical cables, analog radio transmission and spread spectrum, industrial chemistry, solar power, nuclear power, vaccines, genetic research, and much more. Private companies putting together a search engine or uber or deliveroo are sitting on the shoulders of giants.
simion314 5y ago
>Would you like to solve the protein folding problem? Isn't Deep Mind a british company that Google bought it? Seems to me that most of the credit should go to UK society/education and just a bit to advertising industry.
- random314 5y ago
  What was the endgame for Deepmind, a company with no revenue stream? Geoff Hinton was also hired via acquisition.
  - simion314 5y ago
    I do not know, but let's give the credit to the educational system and country of origin too , I see a lot of people give credit to US or SV for shit that is bought by the giants. Or I see arguments like "why there is any non US company for X ?" and the implication is that SV is the best where in fact SV has the money to buy anything that is promising and sometimes kill it but sometimes like in DeepMind case managed it good enough ot to waste it.
    - random314 5y ago
      The point is, ads engineering is about 10% of Google's engineering workforce. The 90% remaining are producing products like organic search, YouTube, maps, shopping, picasa, gmail, deepmind, Google brain, voice recognition that actually works, self driving, automatic translations and captions etc. Google produces the maximum amount of published research by far. How do you think these free products get paid for? By the 10% of engineering that does ads. The engineers optimizing ads are actually getting everything else in the company funded, including revenue less Deepmind, which would have folded years ago unless it was purchased by Big Tech. Sure give credit to the government and educational system, but what is the point of denigrating ads engineers who are funding all of these while also optimizing ads results so that they are relevant to users. Would you rather have them work on cancer research? Do you really think cancer research is desperate for machine learning engineers to help them out? Or would something like DeepMind, Google brain, automated medical report analysis, smart search of medical archives help them better? Edit: as an aside, its presumptuous to say that the best minds of our generation are ML and systems engineers. ML engineers are good at math, programming and optimizing models. This doesn't automatically make them great biologists, oncologists, physicists or even mathematicians.

echelon 5y ago

Outside of protein folding, each of these has been solved for 15+ years. And protein folding isn't exactly solved.

random314 5y ago
YouTube was solved 15 years ago?
- echelon 5y ago
  > February 14, 2005; 16 years ago https://en.wikipedia.org/wiki/YouTube https://en.wikipedia.org/wiki/YouTube It had a lot of content back then. Work since has gone into scaling and transcoding, but much more work has gone into ads, platform stickiness, music rights management, abuse detection, etc. Not really solving problems at the forefront of humanity.

goodpoint 5y ago

The following technologies have been researched between 1900 and 1990 almost entirely with *tax money*: Semiconductors, computing theory, GPS, GSM, satellites, airplanes, X-ray, optical cables, analog radio transmission and spread spectrum, industrial chemistry, solar power, nuclear power, vaccines, genetic research, and much more. Private companies putting together a search engine or uber or deliveroo are sitting on the shoulders of giants.

simion314 5y ago

>Would you like to solve the protein folding problem? Isn't Deep Mind a british company that Google bought it? Seems to me that most of the credit should go to UK society/education and just a bit to advertising industry.

random314 5y ago
What was the endgame for Deepmind, a company with no revenue stream? Geoff Hinton was also hired via acquisition.
- simion314 5y ago
  I do not know, but let's give the credit to the educational system and country of origin too , I see a lot of people give credit to US or SV for shit that is bought by the giants. Or I see arguments like "why there is any non US company for X ?" and the implication is that SV is the best where in fact SV has the money to buy anything that is promising and sometimes kill it but sometimes like in DeepMind case managed it good enough ot to waste it.
  - random314 5y ago
    The point is, ads engineering is about 10% of Google's engineering workforce. The 90% remaining are producing products like organic search, YouTube, maps, shopping, picasa, gmail, deepmind, Google brain, voice recognition that actually works, self driving, automatic translations and captions etc. Google produces the maximum amount of published research by far. How do you think these free products get paid for? By the 10% of engineering that does ads. The engineers optimizing ads are actually getting everything else in the company funded, including revenue less Deepmind, which would have folded years ago unless it was purchased by Big Tech. Sure give credit to the government and educational system, but what is the point of denigrating ads engineers who are funding all of these while also optimizing ads results so that they are relevant to users. Would you rather have them work on cancer research? Do you really think cancer research is desperate for machine learning engineers to help them out? Or would something like DeepMind, Google brain, automated medical report analysis, smart search of medical archives help them better? Edit: as an aside, its presumptuous to say that the best minds of our generation are ML and systems engineers. ML engineers are good at math, programming and optimizing models. This doesn't automatically make them great biologists, oncologists, physicists or even mathematicians.