Making Machine Learning Great Again | Jerry Kaplan | Talks at Google

[@00BillyTorontoBill]

schools should be more impressive than pro sports stadiums.

[@TimothyBates]

Almost everything said about IQ in the opening is false. it matters long after age 15, hasn't been a quotient of mental age/chronological age since Wechsler did away with this approach in 1939. ergo IQ doesn't mathematically plummet with increasing age. Intelligence remains the single best predictor of job performance, likelihood of getting a patent etc. (Schmidt). Critically, it's predictive strength does not end at age 15, nor is there a maximum value after which prediction weakens - it continues all the way up (Lubinski)

[@gabrielgossett9663]

I think you missed his point. He was pointing out IQ was designed as an educational tool, and that what most people generally mean when describing someone as intelligent doesn't necessarily relate to their literal IQ.

[@TimothyBates]

If he meant that, he should have said that. Instead he made multiple incorrect statements about IQ. If he meant to say IQ is an educational tool, he should have noted that, by design (Binet), it indexes ability to learn and solve novel problems, not current knowledge, and that it therefore predicts not only school (correlations around .7) but goes on predicting across the lifespan and for very high scores: both of which he denied. He would not have stated IQ scores are MA/CA, which hasn't been true for over 70 years. He would not then have made the absurd putdown of computer ability and IQ that if a program was 'born' 15 minutes ago it would have to score as brilliant. He set out to dismiss the idea that ability can be measured well, and he's just plain wrong. It's also not the case (regarding what you think his point was) that the common usage of "intelligent" is unrelated to IQ scores. They correlate very well. The main difference is the non-scientific usage is broader. It's a pity if googler's are learning baseless notions about intelligence, especially when they're working on AI…

[@gabrielgossett9663]

Timothy Bates Yea he seems to be referring to the original conception of IQ. He only brought up IQ to point out the cultural misunderstanding of machine intelligence as fundamentally different from human intelligences. I don't think his intentionally trying to spread misunderstanding about IQ as an objective measurement, he's just making a point about machine learning.

[@roganmorrow]

If that was that the point he was trying to make then he should have made it with more clarity. I can't take anything he says seriously anymore after listening to his erroneous ramblings about IQ. The derisive way in which he said "psychologists" was also cringey.

[@mnemonicman4161]

Although he makes good points, he presents them as answers, rather than possibilities. We *do* need to worry about whether machines should be aimed at being "intelligent", what responsibilities we give them, and whether we should try to have them replace our labour. These may not be the huge problems we fear based on science fiction, but these ideas are among the many that need thoughtful consideration in the real world.

[@susannec659]

you know enough to be Cliff Clavin on Cheers with "what everyone needs to know"

[@flinx]

When technology reduces an item's cost or time of production by an order of magnitude, new industries are created. For example waterwheel-powered fabric factories reduced the cost of fabric so much that the fashion industry grew tremendously. However when the drop in unit cost or production time isn't by an order of magnitude, new industries may not be created. For example before Ikea, most available furniture was expensive compared to today. Ikea's flatpack-self-assembly model added new, cheap, but decent furniture to the marketplace. Did it create a new industry? A limited one because worse-quality stuff breaks and needs replacing. But it also put many furniture makers and showrooms out of business. So what happens if transportation costs drop by an order of magnitude? Hopefully a new industry is created. What about burger-making, crop-picking, and cashiering robots? They exist today, but aren't an improvement by an order of magnitude. As development continues to improve them, they'll start replacing people's jobs. But it'll be like Ikea. Stuff will gradually get cheaper, while leaving huge numbers of people unemployed because a new industry wasn't created.

[@slightlygruff]

What kind of research is the Indian guy talking about at 31:36?

[@VPopkins]

Hi Doc

[@susannec659]

wait a minute wait a minute though the machines are going to learn they already do so they potentially could make decisions on their own someday

[@00BillyTorontoBill]

especially if they are made to pay taxes.

[@LuisManuelLealDias]

That title is waycist.