Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Using sequences of life-events to predict human lives

arxiv.org

Post reply on HN
Print view View original

21–30 of 56 posts

Re: Using sequences of life-events to predict human lives

#21

Post by garrisonhh »

zmgsabst wrote: post #16
wharvle wrote: post #9

Earlier quoted context omitted.

Every time we get better at things like this, all the power tends to stay with organizations, and individual people are lucky if they get anything out of it. Amazon, Wal-Mart, MacDonalds, giant banks, insurers, et c., analyze their competition, suppliers, and consumers in great detail with computer data and models. Where’s the widespread automated buyer’s agent for consumers, to balance that power so strongly that it…

Scraping is legal; as are assistive technologies. What’s lacking is the collective willpower to develop social organizations, which could orchestrate and fund such efforts. There’s a complete lack of organization among the public. The problem is neither technical nor legislative: people just don’t care to, and are happier accepting the status quo.

People don't have the time, money, or energy. And they will continue to lose all three until there is no option but to figure out collective organisation

Re: Using sequences of life-events to predict human lives

#22

Post by derbOac »

anigbrowl wrote: post #4

I have mixed feelings about this. It could be a great tool to help people, but on past experience it seems more likely that HR departments/ financial service providers/ governments would use it to write people off if they have too many negative life indicators.

This is a critical problem of our time in my opinion. Persons are approached as static unchangeable objects, as if their trajectories are deterministic and what you're doing with these models are resolving the trajectories like a telescope or microscope. Measuring the substance of the person so you can extract from them.

The other schema is one where you're modeling manipulation effectiveness, to probe and predict how you can best (most effectively with the least resources) achieve some change in the trajectory. Engineering a person. It seems like we really only think of interventions as trying to return someone to some baseline condition, rather than to try to improve people beyond the baseline, to invest in them.

The paper has a bit at the end about using it for personalized interventions but it feels performative, like lip service. I know one of the authors professionally in a distant sense so I think they're probably genuine about it, but I can't help but wonder how the paper might be different if intervention was actually their focus. Wouldn't they be modeling changes in trajectories, the deviations or something?

Re: Using sequences of life-events to predict human lives

#23

Post by walterbell »

reality_inspctr wrote: post #5

I tried to build a startup based on a similar idea, but the traction was not good. The squarespace website is still up https://sundialcalendar.com/ It’s stunning how much of our lives are predictable. We are free will surfers on waves of determinism.

> We are free will surfers on waves of determinism

As the cost of real-time surveillance, prediction and "nudging" continue to fall, the cumulative feedback loop can converge into human archetypes which are as much a function of the observer as the observed.

In an ocean of sensors and stimuli, how can past free will influence our future selves? Any stories you can share from users of your service?

Re: Using sequences of life-events to predict human lives

#25

Post by otikik »

wharvle wrote: post #9
anigbrowl wrote: post #4

I have mixed feelings about this. It could be a great tool to help people, but on past experience it seems more likely that HR departments/ financial service providers/ governments would use it to write people off if they have too many negative life indicators.

Every time we get better at things like this, all the power tends to stay with organizations, and individual people are lucky if they get anything out of it. Amazon, Wal-Mart, MacDonalds, giant banks, insurers, et c., analyze their competition, suppliers, and consumers in great detail with computer data and models. Where’s the widespread automated buyer’s agent for consumers, to balance that power so strongly that it…

Seize the means of computation

Re: Using sequences of life-events to predict human lives

#26

Post by nehal3m »

garrisonhh wrote: post #21
zmgsabst wrote: post #16

Earlier quoted context omitted.

Scraping is legal; as are assistive technologies. What’s lacking is the collective willpower to develop social organizations, which could orchestrate and fund such efforts. There’s a complete lack of organization among the public. The problem is neither technical nor legislative: people just don’t care to, and are happier accepting the status quo.

People don't have the time, money, or energy. And they will continue to lose all three until there is no option but to figure out collective organisation

It definitely feels deliberate and yeah it is a feedback loop.

Re: Using sequences of life-events to predict human lives

#27

Post by SCUSKU »

Part of me wants to be able to encode my life2vec and see what the next predicted tokens are. On the other hand, as much as going to a fortune teller sounds fun to me, I worry that if I go I will simply make what they say come true since I will be anxious about what the fortune teller said.

That said, my life is already on a certain track, and because I'm afraid of change, it's unlikely to be much different than it was before. Maybe this is just a weird call to actually do what your gut says, and to follow your dreams.

Re: Using sequences of life-events to predict human lives

#28

Post by dkjaudyeqooe »

SCUSKU wrote: post #27

Part of me wants to be able to encode my life2vec and see what the next predicted tokens are. On the other hand, as much as going to a fortune teller sounds fun to me, I worry that if I go I will simply make what they say come true since I will be anxious about what the fortune teller said. That said, my life is already on a certain track, and because I'm afraid of change, it's unlikely to be much different than it w…

> because I'm afraid of change, it's unlikely to be much different than it was before

Life changing events are often out of your control

Re: Using sequences of life-events to predict human lives

#30

Post by jedberg »

shadowgovt wrote: post #20
MattDaEskimo wrote: post #8

Earlier quoted context omitted.

I'd like to believe that lives are way too nuanced to properly train and evaluate a transformer this way. Also, I would wager that upon release, suddenly everybody says they eat healthy and exercise daily.

I have an acquaintance who is convinced that Google ads stalk him because the ads make inferences yet shouldn't be able to, like offering flight prices from one city to another he frequents when he's never done that search in Google before. I've tried to explain to him that people aren't actually very unique in large models, and there's probably several signals Google does have from his other searches to conclude he'…

Does he use Gmail? Because it gets all of his flight confirmations and Calendar events. Does he use Google Maps, because it gets his realtime location whenever he uses it. If he frequently goes to those cities, they would certainly know if he uses any other Google service.
Post reply on HN