Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
How many quarters after Q1 2016 till Marissa Mayer is unemployed?
1 or fewer
2
4
Her job is guaranteed; what are you even talking about?
View Results
 
  • Post
  • Reply
Kwyndig
Sep 23, 2006

Heeeeeey


Oh man, and they just can't accept that several percent loss from deliberate misscanning and accidental walk offs can they.

Adbot
ADBOT LOVES YOU

Mister Facetious
Apr 21, 2007

I think I died and woke up in L.A.,
I don't know how I wound up in this place...

:canada:
It's a zero sum game to them; if they aren't taking your money, you're taking theirs.

Leon Trotsky 2012
Aug 27, 2009

YOU CAN TRUST ME!*


*Israeli Government-affiliated poster

Mister Facetious posted:

It's a zero sum game to them; if they aren't taking your money, you're taking theirs.

Isn't a situation where you are stealing something a literal zero sum transaction?

Mister Facetious
Apr 21, 2007

I think I died and woke up in L.A.,
I don't know how I wound up in this place...

:canada:

Leon Trotsky 2012 posted:

Isn't a situation where you are stealing something a literal zero sum transaction?

That's what I'm saying; ANY loss is unacceptable, no matter how inevitable.

No expense will be spared to stop theft!
Spoilage? Enh, not so much.

Magic Hate Ball
May 6, 2007

ha ha ha!
you've already paid for this

Antigravitas posted:

Inserted AI-generated Microsoft poll about woman’s death rankles The Guardian
Speculative AI news poll presented three choices: Murder, accident, or suicide.

:eyepop:

A screencap of the poll, which is bleakly hilarious:

Mister Facetious
Apr 21, 2007

I think I died and woke up in L.A.,
I don't know how I wound up in this place...

:canada:
Anyone remember which doordash/uber eats/etc competitor business was creating restaurant pages without their owners consent?

withak
Jan 15, 2003


Fun Shoe
I thought google was?

Nenonen
Oct 22, 2009

Mulla on aina kolkyt donaa taskussa
I'm surprised that the AI hasn't learned already that the choices need to be: Soros, Hillary, lizards

Mister Facetious
Apr 21, 2007

I think I died and woke up in L.A.,
I don't know how I wound up in this place...

:canada:

withak posted:

I thought google was?

My friend says grubhub, but thanks to your post i did find a lawsuit against google too :cheers:

shoeberto
Jun 13, 2020

which way to the MACHINES?

Leon Trotsky 2012 posted:

My local Wal-Mart has now started staffing every individual self-checkout machine with an employee who scans all of your items for you.

Thus, the technological changes in checkout have come full circle and now the store has to staff every register AND it takes much longer to checkout that before.

On a related note, I saw a commercial for this last night and had a lol: https://www.spectrum.com/cable-tv/streaming/xumo

Sort of feels full circle, they know that no one wants terrestrial cable service with a proprietary set-top box, but if you can still sell it if you call it streaming (still with a proprietary set-top box)

Baronash
Feb 29, 2012

So what do you want to be called?

Professor Beetus posted:

Wouldn't the reliability of AI cars on low traction/ice surfaces be something where, ideally, there would be sensors keeping track of the car's traction and grip to adjust speed and movement accordingly? In other words the car wouldn't need to know that ice was on the road, because it's detecting the loss of traction and not the ice itself.

Unless that's crazy sci-fi tech I thought we already had systems like that to assist with braking and stability, which could conceivably be iterated on to improve performance and sensitivity.

I mean I'm certainly not sold on the feasibility but it seems silly to say "well they can't know about where they saw a car drive into a ditch once" when the car should have much more concrete and reliable information available to assist with those conditions.

I agree, and I find it kind of weird that discussion has trended towards whether self-driving vehicles can learn the uncomputable voodoo of "stopping distance is increased in icy conditions." The real issue presented by snowy and icy conditions is actually far larger than how these vehicles handle skidding. At present, even mild snowfall impedes the ability of these vehicles to understand where they are on the roadway by obscuring lane lines and increasing noise for other sensors. It's a significant problem, and it so thoroughly confounds the sensor packages (typically a combination of Lidar, radar, and cameras) used in most autonomous vehicles on the road that some researchers have given up on incremental software improvements in favor of mapping the composition of every square foot of drivable space and giving autonomous vehicles a ground-penetrating radar system to compare it against.
It's pretty illustrative of how far away these systems are from general adoption that even as the technology improves, both researchers and the companies testing these vehicles on open roads are leaning harder and harder on navigation systems that can only function in areas that have been manually mapped at great effort and expense.

To bring it back around, what I do think we'll continue to see is big automobile manufacturers mining these self-driving companies for a handful of driver assist features every few years. I think you're absolutely right that a system with access to reliable sensor data would probably do a much better job of slide recovery than I as a human would, and there's every reason to expect that some sort of "advanced ABS" is going to join features like adaptive cruise control and lane keeping at some point.

Baronash fucked around with this message at 22:08 on Nov 1, 2023

Professor Beetus
Apr 12, 2007

They can fight us
But they'll never Beetus

Nenonen posted:

If you are going to a curve @100km/h and only at that point realise that it might be slippery, what is it good for?

Well since a human being still has to be inside the car, one would imagine that there could be different options for various weather conditions that you could activate, in theory. I'm going to point out again though that I'm simply imagining what might be possible, as a thought exercise, because some of the posts sound like "well look I know the weather balloons are very good at determining weather but there's nothing that works quite as well as the bursitis in my knee telling me the weather is changing." What isn't possible now could very well be in the future and it probably should be researched and developed (in a responsible and well-regulated way, not what's currently happening).

Dirk the Average posted:

These systems already exist without AI. Not everything needs to be AI.

I never said they did, but it was what was being discussed. And machine learning does seem like a promising avenue for improving those systems.



Baronash posted:

I agree, and I find it kind of weird that discussion has trended towards whether self-driving vehicles can learn the uncomputable voodoo of "stopping distance is increased in icy conditions." The real issue presented by snowy and icy conditions is actually far larger than how these vehicles handle skidding. At present, even mild snowfall impedes the ability of these vehicles to understand where they are on the roadway by obscuring lane lines and increasing noise for other sensors. It's a significant problem, and it so thoroughly confounds the sensor packages (typically a combination of Lidar, radar, and cameras) used in most autonomous vehicles on the road that some researchers have given up on incremental software improvements in favor of mapping the composition of every square foot of drivable space and giving autonomous vehicles a ground-penetrating radar system to compare it against.
It's pretty illustrative of how far away these systems are from general adoption that even as the technology improves, both researchers and the companies testing these vehicles on open roads are leaning harder and harder on navigation systems that can only function in areas that have been manually mapped at great effort and expense.

To bring it back around, what I do think we'll continue to see is big automobile manufacturers mining these self-driving companies for a handful of driver assist features every few years. I think you're absolutely right that a system with access to reliable sensor data would probably do a much better job of slide recovery than I as a human would, and there's every reason to expect that some sort of "advanced ABS" is going to join features like adaptive cruise control and lane keeping at some point.

Yeah this is a great post and that stuff about mapping every inch of drivable surface is wild. And I think your last paragraph is spot on, even without fully automated luxury gay self-driving cars, there are a lot of potential safety improvements that could be developed out of the research into it. I think just because Tesla is led by a collosal moron who keeps cutting corners in ways that should have resulted in losing the right to sell vehicles, isn't a good enough reason to just throw up hands and say "welp this isn't possible"

Nenonen
Oct 22, 2009

Mulla on aina kolkyt donaa taskussa

Professor Beetus posted:

Well since a human being still has to be inside the car, one would imagine that there could be different options for various weather conditions that you could activate, in theory. I'm going to point out again though that I'm simply imagining what might be possible, as a thought exercise, because some of the posts sound like "well look I know the weather balloons are very good at determining weather but there's nothing that works quite as well as the bursitis in my knee telling me the weather is changing." What isn't possible now could very well be in the future and it probably should be researched and developed (in a responsible and well-regulated way, not what's currently happening).

Okay. So tell me this: if an AI model thinks that it's totally safe to drive into a curve @100km/h and there's ice and people die, who's responsible?

Because we know who is responsible when a human is driving: the human driving. If the AI model can't be trained to be safe enough to avoid all hazards then who should bear the responsibility for the inevitable accidents?

Arsenic Lupin
Apr 12, 2012

This particularly rapid💨 unintelligible 😖patter💁 isn't generally heard🧏‍♂️, and if it is🤔, it doesn't matter💁.


Oxyclean posted:

I'm curious how well AIs could be set up to handle stuff like black ice or other loss-of-control scenarios. Humans don't handle these well, but I feel like you have scenarios where humans *can* make on the fly decision making that would be extraordinarily hard to design an AI to do?
A crucial part of driving in snow is predicting which cars around you are likeliest to spin out and having an exit path planned, if necessary. Similarly, "poo poo, I'm being tailgated, better give as much warning while braking as possible". It's not just the ice and snow. It's being aware of the mistakes other drivers seem to be likely to make. To do that, you have to have some (not all) theory of mind.

Professor Beetus
Apr 12, 2007

They can fight us
But they'll never Beetus

Nenonen posted:

Okay. So tell me this: if an AI model thinks that it's totally safe to drive into a curve @100km/h and there's ice and people die, who's responsible?

Because we know who is responsible when a human is driving: the human driving. If the AI model can't be trained to be safe enough to avoid all hazards then who should bear the responsibility for the inevitable accidents?

I mean if there's still some measure of human control, like my example of having different driving "modes" based on current weather conditions, then it should be possible to see if the driver was using a mode that could be considered "safe" for current traffic conditions. If it can be determined otherwise, you could look at assigning liability to the car manufacturer as a mechanical failure, just like any other situation where a mechanical failure is at fault. That's obviously a very broad hypothetical because the technology isn't there yet, but it's the most likely thing I am capable of imagining right now.

I think it's a worthwhile endeavor and if it does turn out to be capable of reducing traffic fatalities by even ten percent, then going through the mess of trying to navigate the legality and liability issues would be worth it. I would rather the US get over its car obsession and push for getting rid of them in cities and massively improving public transit, but in the US that's about as possible as perfecting a self driving car, and less likely to be pursued.

Obviously this is a very US-centric view but I do think that's fairly warranted given the particular dependence on cars that the US has.

PT6A
Jan 5, 2006

Public school teachers are callous dictators who won't lift a finger to stop children from peeing in my plane

Arsenic Lupin posted:

A crucial part of driving in snow is predicting which cars around you are likeliest to spin out and having an exit path planned, if necessary. Similarly, "poo poo, I'm being tailgated, better give as much warning while braking as possible". It's not just the ice and snow. It's being aware of the mistakes other drivers seem to be likely to make. To do that, you have to have some (not all) theory of mind.

And also, knowing when people are going to bend or break the rules. If you're on a three-lane road that's not echelon-plowed, meaning there's snow and slush between all the lanes because the lanes were plowed one at a time, do you follow the lines or do you follow the plowed area/tire tracks? A human knows that you tend to follow the clear areas, an AI doesn't have the same ability to reason about when the lane markers start to matter again.

Ultimately, I think safety is most improved by advanced driver aids without moving to self-driving, because the reasoning of a human being combined with the technical expertise of a good computer system leverages what both components do well. That means I don't get to crack a beer and jack off during my drive home, but it can wait.

Platystemon
Feb 13, 2012

BREADS
I would simply denude the environment in front of the vehicle of any traction‐impeding precipitation.

Mister Facetious
Apr 21, 2007

I think I died and woke up in L.A.,
I don't know how I wound up in this place...

:canada:

Platystemon posted:

I would simply denude the environment in front of the vehicle of any traction‐impeding precipitation.



You're joking, but what about States and municipalities where taxes for roadworks in general and winter operations specifically are considered a valid target for reduction

notwithoutmyanus
Mar 17, 2009

Nervous posted:

:siren:NON TECHBRO DETECTED:siren: Do not be alarmed. A gift grift has been dispatched to the origin point of this post. Please do not move until it arrives.

Neito
Feb 18, 2009

😌Finally, an avatar the describes my love of tech❤️‍💻, my love of anime💖🎎, and why I'll never see a real girl 🙆‍♀️naked😭.

Mister Facetious posted:

Anyone remember which doordash/uber eats/etc competitor business was creating restaurant pages without their owners consent?

It was DoorDash.

Evil Fluffy
Jul 13, 2009

Scholars are some of the most pompous and pedantic people I've ever had the joy of meeting.

Magic Hate Ball posted:

A screencap of the poll, which is bleakly hilarious:



This viral marketing for Silent Hill: Ascension is wild.

Original_Z
Jun 14, 2005
Z so good
https://www.nytimes.com/live/2023/11/02/business/sam-bankman-fried-trial

SBF found guilty on all charges!

Jose Valasquez
Apr 8, 2005


That's not a nightmare at all

Arsenic Lupin
Apr 12, 2012

This particularly rapid💨 unintelligible 😖patter💁 isn't generally heard🧏‍♂️, and if it is🤔, it doesn't matter💁.


PT6A posted:

And also, knowing when people are going to bend or break the rules. If you're on a three-lane road that's not echelon-plowed, meaning there's snow and slush between all the lanes because the lanes were plowed one at a time, do you follow the lines or do you follow the plowed area/tire tracks? A human knows that you tend to follow the clear areas, an AI doesn't have the same ability to reason about when the lane markers start to matter again.

Ultimately, I think safety is most improved by advanced driver aids without moving to self-driving, because the reasoning of a human being combined with the technical expertise of a good computer system leverages what both components do well. That means I don't get to crack a beer and jack off during my drive home, but it can wait.
I did not know the phrase "echelon-plowed". Neat!

You're quite right about following the clear areas. One of the reasons is that you don't know what the first car was dodging around, and you're safer on the track that lots of other people survived :)

Oxyclean
Sep 23, 2007


Jose Valasquez posted:

That's not a nightmare at all
the nightmare is the dozens of other crypto and financial sector people who get away with bullshit everyday with little to no consequence

Vegetable
Oct 22, 2010

His co-conspirators all got plea deals for ratting him out. They were just as bad imo. A lot of injustice is done to get a bit of justice.

Jose Valasquez
Apr 8, 2005

Vegetable posted:

His co-conspirators all got plea deals for ratting him out. They were just as bad imo. A lot of injustice is done to get a bit of justice.

They're still going away for most of their lives. Caroline Ellison wasn't testifying to avoid jail she was testifying to maybe avoid dying in jail

Papercut
Aug 24, 2005
Yeah we still don't know the sentences of anyone involved so it's premature to be declaring injustice for the plea deals

Mega Comrade
Apr 22, 2004

Listen buddy, we all got problems!

Vegetable posted:

His co-conspirators all got plea deals for ratting him out. They were just as bad imo. A lot of injustice is done to get a bit of justice.

He could have plead guilty....

Leon Trotsky 2012
Aug 27, 2009

YOU CAN TRUST ME!*


*Israeli Government-affiliated poster
Caroline Ellison was facing a 110-year sentence.

She could get a max of around 20 years with the plea deal. She definitely won't get the max, but we don't know how much she will get.

There's no parole for newly convicted people in federal prison, so they will have to serve whatever sentence they get.

Potentially getting 20 years for a first-time financial crime is still not a great situation to be in.

SBF is basically dying in prison for bitcoin.

F4rt5
May 20, 2006

PT6A posted:

Ultimately, I think safety is most improved by advanced driver aids without moving to self-driving, because the reasoning of a human being combined with the technical expertise of a good computer system leverages what both components do well. That means I don't get to crack a beer and jack off during my drive home, but it can wait.
Well… the past few years there has been chaos on the roads every first snowfall. Much because people are idiots and forget this happens every year and try to drive with summer tires (which is illegal in icy or snowy conditions) but even those who changed tires in time are sliding out like amateurs. I think this is because of assists like modern traction control and stability stuff, making the surface seem safer than it is, masking the real conditions; if you lose grip because you’re on the limit of what those can do, the car is way beyond what you as a human can save and you have a bad time.

Foxfire_
Nov 8, 2010

Leon Trotsky 2012 posted:

There's no parole for newly convicted people in federal prison, so they will have to serve whatever sentence they get.
Did something change recently?

In the past, you could get about 1/7th of a federal sentence turned into supervised release if you max out your good conduct credits every year (54 days/year). Much less early release than most state convictions, but not nothing

Leon Trotsky 2012
Aug 27, 2009

YOU CAN TRUST ME!*


*Israeli Government-affiliated poster

Foxfire_ posted:

Did something change recently?

In the past, you could get about 1/7th of a federal sentence turned into supervised release if you max out your good conduct credits every year (54 days/year). Much less early release than most state convictions, but not nothing

They eliminated parole for federal crimes under Reagan. You can still get good conduct credits, but you have a fixed minimum of your sentence you have to serve.

You used to be able to get parole after 25% of your sentence was completed, but that hasn't been the case for a very long time.

Blut
Sep 11, 2009

if someone is in the bottom 10%~ of a guillotine

Leon Trotsky 2012 posted:

Caroline Ellison was facing a 110-year sentence.

She could get a max of around 20 years with the plea deal. She definitely won't get the max, but we don't know how much she will get.

There's no parole for newly convicted people in federal prison, so they will have to serve whatever sentence they get.

Potentially getting 20 years for a first-time financial crime is still not a great situation to be in.

SBF is basically dying in prison for bitcoin.

gently caress with rich people's money and you do serious jail time in the US. The courts really, really care about protecting rich people it turns out.

SBF getting 115 years when the median time served for literally being convicted of murdering someone in the US is 13.4 years says a lot.

Mister Facetious
Apr 21, 2007

I think I died and woke up in L.A.,
I don't know how I wound up in this place...

:canada:

Blut posted:

gently caress with rich people's money and you do serious jail time in the US. The courts really, really care about protecting rich people it turns out.

SBF getting 115 years when the median time served for literally being convicted of murdering someone in the US is 13.4 years says a lot.

Sbf got five consecutive sentences or something, so it's more the equivalent of a serial killer sentence.

OddObserver
Apr 3, 2009
I don't think he's been sentenced yet, just found guilty? And good chance the guidelines call for less than the maximum.

Volmarias
Dec 31, 2002

EMAIL... THE INTERNET... SEARCH ENGINES...

Leon Trotsky 2012 posted:

Caroline Ellison was facing a 110-year sentence.

She could get a max of around 20 years with the plea deal. She definitely won't get the max, but we don't know how much she will get.

There's no parole for newly convicted people in federal prison, so they will have to serve whatever sentence they get.

Potentially getting 20 years for a first-time financial crime is still not a great situation to be in.

SBF is basically dying in prison for bitcoin.

Tech Nightmares 6: Dying in prison for Bitcoin

notwithoutmyanus
Mar 17, 2009

OddObserver posted:

I don't think he's been sentenced yet, just found guilty? And good chance the guidelines call for less than the maximum.

This is the first trial, though. Nobody gets the maximums. He still has a second for different fraud, but they could still theoretically sentence him to 50 or 50+ years just on this one. We'll know in March

Mister Facetious
Apr 21, 2007

I think I died and woke up in L.A.,
I don't know how I wound up in this place...

:canada:

Volmarias posted:

Tech Nightmares 6: Dying in prison for Bitcoin

I like it

Adbot
ADBOT LOVES YOU

Sign
Jul 18, 2003

Volmarias posted:

Tech Nightmares 6: Dying in prison for Bitcoin

Missing the obvious
Tech Nightmares 6: Don't drop the Soapcoin

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply