Stéphane ParentSelf Employed / Semi-retired| Leader MakerPrince Edward Island, Canada
Oct 01, 2019 2:19 PM
Replying to Wade Harshman
...
I think that depends on the scenario. Given a predictable format of data to process, the computer will win almost every time. (This includes not only processing, but input and output, which still takes people quite a bit of time.) But in more complex scenarios, or where strategy or creativity are required, AI cannot yet compete.
In the case of autonomous vehicles, they will (because we won't accept them if they don't) out-perform human drivers in normal driving situations, meaning they will stay in their lane, follow the speed limit, and turn when required. Under abnormal conditions, such as in an emergency requiring immediate evasive driving, they cannot be counted on, because we can't possibly program them for every unexpected scenario. People can't reliably make good decisions every time, either, but the point is that the average alert adult still has a better chance of reacting to an unexpected and complex scenario, perhaps even reacting before they've had a chance to think about the situation.
This is why Tesla's auto-pilot explicitly states that it's merely driver assistance, and the driver is still responsible for anything the car does or fails to do.
It's also why I don't yet fear that AI will replace project management.
In the case of an impending collision,my Corolla will brake my car, correctly balancing the needs of the human in front of the car and the needs of the car passengers. I doubt I could calculate the correct sweet spot in my head.
...
1 reply by Wade Harshman
Oct 03, 2019 8:08 AM
Wade Harshman
...
I think you underestimate yourself. Most of us can, in an emergency braking situation, consider what's in front of us, what's behind us, what's around us, and what/who is inside almost simultaneously. We typically do this without even realizing it. That doesn't mean we always make the best decisions (or that we're always alert to dangers), but the human brain is still a very powerful data processor.
Don't misunderstand me, I think this new technology is going to save lives, but I believe the advantage still goes to the human driver when you create more complex situations. What if the danger is not in front of you, but coming at you from the side? What if a potential hazard is actually a person, like another driver looking at his phone or expressing signs of road rage? Can an autonomous vehicle tell whether a child near the road is paying attention to traffic? Does it notice that a trailer has loose items that could fall out? Or that you have a team member who doesn't seem particularly interested in finishing his task on time and that could delay successor tasks and switch your critical path? These are things we pick up all the time, often without consciously thinking about it.
ERL/underground is fearless for me knowing the track’s making it stick to the ground (false sense of security i know!) but driverless cars? May be not (yet)! Saving Changes...
Wade HarshmanScrum Master| GDITIndianapolis, In, United States
Oct 02, 2019 2:07 PM
Replying to Stéphane Parent
...
In the case of an impending collision,my Corolla will brake my car, correctly balancing the needs of the human in front of the car and the needs of the car passengers. I doubt I could calculate the correct sweet spot in my head.
I think you underestimate yourself. Most of us can, in an emergency braking situation, consider what's in front of us, what's behind us, what's around us, and what/who is inside almost simultaneously. We typically do this without even realizing it. That doesn't mean we always make the best decisions (or that we're always alert to dangers), but the human brain is still a very powerful data processor.
Don't misunderstand me, I think this new technology is going to save lives, but I believe the advantage still goes to the human driver when you create more complex situations. What if the danger is not in front of you, but coming at you from the side? What if a potential hazard is actually a person, like another driver looking at his phone or expressing signs of road rage? Can an autonomous vehicle tell whether a child near the road is paying attention to traffic? Does it notice that a trailer has loose items that could fall out? Or that you have a team member who doesn't seem particularly interested in finishing his task on time and that could delay successor tasks and switch your critical path? These are things we pick up all the time, often without consciously thinking about it. Saving Changes...
Wade HarshmanScrum Master| GDITIndianapolis, In, United States
This is an old topic and an older article, but it fit so I wanted to share.
To sum up: the Society of Automotive Engineers have defined 5 levels of autonomous driving. BLUF is that we're still trying to achieve the middle levels of advanced driver assistance; fully autonomous vehicles are currently still the things of science fiction.
Level 1 is the basic driver assistance and safety technology common in many newer vehicles, like cruise control, auto brakes, and lane assistance.
Level 2 is the place we're trying to get to. It lets the car accelerate, brake, and steer, and the key is that the vehicle must be able to do multiple things at once. But the driver is still in charge. This is sort of like advanced driver assistance.
Level 3 places more emphasis on autonomous driving, but the human driver is still expected to over-ride the vehicle to avoid critical situations. This would be where our most advance research is focused, but it's not a near-term reality.
Level 4 is when an autonomous vehicle can continue without human assistance through complex and dangerous scenarios, such as construction zones and hazardous weather. The human driver rarely needs to intervene.
Level 5 is full autonomy. As the article states, this is the level where a car no longer needs a steering wheel.
To relate this back to project management, I suppose this comes down to managing expectations. One of my projects, for example, put a Q1 2020 release on our roadmap. This magically turned into release date of January 1st. That was poor expectation management. The same project has stakeholders with wildly different ideas of what MVP means, so we're still clarifying what will actually be in this new release.
...
1 reply by Luis Branco
Dec 16, 2019 9:42 AM
Luis Branco
...
Dear Wade
Interesting this definition of the 5 levels of autonomous driving.
Thank you for sharing
At what level will we be today?
In some car brands, "parking alone" is a common promise.
If so, we can quickly move to driverless vehicle levels
Saving Changes...
Luis BrancoCEO| Business Insight, Consultores de Gestão, LdªCarcavelos, Lisboa, Portugal
Dec 16, 2019 9:00 AM
Replying to Wade Harshman
...
This is an old topic and an older article, but it fit so I wanted to share.
To sum up: the Society of Automotive Engineers have defined 5 levels of autonomous driving. BLUF is that we're still trying to achieve the middle levels of advanced driver assistance; fully autonomous vehicles are currently still the things of science fiction.
Level 1 is the basic driver assistance and safety technology common in many newer vehicles, like cruise control, auto brakes, and lane assistance.
Level 2 is the place we're trying to get to. It lets the car accelerate, brake, and steer, and the key is that the vehicle must be able to do multiple things at once. But the driver is still in charge. This is sort of like advanced driver assistance.
Level 3 places more emphasis on autonomous driving, but the human driver is still expected to over-ride the vehicle to avoid critical situations. This would be where our most advance research is focused, but it's not a near-term reality.
Level 4 is when an autonomous vehicle can continue without human assistance through complex and dangerous scenarios, such as construction zones and hazardous weather. The human driver rarely needs to intervene.
Level 5 is full autonomy. As the article states, this is the level where a car no longer needs a steering wheel.
To relate this back to project management, I suppose this comes down to managing expectations. One of my projects, for example, put a Q1 2020 release on our roadmap. This magically turned into release date of January 1st. That was poor expectation management. The same project has stakeholders with wildly different ideas of what MVP means, so we're still clarifying what will actually be in this new release.
Dear Wade
Interesting this definition of the 5 levels of autonomous driving.
Thank you for sharing
At what level will we be today?
In some car brands, "parking alone" is a common promise.
If so, we can quickly move to driverless vehicle levels
...
1 reply by Wade Harshman
Dec 16, 2019 1:07 PM
Wade Harshman
...
According to SAE, auto-parking is Level 2. This is limited autonomy, when a vehicle can do some things on its own, under controlled circumstances. So there is still a wide gap until we reach fully autonomous vehicles (Level 5).
Saving Changes...
Wade HarshmanScrum Master| GDITIndianapolis, In, United States
Dec 16, 2019 9:42 AM
Replying to Luis Branco
...
Dear Wade
Interesting this definition of the 5 levels of autonomous driving.
Thank you for sharing
At what level will we be today?
In some car brands, "parking alone" is a common promise.
If so, we can quickly move to driverless vehicle levels
According to SAE, auto-parking is Level 2. This is limited autonomy, when a vehicle can do some things on its own, under controlled circumstances. So there is still a wide gap until we reach fully autonomous vehicles (Level 5). Saving Changes...
