LLM is what usually sold as AI nowadays. Convential ML is boring and too normal, not as exciting as a thing that processes your words and gives some responses, almost as if it’s sentient. Nvidia couldn’t come to it’s current capitalization if we defaulted to useful models that can speed up technical process after some fine tuning by data scientists, like shaving off another 0.1% on Kaggle or IRL in a classification task. It usually causes big but still incremental changes. What is sold as AI and in what quality it fits into your original comment as a lifesaver is nothing short of reinvention of one’s workplace or completely replacing the worker. That’s hardly hapening anytime soon.
LLM is what usually sold as AI nowadays. Convential ML is boring and too normal, not as exciting as a thing that processes your words and gives some responses, almost as if it’s sentient.
To be fair, that’s because there are a lot of automation situations where having semantic understanding of a situation can be extremely helpful in guiding action over a ML model that is not semantically aware.
The reason that AI video generation and out painting is so good for instance it that it’s analyzing a picture and dividing it into human concepts using language and then using language to guide how those things can realistically move and change, and then applying actual image generation. Stuff like Waymo’s self driving systems aren’t being run through LLMs but they are machine learning models operating on extremely similar principles to build a semantic understanding of the driving world.
I’d argue, that it sometimes adds complexity to an already fragile system. Like when we implement touchscreens instead of buttons in cars. It’s akin to how Tesla, unlike Waymo, dropped LIDAR to depend on regular videoinputs alone. Direct control over systems without unreliable interfaces, semantic translation layer, computer vision dependancy etc serves the same tasks without additional risks and computational overheads.
I’d argue, that it sometimes adds complexity to an already fragile system.
You don’t have to argue that, I think thats inarguably true. But more complexity doesn’t inherently mean worse.
Automatic braking and collision avoidance systems in cars add complexity, but they also objectively make cars safer. Same with controls on the steering wheel, they add complexity because you now often have two places for things to be controlled and increasingly have to rely on drive by wire systems, but HOTAS interfaces (Hands On Throttle And Stick) help to keep you focused on the road and make the overall system of driving safer. While semantic modelling and control systems absolutely can make things less safe, if done well they can also actually let a robot or machine act in more human ways (like detecting that they’re injuring someone and stopping for instance).
Direct control over systems without unreliable interfaces, semantic translation layer, computer vision dependancy etc serves the same tasks without additional risks and computational overheads.
But in this case, Waymo is still having to do that. They’re still running their sensor data through incredibly complex machine learning models that are somewhat black boxes and producing semantic understandings of the world around it, and then act on those models of the world. The primary difference with Waymo and Tesla isn’t about complexity or direct control of systems, but that Tesla is relying on camera data which is significantly worse than the human eye / brain, whereas Waymo and everyone else is supplementing their limited camera data with sensors like Lidar and Sonar that can see in ways and situations humans can’t and that lets them compensate.
That and that Waymo is actually a serious engineering company that takes responsibility seriously, takes far fewer risks, and is far more thorough about failure analysis, redundancy, etc.
LLM is what usually sold as AI nowadays. Convential ML is boring and too normal, not as exciting as a thing that processes your words and gives some responses, almost as if it’s sentient. Nvidia couldn’t come to it’s current capitalization if we defaulted to useful models that can speed up technical process after some fine tuning by data scientists, like shaving off another 0.1% on Kaggle or IRL in a classification task. It usually causes big but still incremental changes. What is sold as AI and in what quality it fits into your original comment as a lifesaver is nothing short of reinvention of one’s workplace or completely replacing the worker. That’s hardly hapening anytime soon.
To be fair, that’s because there are a lot of automation situations where having semantic understanding of a situation can be extremely helpful in guiding action over a ML model that is not semantically aware.
The reason that AI video generation and out painting is so good for instance it that it’s analyzing a picture and dividing it into human concepts using language and then using language to guide how those things can realistically move and change, and then applying actual image generation. Stuff like Waymo’s self driving systems aren’t being run through LLMs but they are machine learning models operating on extremely similar principles to build a semantic understanding of the driving world.
I’d argue, that it sometimes adds complexity to an already fragile system. Like when we implement touchscreens instead of buttons in cars. It’s akin to how Tesla, unlike Waymo, dropped LIDAR to depend on regular videoinputs alone. Direct control over systems without unreliable interfaces, semantic translation layer, computer vision dependancy etc serves the same tasks without additional risks and computational overheads.
You don’t have to argue that, I think thats inarguably true. But more complexity doesn’t inherently mean worse.
Automatic braking and collision avoidance systems in cars add complexity, but they also objectively make cars safer. Same with controls on the steering wheel, they add complexity because you now often have two places for things to be controlled and increasingly have to rely on drive by wire systems, but HOTAS interfaces (Hands On Throttle And Stick) help to keep you focused on the road and make the overall system of driving safer. While semantic modelling and control systems absolutely can make things less safe, if done well they can also actually let a robot or machine act in more human ways (like detecting that they’re injuring someone and stopping for instance).
But in this case, Waymo is still having to do that. They’re still running their sensor data through incredibly complex machine learning models that are somewhat black boxes and producing semantic understandings of the world around it, and then act on those models of the world. The primary difference with Waymo and Tesla isn’t about complexity or direct control of systems, but that Tesla is relying on camera data which is significantly worse than the human eye / brain, whereas Waymo and everyone else is supplementing their limited camera data with sensors like Lidar and Sonar that can see in ways and situations humans can’t and that lets them compensate.
That and that Waymo is actually a serious engineering company that takes responsibility seriously, takes far fewer risks, and is far more thorough about failure analysis, redundancy, etc.