You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Deep learning excels at unlocking the creation of impressive early demos of new applications using very little development resources. The part where it struggles is reaching the level of consistent usefulness and reliability required by production usage.
Autonomous driving is the ultimate example. You could use deep learning to create an impressive self-driving car prototype in 2015 on a shoestring budget (Comma did exactly that, using Keras). Five years and billions of $ later, the best DL-centric driving systems are still L2+.
Every app demo based on GPT-3 follows this pattern. You can build the demo in a weekend, but if you invest $20M and 3 years fleshing out the app, it's unlikely it will still be using GPT-3 at all, and it may ever meet customer requirements
The reason why is that parametric models trained with gradient descent make it easy to automate something, but have little ability to deviate from the patterns they've learned. Meanwhile, the real world is full of surprises, and handling it requires the ability to adapt.
"Information Theory" is an example of a "complex system" (math terminology) which attempt to enable work on system variables variables that influence each other in non-obvious ways.
alternatively we just don't test and validate enough early on and assume linear complexity in continuing the solution when the reality is the complexity is exponential and performance may be subject to a glass ceiling.
It does not need to be perfect, just better than humans. Example: AI is now better than humans in diagnosing skin cancer. Not perfect, but better. And this is what matters.
The declaration of Tesla FSD beta as L2 has legal reasons, is independent of its capabilities. Declaring it L3 or L4 just shifts liability, so it will be decided in a talk with insurance, not with regulators. But that will not happen as long as it is declared beta.
Seems like an echo of an earlier AI era, when people would get into conversations with Eliza and such. Or even graphics, where the definition of "realistic" keeps changing. The gulf between good demo and long term *reliable* performance is much bigger than nothing to demo.
George Hotz said he specifically and purposefully wants to remain L2 forever to avoid liabilities. I think Tesla FSD is doing the same thing. This is a problem separate from development, no?
sounds more like Lionel Hutz
What are your thoughts / prediction around companies achieving full seek driving in 5 years?
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
