Might be a very opinionated statement to say, however there clearly seems to be a very clear objective mismatch in the AI community as to what it wants to achieve. The primary goal of AI was to figure out how human mind works and that doesn't seem to be the case currently? I think the pioneers had a lot more clarity in what they wanted this field to achieve. And the present one doesn't have a clue on what it wants, and where it wants to go? Sorry, about the rant. Here's my TC: 20L, INR
The short term goal at the moment is clear: use data to automate stuff to improve productivity. There are teams at deep mind etc that think about longer term goals.
Agreed
I don't think that in theory this would work, the primary goal of google as an organisation doesn't seem to match the goal of the AI field itself.
Who are the present ML godfathers according to you? All the pioneers like Hinton, Norvig and Godfellow, LaCunn has already accepted that the current Deep Learning methods are insufficient for AGI and strongly suggest moving to or at least utilise symbolic processing. There is a vast academic interest in this area and they are all publishing papers demonstrating innovative ways to do AI. Don’t follow wannabes and form an opinion of. Industry is very different from academics.
That's a very great point. Industry is very different from academics! And here is where the problem is, the industry doesn't care about the goal of AI as a field. The industry only cares about short-term returns. And figuring out how the mind works (won't yield returns atleast immediately)
Might be very crude to say, but calling Hinton, Norvig and Godfellow, LaCunn as pioneers is just too early. I understand they did have good amount of contribution to the subject, but their contribution to the vision of the field is not significant enough.
The short term goal is just to use it as a marketing tactic. Every company is like just because everyone is doing, let us also do it, without bothering whether it actually adds value
Coming up next: quantum
That’s the whole point of AI. So they can figure out their own goals. U fool!
I recommend watching Hamming's lectures on AI on YouTube. The first: https://youtu.be/aq_PLEQ9YzI
This is exactly what they want: all they can do is CV and NLP, but implies to VC they can create intelligence as long as you can imagine
Another issue is not many people even after formal training actually know what artificial intelligence and just like bluffing around and using terms like ai , deep learning, data science , chatbot in their slide decks . The situation worsens in companies like cisco where certain people have a great amount of difficulty in writing SQL queries and using easy tools like Tableau , DOMO . but even these people will be so called ai driven data scientist
I feel attacked. I haven't really used SQL in 15 years, and never tried tableau or DOMO. I don't like to call what I do AI, but there's plenty of ML, a bit of deep learning, and yes "data science".
Sincere apologies . I was trying to give a comparison. Sorry
People who do real AI work are currently heads down making things happen. The field is very new and some very smart minds are on it. Some dumb minds are also on it but they are really not passionate about it. The smart ones are focused on making things happen rather than making the news. The ones who make the news are the ones who want some fame in the name of AI. Real innovation is always low profile.
The smart people u refer to are engineers and such. I don't think they have enough philosophers on board. And that's who we need to make it happen the "right" way. Otherwise our tech will race ahead while the heart stays behind.
Perhaps engineers only need to take on the role of a philosopher in this case. We won't magically get one suited for this role otherwise ;)
How do you define AI?
Already did, the primary goal was to figure out how human mind works.
This is more like neuroscience