Skip to main content

What AI labs should be doing now

 By now I guess it is clear that Deep Learning isn't the miracle solution that will finally bring us our dream of Artificial General Intelligence. Now we can leave the hype alone and focus on the reality of the situation, Deep Learning is a sophisticated tool that can help us solve many problems we can't solve through traditional software development.

We should continue its development with the aim of continuously reducing its computational footprint so that it is more accessible to the common programmer without multimillion-dollar resources. But we must stop that unproductive fantasy that somehow deep learning is that path to the biggest dream of humanity, which is Artificial General Intelligence (AGI).

Every AI Lab around the world should be devoting nothing less than 30% of its time and resources to the development of AGI. It is the most important goal humanity can ever achieve because it will be a goal of goals, enabling us to solve every other problem we are currently cracking our brains about. 

It will help us solve physics that will enable us to achieve interstellar travel. It will help us solve biology and enable us to live eternal healthy lives. It will literally give us solutions to the most confounding questions of our existence and enable us to become a truly universal species. 

So why devote only 30% of time and resources and not more, well we have to understand that it is a very risky pursuit with the possibility that we may never solve it within this century because no one knows for sure how to go about solving it. But it would be a disservice to humanity to devote fewer resources to its pursuit because of the possibility that we may solve it in the next year or in a decades time. It is as risky as powered flight and maybe we do not yet have the critical base technology or the math to even make reasonable attempts at the problem yet, but it will be a shame to not try because the payout will be enormous if we succeed. 

Many AI labs have abandoned many of the old research directions that AI took in the past, everybody is devoting too many resources to squeezing one last drop of juice from deep learning they are ignoring so many promising old school research direction.

It should be of note that current deep learning techniques were actually proposed in the 70s and it is only now that we have the computing ability to even try it, which we did with great success as you can see from all the achievements of deep learning from speech recognition to self-driving cars. The question I often ask is that what if there is some deep-buried research that could give us some headway towards AGI, but it is buried in some obscure journal from the 1950s?

On the one hand, it might seem like we should go back to the past and try out every idea ever published but this would be impractical. This is where the intuition comes to play, researchers should go to the past and pore over a lot of ideas and allow their intuition to guide them to what they find promising and they should be given appropriate resources to pursue the direction they have chosen exhaustively, who knows what we could discover. 

It is very important that we strive to build AGI no matter what our definition of the term means to us. We should try like the Wright Brothers, the future of humanity depends on it. 

Comments

Popular posts from this blog

New Information interfaces, the true promise of chatGPT, Bing, Bard, etc.

LLMs like chatGPT are the latest coolest innovation in town. Many people are even speculating with high confidence that these new tools are already Generally intelligent. Well, as with every new hype from self-driving cars based on deeplearning to the current LLMs are AGI, we often tend to miss the importance of these new technologies because we are often engulfed in too much hype which gets investors hyper interested and then lose interest in the whole field of AI when the promises do not pan out. The most important thing about chatGPT and co is that they are showing us a new way to access information. Most of the time we are not interested in going too deep into typical list-based search engine results to get answers and with the abuse of search results using SEO optimizations and the general trend towards too many ads, finding answers online has become a laborious task.  Why people are gobbling up stuff like chatGPT is not really about AGI, but it is about a new and novel way to...

Next Steps Towards Strong Artificial Intelligence

What is Intelligence? Pathways to Synthetic Intelligence If you follow current AI Research then it will be apparent to you that AI research, the deep learning type has stalled! This does not mean that new areas of application for existing techniques are not appearing but that the fundamentals have been solved and things have become pretty standardized.

Human intelligence is more about knowledge utilization

There are two significant phases of intelligence, the first is the knowledge acquisition phase and the second is the knowledge utilisation phase. The knowledge acquisition stage is available by default in many animals but what makes humans unique is that our ability to utilize knowledge is far higher than most mammals.  But even humans individually and as a whole utilize only a very tiny fraction of the available knowledge. For an individual your acquired knowledge base, that is knowledge you have acquired both consciously through study or unconsciously through just existing, is far more than what you are able to utilize.  This is also similar for humanity as a whole, the knowledge we possess both in written form or in other forms far outweighs our ability to utilize it.  The core part of what we could call general intelligence is really more about knowledge utilization and less about acquisition. Because what is useful is only what we can utilize, the rest is just memory...