I think there is little time left before someone builds AGI (median 2030). Once upon a time, I didn't think this. This post attempts to walk through some of the observations and insights that collapsed my estimates. A single invocation of GPT-3, or any large transformer, cannot run any algorithm internally that does not run in constant time complexity, because the model itself runs in constant time. It's a very large constant, but it is still a constant. They don't have any learnable memory about their internal state from previous invocations. They just have the input stream. Despite all their capability, transformers are fundamentally limited.[1] This is part of the reason why asking GPT-3 to do integer division on large numbers in one shot doesn't work. GPT-3 is big enough to memorize a number of results, so adding small numbers isn't too hard even without fine tuning. And GPT-3 is big enough to encode a finite number of unrolled steps for more complex algorithms, so in principle, fine tuning it on a bunch of arithmetic could get you better performance on somewhat more complex tasks. But no matter how much retraining you do, so long as you keep GPT-3's architecture the same, you will be able to find some arithmetic problem it can't do in one step because the numbers involved would require too many internal steps. So, with that kind of limitation, obviously transformers fail to do basic tasks like checking whether a set of parentheses are balanced... Oh wait, GPT-3 was just writing dialogue for a character that didn't know how to balance parentheses, and then wrote the human's side of the dialogue correcting that character's error. And it writes stories with a little assistance with long-run consistency. And it can generate functioning code. Some of this is already productized. This is an architecture that is provably incapable of internally dividing large integers, and it can handle a variety of difficult tasks that come uncomfortably close to human intuition. Could the kind of intelligence we care about be algorithmically simpler than integer division? This can't be literally true, if we want to include integer division as something a generally intelligent agent can do. But it sure looks like tractable constant time token predictors already capture a bunch of what we often call intelligence, even when those same systems can't divide! I'm raising my eyebrows right now to emphasize it!

The top results are long lists of technical terms, named hard skills. Python, algebra, statistics, and SQL are some of the most popular ones. Later, there come soft skills -- communication, business acumen, team player, etc. Let's pretend that you are a super-human possessing all the above abilities. You code from the age of five, you are a Kaggle grandmaster and your conference papers are guaranteed to get a best-paper award. There is still a very high chance that your projects struggle to reach maturity and become full-fledged commercial products. Recent studies estimate that more than 85% of data science projects fail to reach production. The studies provide numerous reasons for the failures. And I have not seen the so-called essential skills mentioned even once as a potential reason.

The top results are long lists of technical terms, named hard skills. Python, algebra, statistics, and SQL are some of the most popular ones. Later, there come soft skills -- communication, business acumen, team player, etc. Let's pretend that you are a super-human possessing all the above abilities. You code from the age of five, you are a Kaggle grandmaster and your conference papers are guaranteed to get a best-paper award. There is still a very high chance that your projects struggle to reach maturity and become full-fledged commercial products. Recent studies estimate that more than 85% of data science projects fail to reach production. The studies provide numerous reasons for the failures.

This is a story of how artificial intelligence, augmented reality, and virtual reality can transform the academic library into a hybrid space. The library becomes a network of digital connections between physical objects. A network that recommends resources based on personal needs, links print resources to multimedia, embeds interactive tools to enhance knowledge, turns the focus on making discoveries rather than looking for them, and so on. I'm going to talk to you today about information delivery, in the context of libraries, and what I see as the future of libraries. We are about to create a new experience.

And so, you want your teams to reflect the humans that you are going to be serving. So, for us, we want our team to reflect the customer population of our robots because we want to be able to ask the right questions. We want to ask the questions that our customers are asking. We don't want to ask the questions that nerds like me want to know. I have a very specific set of things that I would love our robots to have, which you know, big whoop.

What data actually matters for appraising a property? There is a long list of things to consider, this is a complicated process for humans and not much has changed with the process for decades. Something that human appraisers have struggled to consider are all of the unstructured elements on the property. Many of these topics have been too "subjective" for influence on your price estimate. Sure, if there are gross quality issues (damaged flooring, etc..) that can go into it, but your choice in tile for the backsplash?

Think of artificial intelligence (AI) and you'll think of factories of robots doing manufacturing jobs. You may think of logistics and evoke imagery of robots doing their thing for the likes of Amazon. However, whilst the media often focuses on these radical and wide-reaching applications of AI, there are other forms of adoption which are quietly stirring a revolution. We see this in the way AI is used, and being developed, within the insurance sector. What's particularly interesting is that these wide-reaching forms of AI aren't the headline makers.

Complexity and Scale - AI is meant to help solve problems too complicated for human capacities, so when a project depends on too many moving parts or is too large, let AI do the analysis, and focus your team's energy on finding new relevant data or evaluating next steps based on the model results. Prescriptive Solutions Needed - When you need to move beyond predictive analytics and need a true recommendation system that weighs your business needs against model outcomes, you need AI. Whether you actually trust and leverage the recommendations or not will depend on your organization's flexibility. High Stakes and Strategic Need - When there's no room for error and your competitors are closing in, AI can make the difference between pulling ahead and losing revenue. AI models aren't always perfect and require good data input and conscious revision, but if you can establish robust models, you'll be able to get proactive insights quickly and efficiently.

Summary: Even if you're not big enough to have a full blown data science group that shouldn't hold you back from benefiting from AI. The market has evolved so that there are now industry and process specific vertical applications available from 3rd party AI vendors that you can implement. There are just a few things to look out for. Now that we are squarely in the midst of the exploitation phase of AI, pretty much everything you read will exhort you to hire a bunch of data scientists and get busy. This is not to make light of the top level commitment and organizational effort that's necessary to establish an Advanced Analytics and AI Center of Excellence in your company.

Artificial intelligence in banking and personal finance is marching on, and it's taking lessons from institutional adoption of advanced technologies, said panellists at the AI expo in London. There are several ways that AI pops up in retail: predicting customer behaviours from big data, customer interaction with chatbots, or customized experiences based on how users are interacting with the system. Ash Booth, HSBC's head of Artificial Intelligence for the bank's Digital Assets Technology unit, explained that AI can be defined as "this idea of learning, planning, reasoning. Things we associate with human intelligence embedded in our technology." For Wells Fargo, it's about AI's ability to sift through a huge amount of data the bank has collected over the years.