101: Avi Goldfarb - Why Should AI Matter to You?

Are you thinking about how to grow your business with near and offshore staffing, but you're also wanting to make sure that it's reliable, cost effective and really well done? I think docsa talent is the way to go. Here's why. They don't just fill your jobs. They bring integrity, top talent and a people first approach that's hard to find. I've known the people at Doxa talent for over 20 years now and I say that you can count on them. I have. Talk to Docs of Talent and you won't be disappointed. Go to docsatant.com Bill G. Again, that's doxataalent.com Bill G. One more time. Doxatalent.com Bill G. Check it out today.

Well, AI actually matters to all of us right now. I'm not talking about that sci fi stuff where the machine becomes unaware and either destroys humanity or becomes your best girlfriend ever. I'm talking about prediction technology that prediction machines. So today's guest is going to be talking to us all about the kind of AI that you can use in any business today. So listen, listen in. Hello and welcome again to the Scaling Up Business podcast. I'm your host and business growth coach, Bill Gallagher. On the show with me today is Avi Goldfarb. Avi is going to talk all about AI, right? Not the general AI, Terminator machines, becoming aware kind of thing, but the kind of work that that's actually being done in AI, that that matters to you and your business. So instead of thinking that's something else, that's not us. Everybody ought to be listening into this thing because it's really about something else. It's about making great prediction faster and less expensive. Right? So Avi joins us from Toronto where he is and lives today. And Avi is the chief data scientist at Creative Instruction Lab. He's senior editor at Marketing Science and he's a research associate at the National Bureau of Economic Research. He's also the author of Prediction Machines. Avi spoke recently at our Gazelle's Growth Summit at the Scaling Up Growth Summit in New Orleans and he talked about what are the economics of AI, how does it matter? What does it mean for all us as business people? And I think you're going to love it. Welcome to the show, Avi.

Thanks. It's great to be here.

Yeah, it looks good. Most folks won't see the video, but I can see like a beautiful sunny day out behind you there and we're sunny here on this coast. I was just in Toronto recently. I saw you in New Orleans. Like, it's funny how we cross and bump in all over these places, right?

Absolutely.

It didn't look like that when I was in Toronto. We were just finishing up a wicked ice storm. I flew in from Honolulu.

It was quite a shock, but it's pretty nice today.

Awesome. All right, well, so let's get into it. Why don't you tell us your background or kind of like, you know, you come not from the world of computer science, but from economics and business study and have become more interested in this world of deep learning and prediction. So talk to us about your background. You know, did it start when you were a little boy? What's the deal?

Okay, so I guess it started in grad school. So I was, I was living in Chicago, I was at Northwestern, and it was the late 90s and there was this crazy new technology called the Internet. And everyone seemed to be very excited about this technology and we knew nothing about it. And so that seemed like the perfect thing for a student trying to get a dissertation, a doctoral degree to get his head around. And so I spent the next few years trying to get my head around the economics of the Internet with an emphasis on, on competition between search engines. So back in those days there was genuine competition between search engines. And when I was studying them, Google was in the data, but they were number 17. So there were 16 search engines that had higher market share, at least in my data, than Google. And so for the next decade or more, I spent studying the economics of the Internet and with an emphasis on technology, I looked at inequality, I looked at privacy, I looked at crowdfunding and a handful of other things. Then around 2012, we started running a program here at the University of Toronto called the Creative Destruction Lab, which is a program for helping science based startups scale. We had these very early stage startups, usually pre seed money coming in with great science, great scientists, and they were interested in building a business. We helped them build a business. Starting in 2012, there were these companies, there were a couple companies calling themselves AI companies. Then in 2013 there were a couple more. And by 2015 through our program we had almost 50 companies, I think calling themselves AI companies. And now we have over 100 every year. Me and my co authors, all University of Toronto business school professors, decided to try and get our heads around what this technology meant. We got that early lead because it turns out by sort of happy accident, a lot of these technologies, a lot of technology underlying the current excitement around AI. Deep learning in particular was invented at the University of Toronto, like three blocks from where we are. And so their computer science PhD students were the people starting the companies that came through our lab and that we helped scale. And that gave us an early sense of what was going on and the ability to really think through where the opportunity is for these startups to help businesses.

It's interesting in a topic area born of an interest in the Internet, that proximity matters.

Yes, yeah, absolutely. Although it turns out that a core theme of my old research on the economics of the Internet was often the Internet helps cities relative to rural areas, not Hertz, for a variety of reasons. And sort of the old, the old story there is when you build roads, you think you're going to help people in isolated places commercialize. But what it often does is it allows young people from those places to move into the cities and still communicate back and see their parents every once in a while. There's some sense the Internet did the same thing.

It's interesting. A lot of people will say things to me. I coach a variety of companies and although I'm a Bay Area Silicon Valley guy with certain tech background and networking and telecom and people said, oh, that's true for tech companies. Or that's like. And I immediately think, like, didn't you just live through all of this? Every company is a tech company or should become a tech company. When I think about even like I worked with a company for a couple of years that work, that, that's auto mechanics, right? But they're significantly tech enabled, tech driven. And part of the reason that they scale and are doing interesting things is that they can have visibility to where their mechanics are at all times and they can route them and dispatch them. And I mean, every business ultimately becomes a tech business or should be thinking about it and not just thinking about sort of the more pure place, but even those, if you think about like what works around this neighborhood, it's like, all right, pitch me how you're going to Uberize or Airbnb the next thing. Like, what is it that you've thought of that we didn't think of already and how are you going to do it? Right? So ultimately, tech enables and empowers every single interaction of everything. It's, it's the tool or the latest or the lever, right? I remember fiddling with early networks ARPANET and dial in VBS's and corporate networks early on. And it was a great place for certain sort of geek that could get onto them and knew how to play there. And then I remember the first time I said, The Mosaic browser. And I was just like, holy crap, this changes everything. And I knew in a second, like I'd read about it, but when I saw it, I was like, oh, yeah, okay.

Yeah, so that's a great way to think. What the Mosaic browser did is it made the Internet and these ideas accessible to the general population. It took us a decade to figure out really how to commercialize that and really make money off it. But now it's hard to think of a business that's not an Internet business. I think something similar is going to happen with AI once people get their heads around what this technology is and don't sort of fall prey to over promising and under promising. We're going to see every business in various ways is going to use AI tools to solve particular problems.

So why don't you talk to us about. Because we have a wide range of listeners. Right. Some people will be very savvy and understand, but they'll have to pardon us for just a minute. And why don't you talk for a second about what do you mean by AI, what is AI and distinguish it from it? Because I think a lot of people, it's a buzzword that they haven't really stopped to think about. Right, Great.

The current excitement around AI is based on some developments that really happened in the commercialization of one particular technology called machine learning over the last 10 years. Machine learning is in the general computer science world. It's a subfield of AI, but in terms of business applications, it's the subfield, it's the one you should be paying attention to. What machine learning is, is prediction technology. It's taking data that you might have sitting around your company or data you might collect and using it to fill in missing information, using it to predict things to help reduce uncertainty. It turns out prediction is really, really valuable because even though this isn't general intelligence, it's still a lot of the decisions you make will be better if you have better prediction. Why? Because decision making is all about trying to figure out what to do when you face uncertainty, what prediction is. It's reducing uncertainty. It's making you a little more confident about what the world is like, what you see, what's going on. And that can be a useful tool in all aspects of your business. Sometimes it can be totally transformative.

So just to underscore, because you're starting to go right in the meat of it, what we're not talking about a conscious machine utopian or dystopian. We're not talking about something that's aware. And any of that, that's all cool and interesting, but it's not this conversation. This conversation is about the other aspect. Prediction technology and machines learning to predict things that are utility to your business. Right?

Absolutely. So I don't think that general intelligence idea is crazy, but it's not the technology, it's not the reason we're talking about this in 2018 and we weren't talking about it 10 years. What's really changed is prediction technology has gotten better. And some people decided to call prediction technology artificial intelligence. When you hear AI, you should fill that in as a prediction machine, a machine that helps you fill in missing information and reduce uncertainty.

Well, I was working with a sizable ad tech company a couple years ago and they were doing this AI driven ad buying. Right. So they're just applying a machine learning thing. So to think about it, for, for everyone here, they're giving a loose set of rules and outcomes to a machine and they're saying, go figure out how to accomplish this on your own. And, and it's looking at things, a whole bunch of things really fast.

Yeah, yeah.

And produce this outcome. Right. So. Or optimize for this value.

Exactly. So it can, you know, the applications are really varied. You can think about it as sort of a drop in the cost of prediction quality, adjusted cost. This is, some of you may remember some economics. When the price of something is low, we buy more of it. Or maybe that's just basic business, it's not really economics. And so when the price of prediction is low, when prediction machines get better and better and better, we're just going to use prediction all over the place. And so we're already using AI, we're already using machine learning in good old fashioned prediction problems like should we give somebody credit when they go to a bank or should we insure somebody? Because that's a prediction about whether they're going to make a claim against the insurance company and how much that's going to cost. But increasingly there's all these other applications of prediction that you might not have thought of as prediction. It turns out, for example, medical diagnosis. Figuring out what's wrong with you is prediction. You take predictions, information about symptoms and filling in information about disease, recognizing a picture. So the classic problem in security is somebody walks in, whether physically or virtually, and you got to figure out who that person is and whether they're a good guy or a bad guy. Essentially, that's a prediction problem. Facial recognition is a prediction problem. You take in information about the, the visual information, you fill in the context and a label, whether a name or just a, okay, this is a good guy, let him in. This is a bad guy, don't.

Right. And we see this all over the place. Like you go and post a picture onto a social network now and you're likely to be asked, hey, is that Avi Goldfarb?

Right?

Yeah.

And that's AI Operating. That's Billy and yo, here's a picture. What's the label? Absolutely.

Yeah. Really great. So give us some more examples of this so that people can begin to think how might this apply in their business? Let's talk about what do we need to do with it?

Right? So all over the place you can think about, you can start to identify picture problems. Whenever you see uncertainty in your business, you might apply an AI tool to reduce that uncertainty. And sometimes this can mean that you can really serve your customers better. Or you might be doing things that are very expensive and they only exist because of uncertainty. So my favorite example of that is an airport lounge. Okay. So you go to the airport, you might sit in a lounge. The only reason the airport lounge exists is because we have bad prediction about how long it's going to take to get to the airport and then how long it's going to take to get through security. If you know exactly how long it's going to take to get from your home to on the plane, then you will never need an airport lounge. You'll just show up at the airport and walk through and get right on the plane.

Well, the cost of being wrong can be high. Right? So if the cost of being wrong there is low, then I could afford to be late. But the cost of being late could throw a vacation, a key meeting, a conference, like that.

Sure. So, but that's just a statement about our. If we had better prediction, we wouldn't have to make these compromises. When the cost of being wrong is high, we build these institutions around playing it safe. And sometimes if you had better prediction, much better prediction, you wouldn't need these institutions around being safe. You wouldn't need airport lounges, and you wouldn't need another. One of my favorite examples, a good.

Example for that is I use the routing mapping software prediction to figure out tomorrow morning what time do I need to leave to get there by X time. Right. And then I know that I use this clear and this TSA PreCheck. So the security thing is almost immaterial to me. So I'll leave less buffer than I once did because of that. Right?

Absolutely. And that's because your Prediction about how long it's going to take has gotten better.

Right.

And so in terms of mapping and getting places that just generally the prediction technology we have with the maps on our phones, various other tools have really helped reduce the amount of time you hang around waiting once you get there.

Yeah, it's a great example. And then that has all sorts of implication for every business about being just in time or thinking about things that only exist because of that uncertainty. Like the lounge that I have access to, that I use when I am there really early and that I pay a premium for because I travel so much. Does that matter? Does that go away? It exists only on that. Right.

It exists only on that. There's no, I guess there's a handful of people who are transiting through an airport because of they have two legs on the journey. But for the most part, the reason they're there is because of uncertainty. You solve the uncertainty, you don't need it anymore, and you serve your customers better. And so there's all sorts of compromises we make in our business and our daily lives because we're uncertain, we're not willing to take that risk because the consequences of being wrong are too high. And if we have better prediction, we don't have to worry about that so much.

Yeah, And I guess that matters on whatever side of it, whether you're buying something or selling. So immediate things come to mind. Labor demand, commodity pricing, energy, all of those kinds of things.

Absolutely.

And of course the ultimate like just straight investing. Right. Prediction being applied, but then you're competing with everybody else doing prediction. Right?

Right. So in the investing story, once you have everybody doing prediction, then we end up back in the same place we are now. So it's a short term benefit, but in the other cases it might not be. One of the things I think is really exciting right now is thinking about the opportunities around language translation. So one of the reasons you don't do business abroad, there's a handful of them, but one of them is because you don't understand the person on the other side. If you can reduce that uncertainty by enabling translation, it's going to be much easier to do business. We already see some evidence on that. That machine enabled translation on sites like ebay is leading to way more international transactions. Because translation means that people understand each other, they fill in that missing information. You can predict what the other person wants and you can provide it and trust that person.

Yeah, and I suppose, I mean, I think there's a long time we translate stuff and that makes it Way easier to understand. And yet if you're doing business in Japan, you're not going to get a straight no most of the time. Right. So understanding that is a cultural nuance doing business in many cultures. And the price negotiated isn't the final price. The final price continues to be negotiated until there's an exchange of goods and value that's not there in the words that get translated. But we are seeing real time translation now with even mobile technology, which kind of blows your mind.

It's really changing the world and it's making the world a lot smaller. It's very exciting. And then sometimes this kind of technology can really transform the whole way a company operates. So think about Amazon's business model. What does Amazon do? Amazon has a catalog of millions of items, but a catalog, they make some recommendations to you. You look through that catalog, you choose what you want, you shop, then they send it to the warehouse and their warehouse ships it to your door. That means that Amazon's business model is the same business model that Sears catalog had 100 years ago or more. You shop through a catalog, you look through a catalog and they ship it to you.

Right.

Now they do have those recommendations.

It's faster.

It's faster.

It's much bigger.

Yeah. So it's the same business model. They do absolutely do all sorts of aspects of it. Much, much better. But at the end of the day, the business is the same thing, but better. You can imagine moving to a model where the predictions get so good about what you're going to want, but they don't have to wait for you to order it. They're going to know what you want. Maybe not 100% sure, but 90% sure, 80% sure, even 70% sure. That might be enough for them to ship the item right to your door. And then you can open the box at your door and decide what to keep and what not to.

Well, this is a little bit like, to some degree, it's a little bit like listening to music on Pandora. Right?

Right, Absolutely.

I'm in the mood for some Kali Uchis. Right. So now it's going to find all sorts of things that sound like that and play it for me.

Yeah, right. And their prediction engine will try and figure out what you're going to want. And that has the advantage that the only thing you're paying with song by song is your time. But the same idea, it's also a little bit like if you remember the old Columbia House business model for CDs, where they mail them to your door or in records, you get the record of the month and then you could choose to send it back. But it was a much easier prediction problem because they just send you the hit album of the month. This is a harder prediction problem. You have millions of items. They got to figure out what to send to you. I don't know if they're going to do this eventually or not. They do have the patent for it. They thought about it pretty deeply.

Well, they started early on with one click ordering, and the patent on that was pretty questionable. But the notion of making it easy, like, look, stop thinking so much about this, just name it and it'll show up. Right?

Right, Absolutely. And this is taking that to the next level if it happens. And there's lots of industries where you can think through, well, we do all these compromises because of our uncertainty, and if our predictions get better, we might not need these compromises. Another example I like to think about is when you're hiring somebody. When you're hiring somebody, maybe the most painful part of that process is the initial resume screening. You have hundreds of applications, if all goes well, and you have to screen through them. Any human, when they screen through them, they're pretty bad at screening the 186th application. They're pretty good at the first one, but they're quite likely to miss somebody when you go through that huge stack. Now, machines are already doing that, in many cases doing that with better quality and less bias than humans. But we can even go to the next level. Why even call for applications at all? Why even post a job? If your prediction's good enough, you may be able to go to a database like a LinkedIn database, identify the right person for your job, and make them an offer, essentially, or at least to make an offer for an interview without having to go through the painful process of screening and whatnot. And that even goes beyond, if structured right, you can also deal with issues around bias and networks and that kind of thing too.

So two things immediately spring to mind that probably are coming to our listeners as well, and that's that as a consumer, as a business like, look, if I can predict what you want and do and bring great prediction to all areas of buy and sell in my business, then, and make and produce and deliver that whole world, I can tune that and get better. And if I can do that before my competitors, I have an advantage, and so on. On the consumer side, it's nice not to have your world, your inbox, your. Your attention filled with things you don't care about.

Absolutely.

And I Remember, well, years ago, I managed the consumer international business for Sprint. And the international calling was by far and away the most profitable piece of the telecom business at that time. And so we were highly motivated to find these high volume, highly profitable customers and cater to them and so on. And so there's a lot of motivation there. And we got pretty good at it. Like, we could predict who was and who wasn't where you probably called. And we had some fancy ways of doing that. And when we did that, well, it worked really well, but it also creeped people out. So if I sent you a campaign, like a direct mail piece or I called you or something and I said, hey, if you call Europe a lot, we've got a great deal for you, that worked really well, or hey, we finally have something for India and you just get it and you're like, you're grateful. But if I sent you a campaign and we did this one, that backfired massively said because you call Germany a lot. There were other examples, but the Germans got upset first and fastest and loudest because we indicated that we, we knew with some high degree of certainty that Germany was a place they called and they were just freaked out by that intrusion, that privacy, that lack of anonymity. Right. And now we have GDPR in the news right now.

Right. So the core challenge. So prediction machines, AIs need data in order to make their predictions. And so that means that in order to make them work, there has to be. You need to convince your customers to give up some of their privacy. Now, that's a relationship in the sense that if you use too much data and violate the privacy too much, then you'll lose all your customers. But if you can't figure out any way to collect any data from people, then you can't use a prediction machine, you can't use an AI getting this balance right, so that you are both trusted and trusted so that you're able to use data. But that trust comes from using it properly and appropriately and clearly is tricky.

Once again, execution matters.

Execution absolutely matters. And different companies have taken different tasks. At the company level. You have to make sure that you respect privacy and you collect data in a way that consumers can see the benefit they get from it.

Well, so execution matters, the relationship and the trust. If you get it wrong, they'll let you know massively. Which is, I think, why it's useful to experiment before you scale.

Yes, absolutely. Yeah, Small experimentation is always useful. And just getting a sense of how are you making sure that you can communicate clearly what you're using data for is important.

Yeah. And you also talked about judgment in your talk. Right. And so loop in, because I think we're getting into that area of using appropriate judgment here.

Right. So a prediction on its own is useless. A prediction on its own doesn't tell you what to do and doesn't even tell you what. If you have a prediction machine, you got to figure out what to predict in the first place. And so what judgment is, is knowing the things that matter to your organization, knowing what things you should predict. And then once you have predictions, oh, 75% chance this is going to happen, 25% chance that's going to happen, knowing what to do with it. The way I like the example I think is most powerful in explaining what judgment is comes from the old movie I, Robot. I don't know if you remember this movie. Will Smith was a star. And there's a scene in that movie, it's a flashback scene where Will Smith and a little girl are in a car accident. They go over a bridge and they're sinking into a river and a robot comes along and rescues Will Smith and not the 12 year old girl. So the girl drowns and Will Smith is really angry and thinks the robot made the wrong decision. The robot can only rescue one of them. Will Smith thinks the robot should have rescued the girl. So why did the robot rescue Will Smith? He was an adult male. He had a apparently a 45% chance of survival. The girl had an only 11% chance of survival. And so the robot went with the person it predicted would have the higher chance of survival. Will Smith objects because essentially he's saying that little girl's life is worth four times or more as much as his life. And so the robot should have rescued the little girl because 11% of her chance, her survival is more valuable than 45% chance of his survival. And that may very well be true, but at some point that would switch maybe. I don't know if it's 8% her survival or 1% her survival or 0.1% her survival. But at some point we'd have to make a different decision and the robot would have made the right decision to rescue Will Smith. That understanding what to do with that prediction is judgment. Your prediction machine is useless both in big life and death situations like that. But even in smaller situations, or more precisely in bigger business situations, if you don't know what you care about, if you're not willing to explicitly say what you value, then it's going to be very hard to use a prediction machine. And so a key job as we roll these things out is what we call a reward function engineer. But you think about that as somebody who does judgment, somebody who tells the machine what to predict. But then once you have that prediction says, okay, if we're going to get this prediction, then we should do this, and if we're going to get that prediction, we should do that. What do you do once you have the prediction? That's judgment. And that's going to be the most valuable skill as these things roll out.

Yeah, and look, that's not even true. Just a machine learning prediction AI world. That's true in our business, period. And it goes back to people who are familiar with not only scaling up, but Jim Collins work will know that we're very interested in figuring out what the most important metrics are in your business. And, and your profit per X is a concept that we talk about and we coach people on to figure out that thing. And the key of one profit per X versus another is it leads you to very different decisions. If I pick this one and I optimize for that, I start to get this and you just extrapolate it and then you say, well, which one is more valuable to me? And who knows? That'll really be determined by where you're trying to go in the long run and why you exist as a business and like that. Right?

Absolutely. And what this is saying is if you could get away with being a little wishy washy on that before, if you're going to use AI in your business, it's going to be harder and harder and harder because you can't use this tool unless you have the metrics about what you really care about.

Right. Well, and that goes to fundamentally. Well, I remember somebody asked me once, well, had you raised more money for a business, a startup business I had in the 90s, what would you have done differently if you'd start, you know, I started with like a half a million dollars and I said, what if you'd raised 5 million? I said, I would have made the same mistakes but bigger on a much bigger scale. So you bring a powerful new tool like machine learning to bear on something and you make predictions. Now you're going to do things bigger and faster. So it's just a tool that amplifies like a lever, it amplifies that effort. So you're going to get bigger and faster and the fundamental inputs are simply amplified, right?

Absolutely, absolutely. So yeah, this gives the ability to scale and there's good and bad to that.

Well, test. Right. Test, Evaluate as you add gas to things. I had a client we worked with. It's been a couple of years now since we started, and it was a new CEO taking over a business. They bought into this business and so on, and they invested big right away in something that made sense to virtually everyone that heard it. Lots of different folks heard, okay, well, so we're going to drive marketing in this way and it's going to be really great. And so we're going to hire people who are good in this and just give them some money and turn them loose. And it failed massively and cost a lot of money. The company recovered and is doing fine today. But it was an expensive lesson because it all made sense. And what they didn't do was do a little testing and iteration before they added a lot of fuel to that fire. Right.

Yeah. And if you don't test and explore and ask, then things are weird stuff.

Yeah.

No matter what you do.

We've covered a lot of good things. Have we missed anything?

I think we've covered the main ideas. So, yeah, this is.

So get interested in AI, but, but only really AI in terms of machine learning, prediction. So figuring out what you need to predict and then, and then apply it with good judgment to the right place using some testing and iteration before you go nuts.

So the way I think about it, yeah, absolutely. So the thing is, look, if you're, if you, if you're paying attention to things that are going to affect your business in the next five to 10 years or even the next year, then you don't need to worry about general AI. When you hear artificial intelligence, just ignore that. Insert prediction machine. And then once you understand that that's the technology that's driving the excitement and that's the technology that's useful to you almost surely in some application today, this week.

It's brilliant. Well, the book is Prediction Machines. You'll find it on Amazon or predictionmachines. AI. You can get it either way. Avi Goldfarb is the author and our guru. I think it's a useful tool topic that we all need to think about and begin to figure out how to apply if you're not already, or apply it better if you are now. I thank you so much for getting on the show with us today.

Thanks for having me here.

And if you're interested in other matters of scaling up, you'll find us@scalingupbusiness.com where we have free downloads of growth tools and guides to help your team get started. With all the basics of scaling up, how to run your meetings, how to choose metrics, how to develop your strategy, all that kind of stuff. Much of it for free as well as a podcast like this every week with great gurus, authors, CEOs, other coaches with interesting case studies and so on. So every week we bring you something new. Go find us@scalingupbusiness.com thanks again. I want to give special thanks to our original growth guru Vern Harnish. Our show producer is Nio Bade. Audio production is by Podfly Productions, audio editing by Albert Burge and our show notes are compiled by Ein Kodina with proofreading by Tim McGowan. If you got some value out of our show today, please share it with somebody else and let us know the parts that you loved. Thanks again for listening and keep Scaling up.

Are you ready to turn your goals into daily wins?

You should check out our Scoreboard software.

Powered by Align, it's the official software for scaling up. With Scoreboard, you'll get an AI driven coach for priorities and KPIs and it'll help keep your team focused on what matters most. With alignment you can track your goals in real time and you'll even see results on the go. With their mobile app you can rest easy knowing that Align has world class onboarding and support to set your team up for success.

This is really more than just software.

It'S what I actually use to manage.

Growth in our scaling up client teams.

Unlock clarity and results with Scoreboard by Align. To try it now, go to aligntoday.com Bill Gates one more time. Aligntoday.com Bill G that's aligntoday.com Bill G Check it out and try it now.