OT: I forget ...

Investment Analysis Clubs Saul’s Investing Discussions
1

There’s a lot of talk about Big Data here, not to mention ML, AI, CloudThis and CloudThat.
I own companies that are driving these discussions. And yet, while I’m ashamed to admit it,
I must. I’ve dived so deep into some of these technologies to try to truly understand what
I’m betting my future on, that I’ve forgotten one of the basics.

I get this nagging question in the back of my mind lately. Here it is: What the hell is so
world-changing about big data?

Tech companies will build new tech with new tech (and sometimes Big Data, I guess.)
Marketers will have millions of new targets, and they will be increasingly precise.
Everybody loves marketers. :slight_smile:

But what about, say, a Middleby? (They are a leader in commercial kitchen equipment.) Not
too exciting, not too high tech, but try to eat out without them once. How will Big Data
help them? Or the cloud. How will the cloud help Middleby and millions of companies like it?
I can imagine some possible uses for BD and cloud use, but world changing?
What will non-tech fast growers do with these new technologies that will enable them
to dominate their fields? Or, is it just tech companies that will be able to put the bulk
of these world-changing technologies to productive use?

What about a retired investor sitting in the heartland trying to remember what he had for
breakfast? Will all these help him some how? (If I ever knew, it has slipped my mind along
with what I had for breakfast.)

We’ve all considered the day when we each have a chip embedded somewhere under our skin that
will tell (someone) everything they need to know about us. The thought makes me sick, but maybe
now that BD, ML, AI and their ilk are about to go supersonic, maybe we don’t need such old-fashioned
ideas?

Dan

8 Likes
2

I think you almost answered your own question … there are two aspects of Big Data & related technologies that can be world changing (IMO), namely …

  1. It allows new products & technologies. As an example imagine Amazon’s Alexa or Google’s Assistant with 2000’s voice recognition - it’d be unusable (I have a 2-year old VW group car with on-board voice recognition - it’s basically useless compared to say Google’s Assistant.) It probably wouldn’t even recognize what you were trying to say & then whether it could do anything with what it recognized would be debatable. In this case the management & processing of Big Data enables the robust voice recognition these types of devices have as without these technologies Google (for example) wouldn’t be able to easily capture & manage enough samples of speech to capture the various accents (even for only English speakers) & the the various neural nets wouldn’t be able to process all these samples. Another example is photo searching or facial recognition - those nets need to be trained on many many samples & capturing & managing these samples is problematic without the current tool set of Big Data technologies.

  2. It lowers the cost barrier to usable Data. In the past it was always possible to run large databases, the data within just had to be robustly codified which was difficult & expensive. Addresses had to follow various standardsn - if you mis-coded a zip code then that piece of data was potentially unusable for geographic analysis for example & if it was significant then poor decisions could the result. Keeping these databases accurate required robust standards & controls across potentially the whole organization - very expensive. Today using Big Data tools the ability to automatically extract a usable address is trivial, only significant exceptions need attention (vs. continuous attention in the past.) Big Data is also able to combine sources, perhaps one can’t determine the address from the main database, but another file has the latitude & longitude.

Those are two different angles that Big Data can help existing businesses. For Middleby perhaps big data is able to help add additional value via voice control, or by helping to lower food spoilage somehow (perhaps an oven that won’t burn food)? One example I read is how John Deere recently bought a start-up that is working on automated lettuce & cotton growth management (via things like only putting the correct amount of fertilizer & herbicide on the correct plants) without human intervention (the tractor towing the device can even self-drive if required.) Without being able to capture, manage & process the vast amount of data this development requires, this wouldn’t be possible.
https://www.wired.com/story/why-john-deere-just-spent-dollar…

15 Likes
3

I get this nagging question in the back of my mind lately. Here it is: What the hell is so
world-changing about big data?

Great question! Our rational mind has difficulty thinking like our subconscious thinks. The rational mind is boolean: “if this → then that.” That’s quite useless for predicting the future. Evolution works more like “if this → then maybe that, the other, or something else.” Evolution does not plan, it picks from random mutations the one that survives. If you think rationally (boolean fashion) about what I just wrote, it says that we can’t know the future before it arrives.

The Gorilla Game implies as much when it recommends to “buy the basket and sell the losers.” The science of complexity concurs, when you face a power law distribution you have a situation with more than one possible equilibrium point, one (the winner) is selected at random. This fact is a good reason to buy stocks of companies with good (market selected) track records instead of promising winners which is what venture capital invests in. Venture capital might get one in ten right. For individual investors those are not good odds, one in two is better and it needs to be much less of a block buster than venture capital’s one in ten.

With that mindset my answer to your great question is that we don’t know.

My first guess is that AI needs massive data to work like evolution does, picking from multiple scenarios like our subconscious picks from everything it remembers dating back to birth.

My second guess is that progress will continue eliminating labor with labor-saving devices we all love because they make life easier. That also creates a huge problem, how will we earn our keep when we have automated all jobs? So far we have automated mostly brute force jobs with machines, AI will automate thinking jobs.*

My third guess is that companies like Middleby will benefit from the diffuse “state-of-the-art” instead of from “if this → then that” just like you and I benefit from the Internet, for example.

Try mentally separating the universe into users and providers. We might get “if this → then that” right for providers but not for users – too much uncertainty.

Denny Schlesinger

  • I have a solution for that problem about which I have posted elsewhere. Too far afield for Saul’s board.
10 Likes
4

“I get this nagging question in the back of my mind lately. Here it is: What the hell is so world-changing about big data?”

Big data allows companies to make predictions and spot trends.

One of the earliest, best, big data examples was the Netflix Prize. Netflix tracks what you watch, when you watch, and both then makes suggestions and tells you how close a tv-show or movie is to your preferences.

They offered $1 million dollars to a person or group who could come up with a better algorithm for matching past history to future suggestions.

They took their “big data” and anonymized it, and broke it into two data sets.

The first was smaller, and contained the Netflix user, the movie they watched, the grade they gave, and the date they graded it.

The second was a much larger data set, with the above, but the grade removed, and you had to predict what grade the Netflix user would have given the movie.

Netflix knew the grade the Netflix user gave the movie, and they knew what their algorithm would have predicted. They wanted to someone could come up with an algorithm to beat the Netflix version. And a group did - it was 10% more accurate than the Netflix algorithm.

Typically, big-data analysis looks for relationships between factors - one is called the predictor, the other the response. For example, height and weight. Often the taller you are, the more you weigh. Height is a rough predictor for weight. When you add other predictors (gender, shoe size, waist size, age), your response (weight) gets more and more accurate.

Over time, with enough data, you can accurately predict someone’s weight if you have enough predictors. You can imagine how insurance companies would be very interested in determining what predictors lead to cancer, heart attacks, diabetes, etc.

Is that clear as mud?

As an aside, some researchers were able to deanonymize the Netflix data set - they used data from the IMDB website (ratings) to do it.

There’s a wikipedia article on the Netflix Prize if you want to get into more detail.

David.

21 Likes
5

The smart companies are figuring out how to use big data first too. Home Depot is using it to tailor each shopping experience to the individual customer:

At its recent Investor & Analyst Day Conference, Chief Marketing Officer Kevin Hoffman walked through Home Depot’s digital marketing efforts based on its accumulation of customer data and the numbers are nothing less than staggering. Hoffman said nearly 50 million households had been active at Home Depot in the trailing twelve months and that, through those interactions, the company was generating 1.7 trillion data points per week. While the data sets grow larger, the intent is for Home Depot to effectively get smaller so it can interact with its customers on a more personal level.

Hoffman says this immense amount of collected data will allow Home Depot to “tailor” its marketing messages to the individual consumer, rather than appealing to the lowest common denominator among a diverse group of customers. In his presentation, transcribed by S&P Global Market Intelligence, Hoffman said:

“[O]ur data allows us to know our customer at a more individual level. Our targeting capability will help us get to them at the right place at the right time, and we aspire to have all of our messages be tailored to the audience … Customers will expect retailers to speak to them at an individual level. The world of traditional one-size-fits-all messaging is quickly falling behind us. We’ve got diverse customer groups. Here, you see an affluent baby boomer, a Pro and a millennial who happens to be a new homeowner. The reality is … we regress to a lowest common denominator approach. We try to find the one message that somewhat appeals to all of them. But each of these customers has different needs and a different level of expertise.”

So, for instance, when these three different types of customers – an affluent baby boomer, a Pro, and the new millennial homeowner – go to purchase a new water heater each will be marketed with different material and given a different “Home Depot experience.” The baby boomer’s experience will be geared toward having a new water heater delivered and installed that day. The Pro will be given information on inventory, pricing, and delivery options. The millennial’s experience will be heavy on content, so they can decide on what size and type of water heater might be best for them.

Read more at https://www.fool.com/investing/2018/01/15/how-home-depot-is-…

Matt
Long HD
MasterCard (MA), PayPal (PYPL), Skechers (SKX) and Square (SQ) Ticker Guide
See all my holdings at http://my.fool.com/profile/TMFCochrane/info.aspx

3 Likes
6

What the hell is so world-changing about big data?

Let’s look at just one example from the past as to how big data could have made a big difference.

Remember Firestone? Used to be a premier American company rubber and tire company. In 1990 they were purchased at a deep discount by Bridgestone, a Japanese company. The reason Firestone went on sale was because failure with their tires was eventually linked to a number of fatal accidents. They suddenly incurred enormous liabilities. The linkage was accomplished by very dogged, labor intensive research by an investigative reporter operating on a hunch.

Firestone possessed all the data they needed to make this determination long before it imperalled the company. They lacked the tools to discover it. Of course there were business intelligence tools available at the time, but someone would have to look for the information hiding in the data.

Big data tools obviate the need for someone to first think that there’s something important to look for. The trend would have emerged and presented itself for management to consider.

Whether management would have done anything differently is a moot question. The difference is that the information would have been available to them.

12 Likes
7

Hi Dan,

You asked, [I]What the hell is so
world-changing about big data… what about, say, a Middleby? (They are a leader in commercial kitchen equipment.) Not too exciting, not too high tech, but try to eat out without them once. How will Big Data help them? Or the cloud.

What about a retired investor sitting in the heartland trying to remember what he had for
breakfast?
[/I]

And you are right, big data is not going to help Middleby or the retired investor who needs to remember his what he ate earlier. But how much economic activities do they account for?

Middleby won’t benefit from big data. But food delivery can be completely transformed by big data as autonomous vehicles and cashierless grocery stores may change how we consume, shop, and organize for food. Retired investor who cannot remember what he had for breakfast won’t benefit from the big data, but how we care for, treat, and diagnose people who are losing cognitive abilities will benefit from the big data. These activities constitute a significant share of our overall activities, and hence, support the progression and adoption of big data.

In my mind, one can find thousands, if not millions of small and isolated events, organizations, or activities that can go on without big data (or any data). But one will have a lot of trouble finding one significant part of our economy that will not be impacted by big data, from defense, to food, to healthcare, to work.

Just some random thoughts.

Cheers,

FG

4 Likes
8

What the hell is so world-changing about big data?"

This is really a question with two or three different answers.

One has to do with problems, like AI, which require large amounts of data, often poorly structured data and it is the AI which is world changing, not the mere accumulation of data.

The other has to do with the technology of handling large amounts of data, particularly poorly structured data. Relational databases are great at handling well-structured data, but not poorly structured data. Technology like the noSQL databases allows one to handle poorly structured data and it is the technology which is world changing.

A possible third is that, given this technology, we now can accumulate much larger amounts of data in a usable form and this allows us to ask questions that we could not previously ask.

3 Likes
9

Firestone possessed all the data they needed to make this determination long before it imperalled the company. They lacked the tools to discover it.

The issue with Firestone tires was twofold. The first and largest issue as that the bulk of the problem was actually Ford’s fault. The second issue is that issues negotiating with unions resulted in lower quality tires at a plant in Illinois while union workers were still on shift.

The Ford explorer originally called for 35 psi due to the weight of the vehicle. Due to the height of the vehicle and narrow wheelbase, Ford changed the recommended tire pressure last minute to 26 psi to lower the potential for rollover. This was largely fine for Goodyear tires at the time, as they were definitely overengineered. Firestone’s tires were honestly fine at that pressure, except that the one plant with union issues had some problems with their adhesives. The tire heating up due to the underinflation was what caused the tread to separate during the summer typically. Ford had also removed much of the roof material from the explorer to further lower the center of gravity, resulting in an increased chance of fatalities from a rollover. This combined with the narrow wheelbase that caused complete loss of control of the vehicle when a tire went flat, was the primary problem.

Ultimately Firestone took most of the blame for this issue. Note that in the years that followed, tire pressure recommendations for the explorer were changed to 30 and eventually 35 psi. The wheelbase was also widened by a couple more inches. Firestone paid a massive price because of upset union workers at that plant, when their tires from other plants were comparable safety-wise to Goodyear.

4 Likes
10

I get this nagging question in the back of my mind lately. Here it is: What the hell is so
world-changing about big data?

raptor:

Yes Middleby and EVERY company of every industry will use big data…without question. AI is being used in manufacturing already and isolating out weak parts or redesigning structures for improved strength, etc. Billions upon billions of sensors are storing valuable data from appliances and everything imaginable.

Big Data has become so big because machine power now allows feeding ever increasing amounts of data to improve intelligence…wasn’t like that just a few years ago when more data added nothing further to intelligence…now the power, perhaps world domination rests with big data…and the companies that can harness it…AMZN, GOOG, AAPL, FB, etc.

One of the worlds leading authorities is from Stanford with stints at BIDU among others. This lecture is an AI for dummies that you may find interesting…the next “electricity”:

https://m.youtube.com/watch?v=21EiKfQYZXc

8 Likes
11

I can also imagine that big data is/will be important in the food industry to address food safety issues. They would be able to track to source and determine extent exposure of the issue very expediently.

1 Like
12

What great feedback. My thanks and a rec to each.

I didn’t mean to suggest that any of these data techs were not important or don’t have the possibility
of advancing our lives. It’s just that the press seems to believe that it will benefit every person and
every business as long as they get in the loop on big data. I don’t see the reasons for this and some
here seem to be on that same page and some don’t. Maybe it’s a matter of scope. If some biotech makes a
super discovery, it will 1) make that company rich beyond all expectations, 2) save lives–quite life-
changing for them! and 3) won’t directly help those without a certain malady or disease, but help us all
collectively by extending numerous lives and lessen suffering of individuals and families.

I don’t disagree with any suggestions with the exception of one minor nitpick regarding the value of
some perfect voice actuated device to control every electronic device in our home. (I’m sure it will be
fun and interesting, but is that really necessary? Will it really become historically significant?
Really?!)

I guess my thoughts haven’t changed too much. In fact some of my beliefs (not written in stone) have
been reinforced, not what I expected.

LI** Thoughts and Even More Questions

  1. The benefits of the data and AI advances will not be equally beneficial to all individuals nor all
    companies (any or all of which may benefit indirectly and eventually and collectively.) As far as investing,
    Denny is probably right: “Buy the basket and sell the losers.”
  2. A lot of business hopes for advancement lie in increased benefit to advertising and marketing
    expense, which is a very valid use of technology, but maybe not the most valuable of possible benefits
    to mankind.
  3. I’m not sure where all the ‘unorganized’ data is coming from. Isn’t most desired data (outside of
    medicine, maybe) being generated directly by some form of systematic computer calculations or input?
    Credit card data, for example–product, price, location, time, income, other purchases, etc., are
    already available and if they aren’t ‘organized’ well shame on someone. What advancements are needed
    here? I’m not saying they don’t exist, only that I haven’t identified them.
  4. Once a certain level of advancement is accomplished, especially in AI, this may necessitate either
    the end of many companies’ current efforts to make that very advancement, or a seldom seen industry
    consolidation. Isn’t it likely that someone may (probably will?) achieve advances in AI to the point
    where other AI efforts become redundant and worth very little or nothing?
  5. Any huge leaps in data interpretation will surely end the need for many more jobs and possibly entire
    industries than it will create.
  6. How will our growing populations support themselves?
  7. As far as gadgets, am I the only one who doesn’t want a watch to tell me that I missed a 30-minute
    physical workout this week (shame on me!) or that gas is building up in my digestive tract? Am I being
    ignorant to remain unimpressed?
  8. Can we vote on where to put efforts for big data, artificial intelligence and machine learning? If
    so, I vote MEDICINE and elimination of POVERTY. If not, why not?

“Alexa, bring me another Bacardi and Coke, please, and warm up the hot tub.”

  1. Cool! Hey, can I change my vote?

Agradezco la opinión de todos. Ustedes molan. <---------Tengo que amar la tecnología.

Dan

** LI: Low Intelligence

2 Likes
13

All of that may be true. I worked at a big company where the products involved safety and human life hung in the balance. It made absolutely no difference what the source of a quality problem might have been, the company’s executive management was ultimately responsible for quality problems. Period. Full stop. BTW, I was the lead analyst for QA systems for the commercial products of the company.

Aside from that, everything you said has absolutely nothing to do with the points I made. Firestone had the data, but not the information with which to take action. The fact that sold faulty tires, irrespective of what caused the quality problems makes them more culpable, not less.

2 Likes
14

Dan,

What a fascinating topic you raised! I just wanted to offer my humble responses to these questions because you really made me think.

WARNING! I AM
NOT AN
EXPERT
in ANYTHING

1. The benefits of the data and AI advances will not be equally beneficial to all individuals nor all Companies. As far as investing goes, Denny is probably right: “Buy the basket and sell the losers.”
Yup and yup.

2. A lot of business hopes for advancement lie in increased benefit to advertising and marketing expense, which is a very valid use of technology, but maybe not the most valuable of possible benefits to mankind.
Yup and “mmmaybe not”. /sarc

3. I’m not sure where all the ‘unorganized’ data is coming from. Isn’t most desired data (outside of
medicine, maybe) being generated directly by some form of systematic computer calculations or input? Credit card data, for example–product, price, location, time, income, other purchases, etc., are already available and if they aren’t ‘organized’ well shame on someone. What advancements are needed here? I’m not saying they don’t exist, only that I haven’t identified them.
From what I have read about the big data integration wave that’s supposed to change the world (or at least business), “un-structured” data can include email, Powerpoints, videos, phone calls, machine vision I suppose, and god knows what else. It all seems imaginary to me.

4. Once a certain level of advancement is accomplished, especially in AI, this may necessitate either the end of many companies’ current efforts to make that very advancement, or a seldom seen industry consolidation. Isn’t it likely that someone may (probably will?) achieve advances in AI to the point where other AI efforts become redundant and worth very little or nothing?
Maybe so. But what seems more likely to my gut is that just as ubiquitous electricity led surely to the integrated circuit, this board and this topic of discussion, the deployment of big data will lead to the emergence of currently-unthinkable ideas. Abraham Lincoln could never have imagined a phonograph, let alone an iPod. Fred Flintstone thought everything had already been invented and they didn’t even have electromechanical sirens.

5. Any huge leaps in data interpretation will surely end the need for many more jobs and possibly entire industries than it will create.
As robotics is already.
6. How will our growing populations support themselves?
YES!! As Denny noted as well. What are we going to do about that? What happens when “full employment” equals 10%, or 1%, of able bodied adults? How do the rest of us “earn our keep”? It seems to me that we would have to completely re-imagine the entire economy, nay, the entire society. Am I wrong? Where do you start?

7. As far as gadgets, am I the only one who doesn’t want a watch to tell me that I missed a 30-minute physical workout this week (shame on me!) or that gas is building up in my digestive tract? Am I being ignorant to remain unimpressed?
I can say definitively that that you’re not entirely alone… But I can no longer imagine (except in post-apocalyptic nightmares) being without my phone, or this computer. Here again, my thinking is along the same lines as #4 above. (If you want to see some completely wacked-out visualisations of gadgetry, I highly recommend “Black Mirror” on Netflix.)

  1. Can we vote on where to put efforts for big data, artificial intelligence and machine learning? If so, I vote MEDICINE and elimination of POVERTY. If not, why not?
    I’m totally with you. I’d just add elimination of global hunger, and global accessibility to education.

Again, JMHO, with no guarantees of validity or usefulness.

Many thanks to all the contributors to this, what can only be the best investing board anywhere. I thank you all for your willingness to share your thinking with others. I have gained so much in so many ways.

The name’s “Jack”, not “Master”

1 Like