How to Lose Your Shirt in the Restaurant Business

A project log for Using A.I. to create a Hollywood Script.

Use of an inexpensive microcontroller as a creative adjunct, by incorporating principles of an AI. Be VERY AFRAID!

glgormanglgorman 08/20/2023 at 13:060 Comments

Alright, I asked Google Bard to write an outline for a book entitled "How to lose your shirt in the restaurant business".  Maybe I should also try asking it for an outline for "How to get taken to the cleaners in the fashion industry", or perhaps there is another take on "How the rigged economy will eat you alive".  Somehow, I suspect that no matter what I ask it, I will get something similar to what you see here.  So let's begin to tear this apart - line by line, shall we?  Why, you ask?  Well, why not?  Writing "something" using some kind of template is a time-honored hack of sorts, especially popular in college when you know your term paper is due the next morning, and you have been partying all semester.  Used to be that there was a time that many a graduate student could pay their way all the way to their MBA or law degree or whatever, by offering some kind of "term paper" writing service, that is for those who are desperate enough, and who had the $25 to $50 per hour to pay to have the aforementioned grad student (or upperclassman) help get a vulnerable, pathetic freshman or sophomore out of a jam.  Some term paper writers are more ethical than others - do you have your outline on 3"x5" index cards?  Did your instructor already approve your outline and abstract?  If the answer was "yes" then if you have the cash in hand we can proceed, otherwise - sorry I can't help you on this one.  Now, let's get back to bashing Google - so that we can have some kind of idea what should go into an outline, or not.

 First Google Bard says this - which is pretty hard to screw up - no matter how hard they try.

Sure, here is an outline for a book entitled "How to lose your shirt in the restaurant business":

So they continue now with the suggestion that we have some type of  "Introduction" and then  they tell us that  "The restaurant business is a notoriously risky one."  Likewise "Many restaurants fail within their first few years of operation."  They then suggest that "This book will outline some of the most common mistakes that restaurant owners make."

Now let's rewrite what I just said in PERL. (Warning my PERL might be a bit rusty - no pun intended)

$1 = "restaurant"
$2 = "Introduction";
$3 = "The $1 business is a risky one";
$4 = "Many $1"+"s fail in the first few years of operation";
$5 = ... 
print "So they continue now with the suggestion that we have some type of $2 ";
print "and then they tell us that $3.";
print "Likewise $4.";
print "They then suggest that $5.";

O.K. For whatever it is worth - now we have some kind of "template" for re-hashing someone else's material, while we wait for the deluxe pizza to arrive.  Just plug some relevant stuff into the variables $1, $2, $3, and $4, and before you know it - my CGI will call your CGI and maybe they will hook up, or something like that.  Maybe that is what is supposed to happen.  Otherwise, at least this is something that can be programmed on a microcontroller - possibly in MicroPhython for example, i.e., instead of PERL - since I know that MicroPython exists on the Parallax Propeller P2, as well as on some of the better Arduino platforms.  So let's toil on, shall we?

If we are willing to state that pretty much any business can be a risky one - we can play with our analysis script a little further, just in case "The DJ business is a notoriously risky one" also, or the "Wedding photography" business and so on.  It is also a fact that "most small businesses fail within the first five years or thereabouts, that is according to [pretty much every book ever written on the subject, as well as the IRS, and the SBA, as well as quite likely your own local Chamber of Commerce.  So we could perhaps add more PERL or Python code to help create the data for the strings that use $1, $2, $3, and $4.   Now let's get onward with the task of trashing the first Chapter: "Choosing the wrong location"

Alright then - let's paraphrase and expand according to what Google Bard  has suggested: "The location of your restaurant is one of the most important factors in its success." Let's see now, according to Google Bard, location is everything, so, of course, we have to prove Google Bard WRONG!  Where would you rather operate a restaurant, as a concession in a busy food court in a big shopping mall with lots of competition - and high rents that have been socially engineered by the oligarchy to break the bank of the naive investor, even if only for the purpose of increasing the disparity between the disappearing middle class and "them"?  Or would you rather work out of a converted beach house that serves as a romantic bed and breakfast situated near some secluded, but romantic cove?  Obviously, having an accessible location is important, but what is even more important than the amount of traffic is the quality of the traffic, and how that traffic contributes to your bottom line.

Then in chapter two - Google Bard wants to tell us all about "Underestimating the cost of opening a restaurant", I will spare the PERL vs. Python breakdown here, at least for now - since hopefully, the reader has some idea by now about how to create a very convincing, even if seemingly fake AI, that just might do a better job than the real thing - that is - without begging the question of whether there is a method for turning narrative into code that can in itself turned into code!  Nonetheless, Google wants us to think that "Opening a restaurant is a very expensive undertaking", which may or may not be true.  I don't; know if there are any successful Mexican food diners that started out as "Taco trucks" that catered to "construction sites", or "Hot dog stands" that got a concession "not all that far" from some stadium or other venue.  This might be one way to avoid many of the "hidden costs that many first-time restaurant owners do not factor in."

Other ways to go wrong include the subject matter from Chapter 3, which deals with the fact that some people might succeed despite "not having a clear concept" at the very outset, with respect to many of the parameters that might affect their business decisions.  For example - if you are stuck on a question like "What kind of food will you serve?", and think that you are only going to sell Tacos, Hot Dogs, or Chinese, well may or may not succeed, but most likely you WILL fail, and that is because if you can only do one thing well - chances are that there is someone who can, and will do it better.  On the other hand, if you are an experienced chef, and can adjust your menu offerings according to your customer's interests, then you have a much better chance of getting favorable reviews and repeat customers.  So this is also a dangerously slippery slope - that is - if you are thinking that "your clear and focused" plan will be laser accurate, when in fact, you might be headed down a dangerous and delusional rabbit hole that you won't be able to dig yourself out of. 

Which brings us to the next question: "What kind of atmosphere do you want to create?"  Again - the path to perdition is well trodden.  Let's change the subject.   Imagine that you want to start a business as a mobile DJ, and ask the same question "What kind of atmosphere do you want to create?"  Now do you see why you are about to fail?  Yes or No? Do you understand why you are going to FAIL?  The question, if you want to be a DJ, is NOT what kind of atmosphere you want to create - since that will cause you to lose your shirt - and fast.   Imagine that some of your customers are going to want to book a DJ for a wedding, others want a DJ for a college party, and then there is going to be the occasional 50th anniversary.  So now you get a call at your restaurant, which has been empty for a month, and someone wants to know if you can cater Chinese, Italian, and Mexican food for a wedding party with 100 guests, and all they want to know is - how much does it cost, and can their DJ use your sound system.  What sound system you say? Next.   So while "having a clear concept at the outset might seem essential for the success of any restaurant", the fact is that it is more likely than not that you will have to adapt to hard-to-foresee market dynamics, where your "clear concept" just might put you at risk for an unmitigated financial disaster, which sadly might have been foreseen - that is - had the business plan not been drafted so narrowly, or perhaps even dogmatically so,

In Chapter 4 we are supposed to learn about how "Hiring the wrong people" will contribute to your failure and how "the staff you hire will have a big impact on the demise of your restaurant."  Of course, if all you do is to "make sure you hire people who are experienced, reliable, and passionate about food", then your failure is assured, and who knows - you might even end up in jail when the IRS comes knocking.  Consider this:  What if you hire a really passionate bartender who is especially passionate about bringing two or three cases of beer to work every night in the back of his van, which of course he picks up at the local Costo or Wal-Mart?  You show up for "work" the next day and it is like "Wow - we only sold one case last night", and you check the register vs. inventory and it all adds up - so sad that you are losing your shirt in this busy location!  Of course, your "passionate" bartender sold three additional cases to some of the "cash customers",  Did you get that on camera?  In the meantime - the IRS has been looking into your dumpsters, and they are coming after you for under-reporting.  Too bad.  So sad.

Now for whatever Google Bard suggests about chapter five "Not marketing your restaurant" - one theory says that "No matter how great your restaurant is -  if people don't know about it, they won't come", so you have to "make sure you have a solid marketing plan in place to reach your target audience."  This is of course completely wrong, at least most of the time. While some kind of marketing is probably essential, i.e., a website, a "yellow pages" ad, maybe some coupons in a local paper; etc., if you have a good location, some of your customers will just simply "find you", whether that location is in the heart of downtown, or right next to the "last chance for gas - 100 miles" place.  Right-sizing your advertising budget is important, of course, but there is something that is more important that we need to talk about, and that is "sustainability".

So let's say that you own your location somewhere along the lonesome highway and that as a family-run business, your overhead is pretty much zero.  If a customer wants a plate of hot wings you can get a bag of 24 of them for ten dollars, and you sell them six wings for twelve dollars.  Lather, rinse, repeat - now you get to keep your shirt for another day.  On the other hand, if you decide to buy an interest in a major franchise like Round Table Pizza. or Burger King, for example, then chances are that you are going to be spending a lot of money on franchise fees, so planning on doing your own advertising might not be something that you need to consider going right out the gate - unless of course, if some franchises offer some form of "dollar matching" or rebate program like if you decide that you want to spend some of your franchise fees doing local radio ad buys, and so on.  Maybe that sort of thing is more common in the automobile dealership business, i.e. when the factory offers to help the local dealership with radio-tv-print incentives.  Random fact: I used to work in Television!  This is one you NEED to know about!  Ask your franchise HQ about "co-ops", just in case they AREN"T telling you!

This brings us to Chapter 6: "Not managing your finances properly" I could go on and on about how the restaurant business is a very cash-intensive business, or the DJ business, or the wedding photography business, or whatever else you want to get yourself into, and while It is important to have a good understanding of your finances and to manage them carefully; there is something that we might want to reflect on that was discussed in an earlier project, you know the one about "The Money Bomb", and no I am not trying to download a trillion-trillion carets worth of diamonds here, or a teaspoon-full of dense neutron matter either.  No, that has nothing to do with what I am about to say - or actually it has everything to do with what I am about to say - because I want to say a few more things about "sustainability" 

Thus, while in chapter seven we might be admonished about the importance of "Giving up too easily" it is important to contemplate, based not just on the theory of "The Money Bomb", but on something that is much deeper, and that is that there are natural laws in the universe, like the Second Law of Thermodynamics, and a whole bunch of stuff that we could try to model with chaos theory and the like, even if we don't have to work things out analytically, with however many petaflops of tensor flow that predict-o-matic might seem to need to tell us what we think we already know - and that is that the restaurant business, like almost any business, is not for the faint of heart.  So remember, as it is sometimes said that a thousand-mile journey begins with a single step, so although there will be many challenges along the way, it is usually the best and most important approach to persevere and not give up easily.  Now that doesn't mean that at times, and there will be times that so-called "sunk costs" might need to be abandoned, like if you are drilling for oil and it turns out you find out the geologist's report was a fraud, or something similar - that is - in some other field of endeavors.

Now Google BARD provided just a brief outline, of course. There are many other things that could be included in a book on any topic. But hopefully, I have shown that even with a completely screwed-up outline, with a lot of wrong information, we can either provide correct information or at least provide you with a good starting point toward understanding how this sort of AI has the risk of getting us all killed.  All in good fun of course.  At least, Google BARD does have a disclaimer that says "experimental"

So if you made it this far, and if you are aware that I have been experimenting with creating training sets for my own AI, thus far, with my rants about compiler design, or else for trying to devise a scripting program that can accept ordinary English commands like "make deluxe pizza" which can be reformatted into something that looks like Pascal, or C++ or any of a number of languages as in

void do_something()
  make ("deluxe pizza");
  serve ("deluxe pizza");

O.K., then, looking at how we got here, maybe it might make sense to try to turn a document into a program, by analyzing the document for the presence of words like "make", "do", or "If", or "when", and so on.  Then one could search for repeating patterns in the narrative, and look at how to reconstruct the original document from a list of keywords and how they map into the recurring patterns, almost the way that Huffman codes work.  Now if that is ALL that the current bumper crop of LMMs is doing, then something is seriously missing.  Yet there is still something very wried lurking here, since just as it has been noted that "artificial general intelligence" seems to be everyone's goal, on the one hand, it remains the case that all of the contemporary models confabulate, or hallucinate, on the one hand, from time to time, while even the simplest models seem to offer some occasional moments of what I elsewhere described as seemingly "brilliant lucidity", like when Mega-Hal said something like "Somewhere in the sky - even for an AI".

Now obviously, I don't have a bot quite yet that can re-write its own source code, but it is headed in the direction, on the one hand, and thus, even if on the other hand, the theory says that an LMM just simply cannot do general intelligence because it is just random recombinant shuffling of patterns without true reflection, it might, on the other hand, achieve a more seemingly sentient presentation, that is - if the random recombinant techniques can be applied to generating more sophisticated world models, that also operate more like the more traditional Von-Neumann architectures that the neural networks are trying to replace.  So if in effect LMMs make coding easier, then when an AGI is achieved, then I suspect that, for efficiency reasons, there will need to be some kind of hybrid approach taken, so that most of the AGI will actually be implemented in some kind of traditional program, even if that program, in turn, turns out to be a sandboxed implementation of an AI that tries to do its best to execute what is traditionally regarded as "pseudo-code"

Well, for whatever it is worth. I just checked and found out this particular log entry is approaching 3000 words or so, even if it was at 2956 according to Microsoft Word - that is to say before I began this otherwise seemingly run-on sentence.  Not that there is a quota or anything, but if I were to write 3000 words every day for a year, well that obviously would add up quickly - like to just over one million words.  Or put another way, according to O.K. Google, there are just 783,137 words in the King James Bible, which would take just 216 hours to type at 60 words per minute, which doesn't seem all that hard at all.  This might have some significant implications, since if the design characteristics and overall quality of a training set have the greatest determining factor on the quality of the results that are obtained with an AI, then if it also follows that just as we are on the verge of reliably being able to convert standard English text into a more traditional programming language; then this will imply, well, whatever it might imply for the quest for AGI, which is an acronym of course, for "Artificial General Intelligence". 

The continuing theory is being promulgated, therefore, that if the human genome only has about 3 billion base pairs of DNA, which at two bits per base pair would require only 750 megabytes, uncompressed; then that takes us back t the question of just how big does the source code for an Artificial General Intelligence actually need to be, based on a bunch of other numbers, of course - like the actual number of coding genes in the genome, and the percentage of those that determine how the brain is wired.

Yet right now, there are a number of prevailing problems with the current approach to AI, that is, at large.  First and foremost, perhaps is the Intellectual Property issue; which I have tried to address by using the absolute bare minimum of third-party materials.  Then there is the issue of expense.  I read the other day that the burn rate for OpenAI right now is running something like $700,000 per day (!) and up to 36 cents per query!  So their GPU-based approach is horribly inefficient, just as I have suspected all along.  Then there is the issue of turning standard English into traditional source code or perhaps using standard English to create large language models which can create source code, and which, in theory, could allow the creation of neural networks that might be capable of turning tensor flow-based solutions to some of the otherwise hard-to-solve problems into a more traditional algorithmic approach.  That might result in a two or three-order-of-magnitude improvement in the cost of doing AI.  Whether this means that the technological singularity is also at hand is harder to fathom.

Time to create a new training set, that combines all of what has been said so far, and actually try it on a Propeller or Atmega 2560, I suppose.

Stay Tuned.