How much tech skills should I know?

16 replies
Hi,



I spent the last few years in school and am slowly planning to transition back to the private sector over the next few years. The more I'm learning, the more I'm thinking it would make sense to create a tech startup focusing on machine learning a few years down the road.

I have a math degree but I still am missing a lot of the engineering know-how.


The big question for me is : How much of that know-how do I have to fill? And how should I compensate for it?


There are a few strategies I could pursue:

a) Focus on making money regardless of in which field. Learn just enough code to judge technical competence. Throw money at the problem (Trial and error outsourcing).

b)Same as a) but learn just enough code to translate my ideas into a proof of concept or MVP. Sell a few units, prove that I have a viable product with a viable market

and THEN seek a technical co-founder.

c)Work toward a sales job in a large tech corporation and really work on building relationships with the programmers and the techs. After a few years, use those contacts and build that startup.

Thoughts?Which paths should I take?
#skills #tech
Avatar of Unregistered
  • Profile picture of the author palmtreelife
    Hey Social,

    I'll be honest...I felt like puking when reading the first two strategies, so my vote is option 3. Why?

    In both of the first 2 ideas, you say "learn just enough"....you're trying to cut corners and no one has any patience for cutting corners. Why would you want to learn "just enough" to get by?

    No matter what you're trying to do to earn money or make a career, you should be focused on learning as much as possible to become a recognized EXPERT in that field. That's when opportunities will arise and people will be coming to you for start ups.

    The way you're approaching this, I don't know how you will get any accomplished tech expert to join forces with you. What skill will you be adding to the team? A person who cuts corners can be seen a mile away.

    Option 3: sales skills will always be in demand. Great relationships cannot be replaced by any amount of money or education. People do business with those they LIKE.
    {{ DiscussionBoard.errors[11531299].message }}
    • Profile picture of the author socialentry
      Originally Posted by palmtreelife View Post

      Hey Social,

      I'll be honest...I felt like puking when reading the first two strategies, so my vote is option 3. Why?
      This is a surprising reaction...


      In both of the first 2 ideas, you say "learn just enough"....you're trying to cut corners and no one has any patience for cutting corners. Why would you want to learn "just enough" to get by?
      Because there's only a finite amount of time.

      I doubt becoming an expert programmer on my spare time (or in 2 or 3 years) is a viable option given that others have been sharpening their craft since 12 or 13 years of age.

      Programming is obviously a crucial part of tech, but increasingly it's becoming a commodity. Not only this,but it's a highly perishable skill so in my case it's best not to invest more time then strictly nescessary. I'm of course open to contrary opinion but the common advice I got was to do something else then focus on programming or pure tech.

      What skill will you be adding to the team?
      I spent several years in sales and I have a pure math degree. To translate industry problems into math terms and vice versa is really difficult because you have to have broad knowledge of both world so I think it's not a bad position to be in.

      OK now I have one question for you: did you pluck your advice out of a self-help book by any chance?
      {{ DiscussionBoard.errors[11531322].message }}
      • Profile picture of the author palmtreelife
        I was being honest, but I could have used different words to describe my reaction. I didn't mean to offend or hurt you. I always try to provide advice instead of criticize or fight like what happens in so many threads. You hit a nerve I guess, I'm sorry.

        I didn't anticipate you getting defensive though. This isn't self help book advice. This is common sense life advice.

        Put yourself in your potential partner's shoes. Would you partner with yourself if you knew corners were cut and the bare minimum work was put into the developing the skills? If someone cuts corners in developing a skill, they probably cut corners in other areas of their profession as well.

        Because there's only a finite amount of time.
        Everyone has the same amount of time? Yes, others started earlier than you, but the amount of free or inexpensive information available to us today for any particular skill dramatically reduces the time required to become an expert in any particular field.

        I spent several years in sales and I have a pure math degree. To translate industry problems into math terms and vice versa is really difficult because you have to have broad knowledge of both world so I think it's not a bad position to be in.
        ...which further justifies why I think you'll do better with option 3.
        {{ DiscussionBoard.errors[11531643].message }}
      • Profile picture of the author springlee
        I appreciate your answer, thanks a lot!
        {{ DiscussionBoard.errors[11532045].message }}
  • What seems to me, the most viable option for you would be a combination of option 2 and 3, if you have the time for it.

    You'll need to do 3 to get the money to make your MVP and network with potential partners.

    At the same time, you'll need to do option 2 so you can test the viability and the market for your product. If you have enough technical know-how to build the basic features, it would also serve as the jumping point for your future technical partner to create the full product.

    The problem here is you'll need to do most of the bootstrapping yourself, at least in the beginning.

    But having an MVP ready before your bring in a technical partner would help define ownership of your product. If you bring a technical partner early, there's a risk your idea might be lost in the process of development.
    {{ DiscussionBoard.errors[11531402].message }}
  • Hellor Socialentry,

    I research and write for the Technology Space and all too familiar with the industry. My clients range from Tech start-ups, SaaS, IaaS, PaaS providers, IT and MSP companies, RFID tracking for the shipping, freight, and airline industries, as well as Tech Target the online technology news site.

    You have some decisions to make, that is clear. As I read your three options, I sense you're still at the crossroads. What I did not see is your well defined, burning deep down in your soul, "I've Got To Solve This...!" position in your post.

    So, let me ask you, why do we need yet another tech start-up, machine learning company? They're everywhere these days. How will your company fit in with digital transformation? What do you currently see in the marketplace that is not getting fixed from current technologies that your machine learning business can solve immediately?

    Chinchilla
    {{ DiscussionBoard.errors[11531531].message }}
    • Profile picture of the author socialentry
      Originally Posted by ThePromotionalGuy View Post

      Hellor Socialentry,

      I research and write for the Technology Space and all too familiar with the industry. My clients range from Tech start-ups, SaaS, IaaS, PaaS providers, IT and MSP companies, RFID tracking for the shipping, freight, and airline industries, as well as Tech Target the online technology news site.

      You have some decisions to make, that is clear. As I read your three options, I sense you're still at the crossroads. What I did not see is your well defined, burning deep down in your soul, "I've Got To Solve This...!" position in your post.

      So, let me ask you, why do we need yet another tech start-up, machine learning company? They're everywhere these days. How will your company fit in with digital transformation? What do you currently see in the marketplace that is not getting fixed from current technologies that your machine learning business can solve immediately?

      Chinchilla

      You're correct. I don't have much more then a general direction right now. I don't even have an idea which industry in particular I want to apply ML to.

      Even so, I have to take a decision pretty soon. My decision is and is based on this reasoning:

      A professor explained it to me as such: ML is like gold panning. If the data is virgin territory, Claude can find nuggets of gold with a plastic shovel, but when the data has already been mined, the chips are down and the clock is ticking, it is mathematical depth that makes the difference.

      But it is surprising how many people enter ML with relatively little math baggage and I think most don't really realize it because with limited tools, you can still extract something of economic value.


      Paul Graham wrote this interesting article:Beating the Averages
      I will use a quote from him to better explain my thoughts:

      Originally Posted by Paul Graham

      It must have seemed to our competitors that we had some kind of secret weapon-- that we were decoding their Enigma traffic or something. In fact we did have a secret weapon, but it was simpler than they realized. No one was leaking news of their features to us. We were just able to develop software faster than anyone thought possible...

      Our secret weapon was similar. We wrote our software in a weird AI language, with a bizarre syntax full of parentheses. For years it had annoyed me to hear Lisp described that way. But now it worked to our advantage. In business, there is nothing more valuable than a technical advantage your competitors don't understand. In business, as in war, surprise is worth as much as force...

      ...if Lisp is so great, why doesn't everyone use it? [it] sound like rhetorical questions, but actually they have straightforward answers. ... And the reason everyone doesn't use it is that programming languages are not merely technologies, but habits of mind as well, and nothing changes slower.


      The article is from 2001 is about Lisp but the moral really is about old ways of thinking dying hard.


      I am confident that someplace somewhere, if I persevere, I will find an industry that describes this exact situation that Graham is describing.
      {{ DiscussionBoard.errors[11532456].message }}
      • Profile picture of the author savidge4
        Originally Posted by socialentry View Post

        I am confident that someplace somewhere, if I persevere, I will find an industry that describes this exact situation that Graham is describing.
        A little advice from the outside in... As much as these startups are "business" generally speaking structure is a bit different. Your average joe blo can have the drive and ambition and hopefully some amount of capital and start a business.

        Tech start ups tend to be a bit different. there are elements that need to be in place. I would say that your general tech start needs 3 elements. An office manager type that has an understanding of whats going on and can be the middle man to direct and redirect the other parties. #2 would be tech, and generally speaking we are talking about programmers etc. The third element is someone on the inside of the Niche you are leaning towards. If you were creating a medical start up a doctor on the team goes a long way. If it was some kind of Financial thing, and accountant or the like would be nice.

        so how this plays out... The manager and the insider get together and discuss need and demand and all of that and determine a path. The path is then translated to something that can be programmed. we need this this and this and want this to look like this and this. The programmer then creates "The Vision" with the guidance of the manager. Once things progress along the insider is brought back to assess where the project is, what needs to be fixed and what is working.. and this goes back and forth until there is a final product. and this would be a structure that works.

        But now we are talking AI and things change a bit. Yes you need a manager - and hopefully there is some amount of sales ability here, you need the programmer, and you need the insider - again hoping there is some amount of sales ability here. BUT now you need to introduce a 4th element and that is the math.

        Programmers by default are usually given X and fill in A to W to get there, but in AI it is reversed. here is A to W, and we need X please, and that is where the math knowledge comes into place. A classic Form follows function example.

        I have been "witness" to a number of AI sales pitches in recent years. They tend to operate in teams of 3 or so to make the presentation ( we are talking big pitches here ) The guy with the general knowledge, the guy with the specific knowledge and then the Math guy. In all instances the General knowledge guy would answer any and every question he could and refer to either the Insider or the Math guy specifically - even if the question is directed at either the insider or the math guy.. the general knowledge guy would respond and refer the question to the appropriate party. ( and I have not seen this once, I have seen this 5 6 10 times ) and I will tell you I would think it would take some practice LOL

        Right now.. Data simply is growing on trees to be picked... There are literally Millions apon Billions of dollars being made Off of IBM's Watson platform alone as we speak.

        YOU have to find a pain, and solve it - YOU have to find yourself an insider. YOU can be a manager, and YOU could be the math guy - and then bring in a programmer and put it all together. Then it becomes a matter of rinse and repeat.

        So I mentioned above it helps if the manager and the insider can sell... its painfully obvious in a pitch meeting when its not the case. Selling AI has to happen on 2 levels The plain dumbed down version, and the more technical same language type pitch only an insider can deliver. Bringing the math guy along is simply a way to introduce a PHD into the mix lol

        so that's kinda sorta all I got on the subject, hope it helps! and best of luck man, you deserve it!

        Added...

        and Lisp... yes, makes you a better programmer for the rest of your days hahaha
        Signature
        Success is an ACT not an idea
        {{ DiscussionBoard.errors[11532553].message }}
      • Hellor Socialentry,

        savidge4 couldn't have said it any better.

        I only have two recommendations.

        1) Go over to Darktrace's website and comb through their entire site and knowledge base. They are the leading AI company in the world. Cisco can't even catch up to them. Here is their link: https://www.darktrace.com/en/

        Then go to the top of their page and click on "Industries." They cover the top 12 industries in the world.

        Once you find an industry that appeals to you, then dig down into that industry, and I mean really, really deep.

        The thing about Technology, as it evolves, we find gaps not getting filled in every industry all the time. It's almost as if new or updated tech opens the door for more unique problems to get solved.

        2) Find a starving crowd that is desperately looking for what your ML business will offer.

        Also check out Robert Coorey's SaaS Podcast: How to find a starving crowd hungry for your SaaS product.

        Here's the link: https://saasclub.io/podcast/robert-c...tarving-crowd/

        If you have no one looking for your product, service, or data, your business will subsequently fail.

        Chinchilla
        {{ DiscussionBoard.errors[11532740].message }}
  • Profile picture of the author Jason Kanigan
    I don't understand why people don't come up with a simple, structured approach to solving problems.

    I do what I'm about to explain all the time:

    With all the free training on YT, I simply go there and try a few search terms. When I get it right, I watch the first three to five videos that come up. I may not watch them in full.

    I also google the term. I read the first few articles that come up.

    Guess what? I've invested maybe a couple hours and now I know more than 99% of the population.

    And when I need more specific info?

    I hire an expert and interview them!

    It costs $100, maybe $200 (most are delighted to receive any money in exchange for their time) and I get my specific questions, that I now know how to ask because of the basic immersion self-training I did, answered. And I'm not wandering around wasting my interviewee's time.


    What I think the biggest problem for people is...is seeing the big, imaginary block of work that must be done to "git good" at the thing. Just get started. Go learn it. Get immersed in it. You'll find there's a current to the river, and it'll take you along. Within quite a short time you'll have a better perspective than you have today, and you'll be able to ask better questions because of it.

    The option is to get hung up on the "OMG This Is Going To Be SO DIFFICULT" thought and fear from your imagination, and do nothing. Then the day goes by and you accomplish nothing.
    {{ DiscussionBoard.errors[11531544].message }}
    • Profile picture of the author palmtreelife
      It costs $100, maybe $200 (most are delighted to receive any money in exchange for their time) and I get my specific questions, that I now know how to ask because of the basic immersion self-training I did, answered. And I'm not wandering around wasting my interviewee's time.
      I like this a lot. It's my favorite thing to do when learning something new. I learned the most about investing in real estate from paying someone documented in the field to just answer my questions...."explain this and explain that...how does this work step by step...how to negotiate deals, build a team etc"

      I learned more information in a very short period of time than I ever did in any course.

      Like you said, most people are more than willing to receive money to share their tips and lessons. It's surprising more people don't do this...or is it? People are lazy...nevermind
      {{ DiscussionBoard.errors[11531995].message }}
  • Profile picture of the author ewenmack
    Machine learning is just a way to get things done.

    Like Yellow Pages is a way to get things done
    ie get your ad in front of an audience.

    Therefore in of itself, it has little value
    until you bring a solution to a problem.

    Here are some examples you can use to narrow your focus onto problems
    that can be solved for a group of people
    using machine learning...

    Amex wanted their advertisers to de-risk their advertising experience.

    The solution was inside their transaction database.

    Set up predictive algorithms as to what people
    are about to buy in different products and services.

    Now they have it in over 2,000 categories.

    H&R Block wanted to sp[eed up and with more consistent accuracy
    the tax deduction process internally.

    They tapped into IBM Watson and the tax code
    to make the machine come back with the answers.

    Problem solved.

    Side note: You could use that tax code data inside IBM Watson and create your own
    app for business users to do away with the cost of companies like H&R Block.

    IBM makes it really easy to tap into Watson.

    Uber wanted to make ordering a taxi easy and much more reliable.

    They created an easy to use interface which consisted of driver id,
    GPS and a payment gateway.

    All apps that existed but not combined for the taxi industry before.

    Look for problems to solve on a daily basis.

    The technology, people and money are all available when a good idea is ready.

    Best,
    Ewen
    {{ DiscussionBoard.errors[11531621].message }}
  • Profile picture of the author savidge4
    An interesting question with some interesting answers.

    I am going to suggest D. Based on my own personal experiences with start-ups, they are basically relationships. Pretty much no different than dating and ultimately marriage, you need to get out and mingle and start dating.

    I might suggest staying at University part time ( a class or 2 ) and start taking programming classes. You want to start relationships with programmers, and that for you is probably the path of least resistance in doing so. Again no different than dating. You decide you want to date a "nice" girl, adult bible study groups would be an option... or you simply want to shag.. you goto the local bar or dance club.

    Really no different than anything else we do, you need to be in the right place to be in front of the people you need, be it to date, or to employ, to sell to, or to start a partnership in a start-up.

    So lets migrate to the technical side of the scenario... Data in recent years has gotten interesting. It has fallen into 2 primary categories those that understand it and those that don't. It then continues the split, to those that have it and those that don't. A quick look out on the main board of this forum, and you can see these splits, primarily those that don't understand.

    So you are left with... How can I sell my knowledge of data? There are basically 2 paths. You can create an Algo that reaches the masses with a set data pool and does what seems to be complicated math and produces an answer. or you can sell efficiency as a service and gather / collect data from a business and through AI show how they can improve their business and profits - IE Consulting

    My 2 Cents anyways
    Signature
    Success is an ACT not an idea
    {{ DiscussionBoard.errors[11531882].message }}
  • Profile picture of the author Medon
    My friend, I am surprised you found your way here, Perhaps it is your passion or instinct that brought you to here. Anyway, I just want to think that you are here because of a reason. It is the best home I have found on the internet. Ask any question and we will help you catch up.
    {{ DiscussionBoard.errors[11532491].message }}
  • Profile picture of the author ewenmack
    Here's another example of multiple unstructured data
    been brought together
    to make it more valuable and usable,
    https://trestle.corelogic.com/Home/Providers

    It's like a supermarket and you can pick and choose.
    . many data products to make your own unique app
    that serves your core customers.

    They make it so easy by setting up contracts with
    all real estate data providers
    inside Core Logic's platform.



    IBM is the granddaddy of them all.
    {{ DiscussionBoard.errors[11534256].message }}
  • Profile picture of the author cearionmarie
    If you already have an above-average knowledge over something then that should be a start. However, you will still to continue learning to be able to contribute more knowledge. There is no limit to learning.
    Signature

    Cearion Uy - Marketing Advisor
    www.influencerauditor.com

    {{ DiscussionBoard.errors[11535643].message }}
Avatar of Unregistered

Trending Topics