Upskilling and Reskilling for Artificial Intelligence  ## Metadata - Author: [[Ray Villalobos]] - Full Title: Upskilling and Reskilling for Artificial Intelligence - Category: #articles - Document Note: --- tags: [#AI, #Upskilling, #Reskilling, #Python, #MachineLearning, #CareerDevelopment, #OpenSource] - #AI - #Upskilling - #CareerDevelopment Author: Matt Harrison (Interviewed by Ray Villalobos) Source: YouTube – *Upskilling and Reskilling for Artificial Intelligence* Reference: [YouTube Transcript Link](https://www.youtube.com/watch?v=otnADHXH-KM) date created: 2025-08-22 19:30 --- # Title: Upskilling and Reskilling for Artificial Intelligence – Insights from Matt Harrison ## Table of Contents: - [📖 References](#-references) - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [🗣️ Paraphrase](#-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-diversity-equity-inclusion-and-belonging-assessment) - [✅ Logical Validity](#-logical-validity) - [🔍 DARVO Analysis](#-darvo-analysis) - [🔍 SEO Evaluation](#-seo-evaluation) - [🎯 Conclusion](#-conclusion) --- ## 📖 References - Harrison, Matt. *Effective Pandas* - LinkedIn Learning Courses (Finance & Python) - NumPy, Pandas, Scikit-learn libraries - Python Package Index (PyPI) --- ## 📝 Summary Matt Harrison, a Python expert, trainer, and author, outlines how developers can stay relevant in the age of AI. He emphasizes Python’s role as the “second best tool for everything,” the importance of continuous learning through teaching and projects, and the need to understand both the power and limitations of AI tools. Developers who embrace AI for brainstorming, documentation, testing, and productivity will thrive—those who ignore it risk obsolescence. --- ## 📖 Main Content The interview blends career advice with practical coding insights. Harrison’s perspective is shaped by decades in Python, natural language processing, and corporate training. - **Career Transition**: From early NLP work to tutoring, speaking at conferences, writing books, and corporate training. - **Python’s Strengths**: Accessibility, ecosystem size, and NumPy’s impact on performance. - **AI Integration**: ChatGPT and Copilot as productivity amplifiers, not replacements. - **Learning Approach**: Motivation matters—choose projects relevant to personal or work interests. - **Ethical AI Use**: Be mindful of hallucinations, data privacy, and unintended consequences. - **Future Outlook**: Developers must leverage AI tools while focusing on uniquely human skills like teaching, empathy, and contextual problem-solving. --- ### 🗒️ Notes - [00:03](https://www.youtube.com/watch?v=otnADHXH-KM&t=3) AI reshaping developer skillsets; Python central. - [02:21](https://www.youtube.com/watch?v=otnADHXH-KM&t=141) Transition into training; importance of teaching as learning. - [07:19](https://www.youtube.com/watch?v=otnADHXH-KM&t=439) Good teaching = empathy + clarity. - [14:12](https://www.youtube.com/watch?v=otnADHXH-KM&t=852) Ethical concerns: IP leakage, model training via user input. - [16:32](https://www.youtube.com/watch?v=otnADHXH-KM&t=992) NumPy unlocks C-like performance in Python. - [22:48](https://www.youtube.com/watch?v=otnADHXH-KM&t=1368) Motivation drives learning success; like fitness training. - [28:22](https://www.youtube.com/watch?v=otnADHXH-KM&t=1702) AI boosts expert productivity; beginners must still learn fundamentals. - [34:25](https://www.youtube.com/watch?v=otnADHXH-KM&t=2065) AI aids code commenting, testing, typing. - [40:43](https://www.youtube.com/watch?v=otnADHXH-KM&t=2443) Future-proofing careers = embrace AI while honing human skills. - [46:57](https://www.youtube.com/watch?v=otnADHXH-KM&t=2817) Coding interviews may need reinvention as AI aces LeetCode-style tests. - [52:45](https://www.youtube.com/watch?v=otnADHXH-KM&t=3165) Python in finance; practical courses bridging Excel to pandas. - [56:56](https://www.youtube.com/watch?v=otnADHXH-KM&t=3416) AI supports job search (cover letters, resumes, brainstorming). --- ## 🗣️ Paraphrase 1. Original: *“Python is the second-best tool for a lot of things.”* **Paraphrase**: Python may not always be the perfect choice, but it is consistently a strong option across domains. 2. Original: *“If you don’t try AI tools, you risk being replaced by those who do.”* **Paraphrase**: Avoiding AI experimentation is career risk; those who adopt it gain the edge. --- ## 📰 Newsletter 🚀 **Upskilling for AI: Python, Projects, and Practical Ethics** Python trainer Matt Harrison joined Ray Villalobos to unpack how developers can future-proof their careers in the AI era. Key takeaways: - 🐍 Python remains the universal “second-best tool”—versatile, supported by NumPy, pandas, and 400k+ libraries. - 🧑‍🏫 Teaching accelerates mastery; explain concepts to truly learn them. - 🤖 AI tools like ChatGPT supercharge productivity but require expert oversight to avoid errors and hallucinations. - 🔐 Handle AI ethically: beware of IP leaks, biases, and careless data use. - 📈 Future interviews may shift beyond LeetCode tests as AI trivializes standard coding challenges. Catch the full talk here → [Upskilling and Reskilling for AI](https://www.youtube.com/watch?v=otnADHXH-KM) --- ## ✅ Action Items - Build WordPress contributor workshops highlighting **AI + Python learning paths**. - Position OSPO training as **career accelerators** for reskilling contributors. - Publish contributor-facing guides on **AI ethics and data privacy**. - Integrate AI-assisted **documentation/testing exercises** in onboarding. - Advocate for teaching-focused mentorship that mirrors Harrison’s student-first perspective. --- ## 🧪 Hypotheses 1. AI-augmented developers outperform peers by 30–50% in project delivery speed. 2. Developers with teaching responsibilities retain knowledge longer than those without. 3. Ethical AI literacy directly improves adoption confidence among organizations. ### 📊 Methodology - Controlled comparisons between AI-using and non-AI-using dev teams. - Surveys tracking contributor knowledge retention after teaching. - Case studies of companies with ethical AI policies. ### ❓ Questions - Should basic AI/ML concepts be mandatory in high school curricula? - How can OSPOs embed AI literacy without overwhelming contributors? - What future replaces coding interviews if LeetCode is obsolete? ### 🔮 Future Research - Longitudinal studies on AI adoption in open source contributor pipelines. - Examination of cultural differences in AI acceptance. ### 💡 Implications - Reskilling strategies could reshape contributor pathways. - AI ethics training might become core OSPO responsibility. --- ## 🧠 Learning Styles Assessment - **Visual**: Few diagrams → add infographics for NumPy & AI workflows. - **Aural**: Video interview supports auditory learners. - **Verbal**: Rich analogies (e.g., Python as “second-best tool”). - **Kinesthetic**: Weak → add coding labs, exercises. - **Logical**: Strong frameworks (NumPy, motivation hacks). - **Social**: Meetups, training sessions. - **Solitary**: Encourages reflection via projects. - **Naturalistic**: Missing → could link AI energy use to sustainability. --- ## 📊 Accessibility Assessment - Transcript provided = ✅ - Needs alt text for visuals = ❌ - Navigation: Long transcript blocks may overwhelm screen readers. - Improve by breaking into smaller segments, ensuring keyboard navigation. --- ## 🌍 DEIB Assessment - **Strengths**: Inclusive teaching tone, international audience participation. - **Weaknesses**: Few diverse case studies; AI ethics mentioned but not tied to DEIB. - **Improvements**: - Showcase global open-source stories. - Include neurodiversity-conscious teaching strategies. - Add culturally diverse coding examples. --- ## ✅ Logical Validity - **Strong Deduction**: NumPy → C-like speed in Python = valid cause/effect. - Summary: Times of change are always full of opportunities for those who are ready, and right now there is a great shift happening in the technology industry towards Artificial Intelligence. The industry is looking for those with the right experience, so let’s talk to an Artificial Intelligence expert Matt Harrison, who helps teams and individuals level up Python, Data Science and Development skills. Matt is the Co-Chair of the Utah Python User group and a speaker and presenter at conferences like PyCon, OSCON, Strata, and more. He’s trained companies like Hewlett Packard, Adobe, Cisco, Samsung and others in everything from drone, python, web development and e-book production. Index ====== 00:00 Intro 00:57 Matt's Developer Journey 02:37 Transition to Corporate Training 04:47 How Matt Stays up to date 08:56 What he learned During Covid 11:43 What Should Developers Learn from GPT and AI? 16:04 Why Python is so popular 21:36 What projects should people build to get started 28:15 How Matt uses GPT and Copilot 40:46 Effect ... - URL: https://www.youtube.com/watch?v=otnADHXH-KM&ab_channel=RayVillalobos with most of Technology focused onartificial intelligence you might bewondering if it's time to take what youknow as a developer and level up to stayrelevant so I've got just the expert forthis today I'm talking to Matt Harrisona sophomore engineer corporate trainerand expert on python he's written anumber of books on machine learningpandas and even something called XGboost I have no idea now he's a popularspeaker at conferences like pycon osconand has trained companies like HP Adobe Samsung and others Matt you have a lotof experience with python but I alsonoticed that you work with tools likereact native Django Docker and lots ofother stuff can you talk a little bitabout your developer journey and maybethe state of some of the opportunitiesfor the different programming languagesin this age where everybody's justreally talking about AIsure yeah so my background is I have acomputer science degree andum out of school I I worked kind of as a data scientist but that was back in 2000before data science as a name reallyexisted but I was doing natural languageprocessing for a small startup thatcompeted with Google andum and then over the years I'veI I did some back-end development workum at that first job I I came acrossPython and I kind of fell in love withthe language at the time I was more of apearl person and I was tasked to work with a person who used the ticklelanguage TCL and neither one of uswanted to use the other tool and sopython was a compromise and I've it sortof fit with my brain and I've I've stuckwith it ever since and it's I thinkpython is in a great place right now youknow is it is it the one best tool foreverything I like to say that pythonwell a lot of people say this python isthe second best tool for a lot of thingsand and so because it fits in a lot ofRealms it happens to be quite useful that's pretty awesome so talk a littlebit about maybe uh your transition frombeing a developer to sort of a corporatetrainer uh or getting into that I knowthat for me it's just a personalchallenge to stay current as somebodywho trains for a living to stay currentwith everything that's going on so howdo you manage like staying up to dateyeah so so let me talk about mytransition first and then I'll talkabout uh staying up to date um souh as I was working in various companiesI I was like I want to be the I don'twant to be the person in the chair atthe conference talk I want to be theperson giving the conference talk andthat I guess that's just myum personality and and so I I startedum doing talks I started doing a bunchof training atum conferences giving tutorials that sort of thing and that led to my firstbook I I gave a tutorial and I and I wasaccepted to give that tutorial again andI was like well instead of justrewriting my slides I'll write a bookum and so uhthat book happened and that that was atthe time where self-publishing was justuh starting to happen that was in 2011.andum you know I think there's just aConfluence of like I was being at a lot of python conferences writing booksabout it and teaching and so therestarted to be an influx of interest inpeople wanting training and that's how Igot into trainingumso I've always liked helping people likeI paid for college doing tutoring andand my mom is a teacher so I I reallyfeeluh a lot of satisfaction like seeing people learn how to do things in betterways or or helping them advance and andso that was my transition to corporatetrainingum you know along the way I have done alot of Consulting as well and and sothat Consulting can keep you uh uhsomewhat up to date or you canspecialize in certain things but stayingup to date is a challenge I don't thinkthis is a challenge that's unique tocorporate trainers I think this is a challenge for everyone right uh you knowI like you know as a corporate trainerI'm looking at education how to teachpeople sort of the latest and greatestbut you know I saw in in likeuniversities there are a lot ofuniversities are like oh we need tostart teaching Python and they will justtake like their Java their C plusmaterial and like change some things soit's python-like but they're not reallyteaching python per se they're teachingyou know some weird version of python as if it were a job or C plus plus as if itwere python so you know finding someonewho really wants to dig into to you knowthe the bits of stuff can be a challengeum souhyou know there is a balance what what doI want to focus on what do I kind ofwant to ignoreum as I said I did do a lot of back andstuff like I've done Django I've done some react react native stuffum and I've sort of let that slide likeI I don't know that my brain is bigenough to like stay on top of uh sort ofthe back end uh web development typestuff and do more of the data sciencymachine learning type stuff thatinterests meumI think a lot of people need to uh findfind what motivates them and there there are some tricks and and and things thatyou can do to to find what motivatesthemum I think one tool that that's beenuseful for me is writing books and andcreating content right andumyou know sharing that content withothers is a great way to learn ityourself but also uh there's a certain Iguessattitude online where where you know if something is incorrect people like tocorrect that and so you will becorrected uh pretty quickly if you startsharing things that aren't correctonlineum so uh yeah I I think my my mechanismfor staying up to date is uh a lot ofresearch and then just sharing that uhcontent with othersyeah I can certainly relate I think thatum I always say that you don't reallyknow something until you try to at least explain it to someone else and that sortof um I mean I think that's the theteaching and the coding are two separateskillsum I feel like most people I I've goneto a lot of talks where the the theperson who was given the talk is veryknowledgeable and especially when Iunderstand what they're talking aboutit's difficult to understand it in theway that they are presenting it so I think that like actually knowing thisstuff and being able to teach the stuffdon't alwaysgo together or they're not the samething and so it's really great um youknow I think that passion of trying tohelp others is really something that isgoing to help you I don't know become abetter better at what you do so I thinkthat looking for opportunities and youcan get started by just you know there'susually lots of meetups that people doeverybody's always looking forum you know people to give talks that's a greatopportunity for anybody who's trying tolearn things make yourself explain itand even you know as somebody who's youknow has a ton of courses I think it'sover 100 courses that I've doneum I feel like when I make myself teachsomething that's really when I know itdeeply because you know as you said youyou the worst thing that can happen isthat you teach something that's wrongpeople will definitely catch you catchyou on it so you know having thatdedication to really research it is really what makes a great teacher andalso I think in a way remembering whatit's like not to know something I thinkthat that's something that I thinksometimes is missing that yeah youforget what it was like when they didn'tknow something I don't know if you'verun into those kind of situations yeah II mean as someone who like especiallyduring covet I started geeking out a loton like the cognitive side of learningbecause I had to do a lot of you knowgoing from like in person to a remote I wanted to like double down on that butbut I think there is aa skill to teaching where it's not justif you know something a lot of peopleare very knowledgeable on something butthey don't really come to it from astudent point of view and so they canglaze over things that people don'tunderstand or just make assumptions thatyou knowthe student students they'll be completely lost very quickly because ofof assumptions that someone who's notputting themselves into into the foot ofa student will have so that is somethingthat that I think is you know thatsaying Teach it to a five-year-old youreally do need to kind ofnot just gloss over certain aspects butmake sure that uh youum that's a great tragicyou need to know your audienceyeah I'll say that yeah make sure you know yeah and I think that I had oneperson one time tell me uh don't youwant like the person that knowssomething the best to teach it to youand I said I think you want the personthat is going to help you understand itthe best like that's the best kind ofperson to learn from somebody that knowsit but they understand how you need tobe able to consume it and so that um youknow that's I think really important soa lot of uh people checking uh in fromall over the place I just want to acknowledge some folks here so there'sMarika from Boston somebody from Londonall over the place so this is prettycool let's see umwe are somebody somebody calling in fromDubai so Ed Bend and Victoria NorthCarolina India Romania we are gettingaround here all over the world so and Iknow you're in Utah and I'm in Floridaso and uh there's more people droppingin from lots of places so and you knowVictoria said something nice that she completely agrees I learned a lot uh youknow training teams and having a littlebit of input is priceless yeah you knowif you really want to know somethingexplain it to somebody else teach itwrite it write a blog post uh you knowdo something that helps you have toprocess it in a special way so we havemore people dropping in from Germany NewYork City Egypt wow everywhere so that'sgreat it's good to see I appreciate it appreciate the attendance so this iscool Serbia so let's keep going what doyou think are talking about DBT we kindof got a little bit started with it sowhat do you think our some of the mostimportant things that a developer shouldlearn today about GPT and generative AIyeah I thinkumwell I I mean this just just apply Ilike I T I'm doing a machine learningcourse for one of my clients this weekandI think that everyone needs to be able to uh write code to do machine learningno I don't but I think it it is usefullike I wish that people in school likein high school were required to takelike they were cried to take math Ithink they should be required to take atleast a basic programming Class Astatistics class and I think a machinelearning class at a basic level wouldactually be really useful because theywould understand the world that theylive in right I mean what when you go toYouTube or whatever they are basically running the program that's feeding thealgorithm and and like it or not we'reliving in that world so I think a basicunderstanding of how these things workum you know how that phone works infront of youum is usefulumyou know I also think people need to beaware of like where the data is comingfrom and what biases are in the dataum so you know for you know if you're using like GPT like at least you knowthat like cert you don't haveinformation after a certain point intime at least right nowum andthat can impact the answers that it saysanother thing that's important to beaware about you know like is that howthese things work but also like they canhallucinate right and and the mechanismuh for delivering a result is basically like we're aggregating all thisinformation that we've read to thatwe've read and now we're trying to likebasically predict what comes next andthat prediction might look good and readgood and sound very confident but mightbe completely incorrect so I thinkthat's a place where a lot of people canget stuck or is especially if you if youdon't have that expertise of what'sgoing on that can lead you down aum a a slippery slope and and then goingback to your other point like this thisstuff is changing so quickly right thatit's like you've got to bestaying on top of that is a challenge inand of itselfyeah totally and yeah I like that you domention people have to sort ofunderstand what the issues are andperhaps make sure that they know how totalk to the GPT and understand that it'sreally I mean they aptly named it Chad GPT because it is a conversation thatyou're having one of the things thathappened recently is this issue withsome Samsung workers accidentallyleaking Trade Secrets via Chad GPT thatwas kind of something that's been in thenews lately uh and just to understandthat the conversations that you'rehaving are also training are used to beused to train the models and that meansthat you know you're basicallypotentially leaking some internalcompany IP if you know you don'tunderstand that it you know that there is somesharing of information happening with avery experimental type of model thatthat's why you need to start thinkingaboutyou know ethical issues uh you knowstart using it of course because it isreally useful but understand thepotential issues I think um that'sprobably something that people don'tthink about they they see the new thinguh and and you know there's other typesof generative AI out there like mid-journey so understand that you're ifyou're uploading photos that you'resharing that data with everyone and thatdata can Surface in certain places lateron because you're using you're basicallynot just interacting with the model butyou're also training the model and thoseare really important so a lot of thestuff that's been happening it's reallybeen coalescing over languages likePython and the AI industry as a pythonexpert and professional why do you think python is so I don't know it seems to bethe first thing that the open AI API youknow was you know has examples for whywhy is it Python and not somethingdifferent like rust or C or anythingelse yeah yeahum I I would say you know in a word myanswer to that would be a library callednumpy andyeah let me sort of explain thatum python is a slow language so so how the object model Works in Pythoneverything is an object there's a lot ofoverhead to thatum and that makes it pretty easy to usebut it also makes it relatively slow andso this numpy Library which came out inthe mid 2000s is basically saying hey wewant to use Python to do numericalcalculations but python is a language isslow so how can we get around that wellwe can use python as an abstractionaround a buffer of memory and instead of having a bunch of python numbers we'lljust say we have you know 10 000 8-bitintegers or 64-bit integers and usenumpy as a way to take advantage of likemodern CPU architectures and simdinstructions and then when we do mathoperations on that numpy array if we'reusing the right abstractions we'rebasically approaching C speeds and Cmemory usage with that and so numpy allows you tobasically get very close to C withpythonand then there's been a whole bunch ofinfrastructure or whatnot built on topof that so things like pandasscikit-learnEtc built on top of that such that a lotof people were like oh we we now havethis language that's relatively easy toapproach uh Python and we have librariesthat are relatively easy to use and we can use them writing python code wedon't have to write C code or Fortrancodebut we we basically get a pretty goodspeed and memory usage from doing thatwe don't we don't pay that pythonpenalty too much but by using that andso that that's I think put us in a placewhere we are today where where python isreally popular among data scientistsdata analysts people who are doing dataI mean Python's also popular like I saidit's kind of the second best languagefor everything you can do data science like things like uh Rum you can do web applications or backin with you know JavaScript or Java youcan also do that with python you cancontrol uh websites and using tools likebash or Ruby but you can also do thatwith python as well so Python's kind ofin The Sweet Spot where it does a lot ofthings that developers and Technicalpeople and even people who who aren't developers per se it enables them to doand at this point in time python like ifyou go to the python package indexthat's peypi.org you'll see that there's that'show I check there's over 400 000different packages available for pythonso literally uh yeah right there andthat number wow the lower left yeah youcan see it says like 400 yeah rightthereum 400 000 packages plus that you can you can leverage right so these you knowsome of these are academic some of theseare toy projects but a lot of these areare packages that are being used inindustry for uh Cutting Edge stuff soyeah it kind of makes sense a lot of thedata scientists are using python uh foras sort of the lingua Franca of data andit makes sense to just sort of stay inthat environment uh rather than toswitch to different languages if you don't have toso that's actually a fantastic secondlanguage for somebody to learn I agreewith you I I haven't used it clearlyanywhere nearly as much as you have butI find it very approachable I'm not abig fan of the indentation thing thatalways kind of frustrates me but youknow you get over it and if you usesomething else like you know I've doneshell scripting uh that language is socrazy that I'm finding myself use it for the type of things I would do for Shellscripting because it seems to be alwaysaccessible and available everywhere uhPython and um it's it's super easy topick up it makes sense it's really solidit's it's probably like you said I lovethat analogy if it's the second bestlanguage for for anything I know that Icome from a newspaper background so Iknow that Django is like a really greatplatform for building like web typethings um and um yeah it seems like a just aneasy to learn language and it can do alot and and you know now it's probablyyou know burst into the Limelight withall the AI things being done in thelanguage probably because of that thosesame reasons that it's just easy forsomebody to learn like you said itapproaches C likes speeds and C is not afun language to get started with whereaswith like python you can get startedpretty quickly with some simple things so speaking of that I always suggestthat people learn by doing somethingstart creating a project so what sort ofthings would you say might beapproachable for somebody who perhaps isjust getting started with learningpython yeah yeah that's a great questionI I mean like Ione of the classes I teach is like afundamentals of python class and I tellmy students like after these three dayslike if you were to go and evaluateyourself on like popular evaluation programs online you you would beevaluated as highly proficient in pythonand a lot of them like uhthe knowledge that you've had and andbased on my interactions with you inthis course indicate that that is thecaseum however you might not feel likecomfortable like going out and likecalling yourself a python programmer perse right and I say like now that youhave this Baseline knowledge your yourchore is to go on practice right and and so you know I get DMS all the time frompeople are liketell me what to practice on and I I geta little bit of frustrated with thatbecauseum I think this is a question ofmotivation and this is this is one ofthose cognitive areas where you need tolike be able to find something that willmotivate you so so this is my generalrule of thumb but again likesort of the same thing like if I want towork out how what what's a good way to work out well I mean you can sort offorce yourself to work out for a whilebut like there are hacks that you can doto make yourself work out uh betterright like you can pay someone uh to bea your your trainer right and if you'veif you invest in money into somethingyou're going to take it more seriouslyor you can get a social group andum you know say we're going to gorunning every three days a week rightand we've set a date and there's goingto be three of us there you don't want to let down the group and so you cankind of do similar things uh with uh youknow learning something right you canyou can pay someone to do something youcan do it with someone elseum those might tend to be a little bitmore uh money intensive or or timeintensive than than people might uh wantto accept but but I think you know ifyou really want to learn somethingyou've got to make the investment to dothatum let me maybe give some things that aren't uh maybe as expensiveum one thing I'll tell uh students isum find something work related right sooftentimes I have students who are usinglike a different tool at work and andnow they want to use Python I'm likewell this is a great great thing to dobut you mean your work just paid for youto Learn Python uh your next projectinstead of doing it in whatever tool youyou prefer to use or you've used in thepast do it in Python and it might be a little bit slower or you might come intosome bumps along the way butum that's where you start practicing itand once you start practicing it that'sgoing to make it easier so I thinkthat's a great way like workum what's the incentive with work theregenerally you have a deadline right andand so that can be a good sort of carrothanging over you to uh force you to likeuh learn that another uh thing that Ifound that that people like to help themlearn is finding something that that I would call Hobby related right and andthis is why like I don't like peoplecoming and asking me like what should Iwhat should I look at and I'm like Imean I could tell you what's interestingto me but that might not be at all beinteresting to you right and and so likeI live in Utah I like to ski so I'lloften like pull up like snow data andlook at like snowfalls and like look howski seasons have looked over time rightwhich to me is interesting andfascinating but like to people who who live in a tropical climate or whatanother like who cares right and that'sthat's a good in Floridanot so useful in Florida right yeah soso find something that has an intrinsicinterest to you right and so you know ifit is skiing I mean maybe it's likedoing some reporting or some machinelearning modeling on skiing but ifyou're not interested in that you knowmaybe you're in Florida you need to lookat Hurricane models or something likethat or BeachBeach whatever that that would be rightfind something that's interesting to you that will will allow you to keep doingit rather than like here's somethingthat someone told me to do I'm going tospend an hour looking at it one time andit's like I'm not interested in that I'mnot going to doyeah yeah I like that like solve solvesome sort of problem that you havepersonally that will really motivate youum I I know that whenever I would needto learn a new program or I wanted tolearn a new program I would always youknow back when I did more desktop publishing and graphic cards I wouldalways say okay the next project I'mgoing to do in whatever program I wantto learn and it would just force you toto learn the things you needed to and Icall it like it's always easy to learn Idon't know 15 of something that youalready know how to do so if you are youknow sort of solving a problem thatyou've solved with like a bash scriptbeforeif you if you know how to solve thatproblem but now you make yourself do itin something like python that's fantastic because you're already youalready know how to do it it's just thatyou need to learn the lingo and a lot ofthese programming languages is justlearning how do the loops go with pythonit's like how does the indentation uhwork you know uh you know bash is sofunny because it's like uh you programwith like if and then you do fee and theloops are crazy so I think that you'llyou'll find that it's a lot moreenjoyable when you have some sort ofTarget uh if you're doing any sort ofautomation I mean python appears to be fantastic at processing large amounts ofdata so I I definitely am going to uhI'm doing a project where I'm I'm doinglike a GitHub action that's basicallycreating a um you know like the RSS feedfor uh these live streams that I do youknow so that I can put them up on iTunesthat's a great sort of project like youhave something that you need to do youknow you can do it in like three otherlanguages but try it on this and I think that the amazing thing about it is thatyou'll find that it's super easy to easyto pick it up pick up as a you knowthat's why people love it so much sowhat about uh tools I don't know ifyou've been using things like Chad gbtor co-pilot or other coding assistantswhat are your thoughts on that wholeworld have you used them do you findthem useful at all what do you thinkabout a beginner maybe looking at thatum in into those toolsyeah a great question um so have I used them yes I have usedthemum are they useful yes I mean that thesecan certainly augment and help youum so let me give an example uh like Isaid I I teach a fundamentals of pythonclass and one of the projects that we doat the end of that is a little bit moreinvolved it's likeum almost 200 lines of code and we makea a little tool we make a command linewrapper for it it talks to the internet we write tests for it it downloadsthings from the internetand so I was able to use chat GPT tobasically write that for me right and sothat this tool that takes a few hours towrite in the course of a few minutes Iwas able to prompt chat GPT to createthat whole code for meum so in the anecdote might be like okay so will programmers not exist in thefuture and I I don't buy that at all Idon't I don't think people need to beafraid of losing their job per seum what I what I do think this means isthat if I am an expert I can leveragethis to be more productive and againlike I said I sort of deep dove intolike how brains work and how learningworks when covet hit just to like doubledown and make myself like a more effective teacher if I could and youknow most people have very limitedamounts of capacity in their brain likeyou're working memory there's thisnotion of seven plus or minus two uhlike seven pieces of information plus orminus two is like how much you can storein your brain at a certain time andthere's also a limited capacity of howmuch deep work you can do and so if Ican offload some of that capacity intochat or co-pilot or something like thatthen that's awesome right it just gives me more time uh and more brain power toto work on things that are betterhoweverum if I'm not aware of what it's doingor if I'm not able to prompt itcorrectly if I don't know sort of thetricks or understand programming suchthat I can do that I can spend all mytime fighting with it to createsomething that will compile or work orrun or whatever that might never workand it might be uh you know more effortto to do that than it than it is to actually just sit down and write it onmy own soum I do think the tools are useful likeI said I I use themum as a beginner I would be carefulabout the notion that like I don't needto learn anything now because chat GPTwill do this all for meum I I do think you should uh find andand maybe maybe if you're creative aboutabout prompting um chat GPD can like walk you through and teach you all thesethings right I II would think that you'd have to be verycreative and about that but I would sayfind a resource that's going to help youunderstandumpython or whatever you want make sureyou have a basic knowledge of that andthen start leveraging these tools uh toaugment your capabilities rather thanusing them as sort of like I don't needto learn anything I'll just use AI to do before meso I'll tell you how I sort of have beenusing it recently I think that it's veryuseful at the beginning because it is achat engine to sort of ask it tobrainstorm ideas with you so for anapplicationum so I tend to use it I tend to use thechat apt part at the very beginningwhere I am thinking about how to buildsomething and just have a conversationwith it I also have used it successfullyto do things like um when I need to like mock up some data I just need some datato sort of test out some sort of thingthat I've already built or I need it tojust uh it's fantastic for brainstormingideas so you know where we like you youmentioned where it's hard for us to comeup with like 10 ideas for somethingwhereas GPT even if they're bad ideas itcan do a hundred no problem and then youcan say give me another 200 and then youcan start to like have a conversationabout what those hundreds ideas should focus on or you can take like here's agears one that I like and we let's talkabout this idea uh and then you canstart you know maybe doing somepseudocode and learning to build with itso it's fantastic for brainstorming alsosummarizing I think a lot of times youknow I don't know if you've run intothose kind of things uh yeah I have Ican give I can give some concreteexamples of where I I've used it rightum one is uhtaking existing code and writing it in a different language right so you couldsay I want to rewrite this bash scriptas python it tends to be pretty good atdoing things like thatanother is like you said summarizingso one of the Eternal debates ofdevelopers is how much comments you putin your code some people are like youneed to comment all your code otherpeople are like the code is the commentumand and so I mean chat makes that likeyou want to comment your code you drop in the code there and say write commentsfor this it will write comments for ityou want to write python has a notion ofa doc stream where which is adocumentation that you can attach tomethods or functions and then you caninspect them uh from your editor you cansay write the doc string for this itwill write the doctrine for that forthose who are using uh typing in pythonpython 3 has a notion of a type systemthat you can make type annotations youcan say annotate this code with types it can do that relatively well writingtests for code you can say write a testfor this function of code it can do thatrelatively wellum so so those are all places where inaddition like as a a you know person whowho creates content online yeahcertainly brainstorming helping tocreate content that sort of thing butbut from a pure coding point of viewthere are lots of asp aspects aroundcoding the documentation testing yourcode understanding your code in addition to just creating code where it is superusefulyeah that's definitely the case for meas well that I I hadn't thought aboutcommenting and it is really good atcommenting code it's really like yousaid it's not something that you maybe Idon't know uhsometimes you feel like it's the sort ofthing that is a pain to do but it's ifyou wanted to help you there I thinkthat that's kind of like any sort ofrepetitive task that you have that's the kind of thing that GPT can really helpyou and commenting is great I don't knowif you've used it for testing butperhaps that could be another futuresort of win that it can help peoplelet's talk about the the you know thethe fact that we have all these co-pilottype tools now and they're coming out inthings like Excel and you know word andother things you know you are used totaking a lot of data with things likeJupiter notebook and and Python and being able to process that datathrough computer programming how do yousee AI affecting something like thatbecause with things like Excel gettingcopilot I see perhaps a time where a lotof data analysis is the kind of thingthat something like GPT can can doyeah and um I I think yeah certainly uhyou could sayPoint GPT at like a database schema and say like write a SQL query that willanswer this question right or you couldprobably say write pandas code that willanswer this question or you couldprobably say take this pandas code andrewrite it as SQL I haven't tried thatbut I would imagine that it wouldprobably do a decent job of thatum so I mean I'll say this like no codeSolutions have been around for a longtime and the idea that like someonewho's not a programmer can like whip upsomethingum and have it work for them um I've had like clients that I've donework with who have sort of like gonedown the no code route for simple thingsthey can often work and make your lifeeasy for for more complex thingsum they kind of make it more difficultand or not work at all and I think rightnowum chat GPT can do a lot of the simplestuffummore some of the more complex stuff uh it might hallucinate or or give you ananswer that's not correctum I would imagine that it can help thatanalysis in the future and uh probablycan be a great like brainstormingpartner or uh likesort of like a co-pilot in that like youcould say oh did you cons it might saydid you consider doing this or you mightwant to go down and check out this uh path as wellumI I think those things are useful againI'm I'm not of of the idea that likechat will make programming go awayum I will make data analysis go away Idon't think so either like I I thinkthat will still be aroundum generally like II use a calculatorum I I use spell Checkers uh and so why wouldn't I use something that augmentsmy my output and makes my output betterum and then be able to work on thingsthat are more difficult or more of achallenge or drive this to automateother things for me so Ium yeah do do I see uh tooling come Imean certainly there are tools that willcome out and do analysis for youumI guess that the challenge is going to be like how how do you integrate thoseum how do you make sure that they'redoing the right thingumthey they can be useful but right now II still think that you want to be anexpert and make sure that uhyou can backstop or make sure that thehallucinations or this this isn'tdriving you down the wrong path orgiving you the wrong answer and you'rejust taking that for grantedI think that um definitely if you are concerned about like maybe the futureand what developers are going to bedoing you should try some of the toolsbecause you can definitely see withhallucinations that they're not perfectbut also if you try to get a specificresult uh especially with some of thegenerative AI like artwork type programsif you ask mid-journey to give you thething that you have in your head whatyou'll find is that it gives yousomething but it's not the thing thatyou had in your head and that I feel that the problem with that is that a lotof times what you really want is what'son your head and to do to get those sortof results those are only going to comefrom somebody who really understands thequestions I think that it could save alot of time it can get you you know partof the way there but if you really wantspecific answers you're going to needstill need a programmer maybe you'll beable to get to those things faster butum if you don't it's better to understand the problem and learn how toget the computer to do some things foryou or brainstorm with you or summarizefor you or get you to a place in thedata and maybe write some starter codebut if you get you know say like eightypercent of the answer you you're gonnaneed that 20 to come from somebody whoknows what they're doing and I thinkthat that's kind of where I see thingsgoing so let's talk specifically aboutthat what about the future what's your sort of uh I don't know likeforget about where it is now but maybethink about like five or ten years fromnow which Developer jobs or tasks do youthink are going to be say the mostaffected and what should thoseprogrammers start thinking about doingyeahumIthings are moving very quickly right nowso I I mean I I don't know that I feelconfident predicting the futureum I I I I guess but but my advice would be likeumI would want to make sure that Iunderstand uh what's going on and so ifI'm a JavaScript person or if I'm areact person or if I'm a databaseadministrator person or you knowwhatever might what I'm interested in orwhere I see my career going I would want to understand what uh the AI can do forme right now what are the limitations ofthat and thenaugment my skills around thoselimitations and then leverage the AI forwhat it can optimize for me right andand soum right nowumlike I said AI is good for coming outthings um giving you sort of boilerplate forstarting out uh can you say like writethis complete app that does thisum I mean a lot of us saw the thedemonstration where they said here's anapkin with like a a website go andwrite that websiteum which was certainly impressive butumthey didn't write like here's a YouTubewebsite go and write this right and itdidn't go out and write YouTubeum right so so I I think we do need to be awareof the limitations uh and and I I woulddouble down on sort of the higher levelskills that and and the limitations andmake sure I understand those it I thinkmaybe the more challenging part is formaybe someone who's a junior or or evenlooking to get their first jobum and I might you know if I can uhif if I'm telling a junior to go out and and do certain tasks and like I'mwriting this in an email it's like Icould write that I could send this emailto chat and it would do this for meumthat that might seem a little bit scaryright and so uhI I do think people who are like uh I'lljust use chat and I won't learn anythingthat that's a scary proposition I dothink you want to make sure that you arelearning um doesn't mean you have to be learningeverything but but I would be exploringchat or exploring AI I keep saying chatum but exploring AI exploring sort ofthe periphery of of what it's able to dowhat it's not able to do and thenum I would double down on things thatare somewhat unique to humancapabilities uh to make sure that I canthat my job is an obsoletethat's right yeah so definitely regardless of what language youunderstand you need to be exploring howthis thing can help you you know getthere quicker how it can assist you uhwhat it doesn't do so if you haven'ttried itum you know even something like copilotseeing what it can do I think those arejust you know I mean this is why youlearned all kinds of languages to stayrelevantum you know you don't want to be I don'tknow the last person using punch cardsor something like that yeah I mean if your head is in the sand yeah if yourhead is in the sand and you're like yeahI'll just sort of wait and see how thisplays out right now I I think you youyou really should uh consider liketrying these things out and seeing whatthey can doum I I get that like you know authors oror you know there's there's uproarersabout people writing books and like ohthis is going to put authors out ofbusiness or people wrote books that haveart it's completely generated by Ai andand the artists are upset that they'll go out of business or whatnotum yeah you should you should see whatwhat it can do and and and also see howit can help you right I mean thesethings are are very powerful uh and ifif you aren't taking advantage of themor learning how to take advantage ofthem I think you will be replaced bysomeone who can take advantage of themthat that would be that's my my thoughtaround thatyeah Mike actually mentioned that youwill not lose your job to AI it will lose it to some developers who areleveraging all of these tools because itit does make you faster and able to domore and the way that I describe it isuh you know remember whenshipping used to take like four to sixweeks and now she if if you don't getsomething within a couple days it'sreally a horrible experience so what wegot when we figured out how to doshipping better was more shipping in thesame way like you know when we have tools that make usdo more developer we're gonna get moredevelopment like more uh so I thinkthere could probably be more of a demandthink about how many more people areemployed in delivery now than were everemployed before it was just one companythe the post office doing everything andnow we have lots of people deliveringdoing last mile delivery that sort ofthing I think that this could open up anopportunity you know think aboutspreadsheets even like when spreadsheetsfirst came out did we get less accountants no we got more we got morelike we solved bigger problems so Ithink that those are some of the thingsthat we're going to see and that's whylike you I think I have a sort of apositive outlook in the future so whatabout this uh sort of ability of thegpts and other AIS being able to passthings like lead code tests and I was ateacher and somebody who trains in Corpin the corporate environment so where doyou see like skill validation going is that going to be harder is it what areyou know what how's that going to workin the futureyeah I I mean part of your question islike how interviews work and andum you knowtypically job interviews have thisnotion like we're going to put youthrough the ringer and make sure thatyou can do these certain types of codingwhether or not they're actually relevantor applicable to the job that's justwhat everyone doesumyeah and so um at least my understanding and Ihaven't played around specifically withthis is the like at least for the themore simple types of these leak codeproblemsumAI can can do them relatively easilyumand I mean cheatYou'll Always Find people who arecheating right and so IIthink eventually that will be exposed soso I don't know that like younecessarily want to go out of your way to likeumfocus on on the cheating per se butum yeah I think you can't really preventpeople from using these tools I see thata lot like uh universities and peoplesaying oh we're going to ban GPT you sawItaly actually ban GPT or chat or openAI Chi gbt from from its usage I meanlike you said you don't want to stickyour hand in the sand uh you want to seewhat they can do see how they can helpyou know yeah put guardrails aroundthings I think that that sort of a I don't know desire to control it's notgonna it's I'm never gonna work it neverworks yeah well I mean the sad truth isis that like chat can write an essaybetter than 80 percent of of high schoolseniors so I mean if you have a toolthat in in 30 seconds can write an essaybetter than most of you and your friendsum why would you not use that and I Ithink you know sadly a lot of uh Techinterviews are sort of similar to that where we put people in a situationthat's not at all so how their work willbe right like you're developing this onyour own using your own brain powerwithout any outside assistance where inpractice you're working with othersyou're collaborating with them you'llhave access to AI tools to do this aswellum so I'll try and uh go not not reallyI just say that is what it is rightum I'll say more as like a teacher and a trainer like how do I see uh validationgoing and again from my point of viewum you knowI'm gonna say like I think it'simportant to have the base understandingI think if you don't have baseunderstanding of certain technology likeusing AI tools isn't necessarily goingto help you so much so you do want thatso I don't see that that's putting likeme as a corporate trainer out of a jobum uh as as far as like skill validation any skill validation itkind of a hard thing and of itself andumI think people will gain whatever uhmetric you want I I I do like to saylike you know when I'm teaching inperson it's pretty easy for me to likeafter a day or so of of being with aclass I can rank order people and saylike these people are the mostproficient in what I'm teachingum does that mean that they're going tobe the best colleagues to work with I mean there are other things that go intothat as well right and so I I do think awell-rounded developer or an employeeum you know is someone who not only cantake advantage of technology and isefficient with their time but it is alsosomeone who works well with others rightand I don't necessarily know thatumyou know just because you can use uhchat GPT that that's necessarilyindicative of you being a good employee per seyeah that's true and um I think thatit's definitely changing everything andfor me I I find that I use it best uh todo things around my training you knowhelp me promote courses sometimes it'sgreat for like a thumbnail so yeah youuse it as much as you can and see um seewhat it does for you soum talk about some of the things thatyou uh I know I know you have a websiteI guess meta snake.com is the website where you have a lot of your stuffum what do you have there and what aresome things that people should look uhyou knowwhether it's you that that are doing itor other things that you're trainingyou've written a bunch of books like forexample so uh what kind of what kind ofwhat what do you want people to knowaboutsure yeah uh metasnake.com that's sortof my corporate siteum so folks who are looking forcorporate training uh getting their teamup to speed that sort of thing can go there and then I have a store attachedto that and so folks who want one-off ondemand courses or booksum I I have that around there and thenI'm also available on other platforms aswell like LinkedIn learningum just released a course on LinkedInlearningum actually looking to do another courseon LinkedIn learning as well soum so that's why I'm really interestedon LinkedIn it was like a course that isfocused on finance can you talk a little bit about kind of your approach thereand why you kind of went after thatspecific Marketyeahum so as we as you mentioned python ispretty big in a lot of places and oneplace where there's a lot of data and alot of like interest around modeling andand machine learning is uh uh in infinance and so there's a lot of use of Imean some of the biggest uh deploymentsof python are financial institutions so um that's a place where there's a lot ofdemand for Python andum my experience likeum so one of the books I wrote is thiseffective pandas book which uh this bookuhpandas is a tool for manip it is kind oflike Excel for python uh manipulating uhtabular data and uh I have some prettystrong opinions about uh how people should use pandas and and so you cansort of think of that course as meapplying my opinionsum around a particular use case of likelooking at financial data right and so Ithink people who are interested infinancial data or sort of looking tomove from Excel to python are going tofind that course useful also people whoaren't necessarily even interested infinancial data but are are reallylooking at like how does someone who's looking at this data what's theirthought processes are going through ithow do they optimize their code how dothey debug their code uh there's goingto be a concrete example of doing thatfrom a finance aspectyeah it sounds like a real practicalsort of use for for this type of uhtechnology so I really enjoyed thatabout that approach I never never seenthat before uh and you know havingworked at a newspaper I know that one of the strongest things that we did wasData analysis and then you know ourGraphics Department was able to you knowconvert that and make it into somethingvery visual so I really like that thatcourse existed so I can't wait for uhyour next course uh as well uh fantasticcontent I watched a little bit of yourFinance course but it wasn't it wasn'tuh enough for me to really understandlike I don't know enough about Jupiternotebooks although that looks amazing and I didn't know enough to I think getthe most out of it so super cool thankyou so much uh for coming I think wehave maybe a couple of questions that wecan answer before we run out of timeokay yeah I have um I have first of allum Adam just wanted to say you found afellow Utah in here soI happen to see people yeah good to seepeople coming in from different placesand let's see um you know talking about being able towork on projects that me are meaningfulto you I like space stuff and so he dida project using a meteor beater strikedata set that's super interesting it'sbasically use any practical purposes butit was a great way to learn five ofdifferent things that's that's awesome Imean that's that's actually exactly whatyou were talking about right findsomething that you're passionate aboutyeah useless for others but likeprobably super useful for Adam asthey're trying that out yeah yeah and I think it's useful if it's helping youyou know learn something that's that's afantastic useyeah we also had a Victoria alsomentioned that uh let's see she islearning how you switch app gbt can beand learning how you switch at BTD canbe in job hunting it really helps whenyou get a mental blog writing coversletters so talk a little bit about youknow howhow relevant do you think it would befor somebody looking for a job out thereto have something like not necessarily chat GPT but all these new things thatyou're that you're doing you fightyourself on a huge market for for thesejobs nowuh as far as like is there a market forprompt engineering or or can yeah whatwhat are what kind ofyeah well Chad gbt yeah I guess it'slike a really cool thing for actuallyjob searching I know that we released atool now that helps you on LinkedIn likeyou know one of the things that I struggle with like how do I write myaccomplishments that's a fantastic thingI mean you know for somebody like me whoyou know who's uh English isn't my firstlanguage so that's really helpfulbecause that's one area where it candefinitely uh do a better job than meum you know even even where English isyour first language right and and I'msomeone who's written multiple booksit's likeum you know if I need to write copy orif I need to write a cover letter rightit can it can crank that out and it looks pretty good like I said it's goingto be better than most uh High Schoolseniors could write but you can alsolike Riff on that you say like welllet's let's write this in a morepersuasive way or in a moretongue-in-cheek way and it will do thatfor you and and so uh does that meanthat you should just wholesale copy andpaste that I mean maybeum generally when I've created contentuh derived from AI uh it's been I wouldsay heavily inspired by the AI but havealso generally heavily edited afterwards um to put sort of a spin on it that thatI likeum yeah so I mean but I I think you canalso as far as like if you're interestedin like job hunting like youfrom the brainstorming aspect right youcould say like let's list uh 20companies that are relevant aroundwhatever aspect right and thenumand sort of use that to brainstorm maybecompanies that you should apply to uh you could use that to to do someOutreachum as wellumcertainly things that you could do II'm a huge fan of of publishing contentand getting content out there not justlike for job searching but I think youryour job search will be more effectiveif you have something you can point tolike a portfolio or a website or or thatsort of thing and and so if you canLeverage you know any of these things that canlike minimize that friction to getstarted on a project that's going tomake your your job prospects betterum you know I yeah definitely likepersonalizing like maybe your coverletters to the you know specificindividuals I think they always say thatI I know like a couple of weeks ago Imet with Logan Kilpatrick from from openAi and he mentioned that he he applied500 times to his first like a machinelearning job and uh so I can imagine hewould have found it handy to have something like GPT that can help with alot of that writing yeah and apply socool awesome well thank you so much Ijust want to mention next week uh wehave Burke Holland uh coming into theshow he's a principal developer advocateof Microsoft who's focusing on VisualStudio code so we're going to be talkingabout how that product is developed andalso what's happening as they attempt tointegrate tools like chat GPT so one ofthe big things is I don't know if you'veseen uh copilot X is the sort of the next generation of co-pilot basicallyadding the chat GPT features into thepopular editors that should be prettypretty awesome so thank you so much Mattfor coming I appreciate your time and uhmake sure you check out Matt's LinkedInpage our URLs are there at the bottomand we'll see you next week take care mypleasure thanks Ray, Upskilling and Reskilling for Artificial Intelligence, Ray Villalobos, 01gya4zf5ge15q0xax03txh3mk.